Artificial Intelligence-Based Battery Management System Enhancement Method and System

Abstract

This invention discloses an artificial intelligence-based battery management system enhancement method and system. The method includes heterogeneous data fusion acquisition, dynamic digital twin modeling, dynamic charge-discharge enhancement, multi-dimensional safety situation awareness enhancement, and battery management system enhancement. This invention relates to the field of battery management enhancement technology, employing a comprehensive battery management system enhancement method that combines dynamic digital twin modeling optimization, dynamic charge-discharge strategy optimization, and multi-dimensional safety situation awareness optimization. Through multi-dimensional enhancement and optimization, it improves the dynamics, prediction accuracy, and overall strategy adjustment efficiency of the battery management system. A four-level digital twin ensemble model is used for dynamic digital twin modeling; a reinforcement learning method combining dual-delay deep deterministic policy gradient and reward function updates is used for dynamic charge-discharge enhancement; and a spatiotemporal graph convolutional network incorporating the battery microstructure evolution map is used for multi-dimensional safety situation awareness enhancement.

Description

Technical Field

[0001] This invention relates to the field of battery management enhancement technology, specifically to a battery management system enhancement method and system based on artificial intelligence. Background Technology

[0002] The AI-based battery management system enhancement method and system optimizes battery management and monitoring by combining advanced artificial intelligence technology. Its core function is to enhance the overall performance of the battery system. The system can effectively improve battery efficiency, extend battery life, predict potential risks, and adjust the battery's operating state in real time, ensuring battery safety and reliability. It is widely used in electric vehicles, energy storage equipment and other fields.

[0003] However, in existing battery management systems, traditional systems mainly monitor voltage, temperature and current and perform charge and discharge control to achieve balanced management and safety optimization. However, existing technical solutions are not accurate enough in predicting battery aging, and at the same time, they are not adaptable to dynamic operating conditions and are difficult to cope with changing battery working environments. Furthermore, their charge and discharge strategies are not efficient enough, which leads to technical problems in the intelligence of existing battery management systems that need to be enhanced and optimized.

[0004] In the existing dynamic digital twin modeling process, there is a technical problem that traditional battery digital twin modeling takes into account less of the internal and external conditions of the battery, making it difficult to fully reflect the changes in the battery's state and health status. As a result, the modeling at both the macro and micro levels is unable to reflect the dynamic nature of the battery.

[0005] In existing dynamic charge and discharge enhancement methods, there are shortcomings in the intelligence and dynamism of the battery charge and discharge adjustment strategies combined with dynamic digital modeling. Furthermore, this leads to the technical problem that changes in the battery operating environment are faster than the battery management adjustment strategies.

[0006] Among the existing multidimensional security situation awareness enhancement methods, there is a problem that most existing battery safety management systems rely on traditional prediction methods, such as lifetime prediction based on data during battery charging and discharging. However, many traditional methods cannot effectively capture changes at the micro level of the battery and early signals of faults. These systems lack the technical ability to dynamically and intelligently perceive and analyze micro changes throughout the entire battery life cycle. Summary of the Invention

[0007] To address the above issues and overcome the shortcomings of existing technologies, this invention provides an AI-based battery management system enhancement method and system. Traditional battery management systems primarily monitor voltage, temperature, and current for charge / discharge control, achieving equalization management and safety optimization. However, existing solutions lack accuracy in predicting battery aging and exhibit poor adaptability to dynamic operating conditions, failing to cope with changing battery environments. Furthermore, their charge / discharge strategies are inefficient, leading to a need for enhanced and optimized intelligence in existing battery management systems. This invention creatively employs a comprehensive battery management system enhancement method combining dynamic digital twin modeling optimization, dynamic charge / discharge strategy optimization, and multi-dimensional safety situation awareness optimization. Through multi-dimensional enhancement and optimization, it improves the dynamics, prediction accuracy, and overall strategy adjustment efficiency of the battery management system. Furthermore, existing dynamic digital twin modeling methods often lack integration with the internal and external conditions of the battery, failing to fully reflect changes in battery state and health. This results in both macroscopic and microscopic modeling failing to capture the battery's dynamics. This invention creatively addresses this issue. This paper employs a four-level digital twin integrated model for dynamic digital twin modeling. By introducing multi-scale coupling in electronic structure calculations, non-uniform potential distribution correction in the phase-field model, and methods for calculating double-layer effects and dynamic equivalent resistance from the perspectives of single-cell and system integration, the dynamics of the battery digital twin model are enhanced, forming an adaptive and dynamically adjustable complete digital twin system. This facilitates subsequent battery health management, charge / discharge optimization, and fault prediction. Addressing the issue that existing dynamic charge / discharge enhancement methods lack sufficient intelligence and dynamism in their combined dynamic digital modeling battery charge / discharge adjustment strategies, and that this leads to faster changes in the battery's operating environment compared to battery management adjustment strategies, this paper creatively combines a reinforcement learning method with dual-delay deep deterministic policy gradient and reward function updates for dynamic charge / discharge enhancement. Through a dual-delay update strategy combined with refined reward function improvements, the intelligence and dynamism of the battery charge / discharge adjustment strategy are enhanced from the bottom up. This not only helps extend battery life and improve charge / discharge efficiency but also ensures the battery operates within safe limits, improving the overall performance and reliability of the battery system and battery management system.To address the shortcomings of existing multidimensional safety situation awareness enhancement methods, which largely rely on traditional prediction methods such as lifetime prediction based on data from battery charging and discharging processes, many traditional methods fail to effectively capture microscopic changes and early signs of failures. These systems lack the ability to dynamically and intelligently perceive and analyze microscopic changes throughout the battery's entire lifecycle. This solution creatively employs a spatiotemporal graph convolutional network combining battery microstructure evolution maps for multidimensional safety situation awareness enhancement. Through dynamic monitoring and multidimensional comprehensive analysis, it provides more refined, comprehensive, and dynamic safety monitoring and management capabilities.

[0008] The technical solution adopted by this invention is as follows: The battery management system enhancement method based on artificial intelligence provided by this invention includes the following steps:

[0009] Step S1: Heterogeneous data fusion and acquisition;

[0010] Step S2: Dynamic digital twin modeling;

[0011] Step S3: Enhanced dynamic charge and discharge;

[0012] Step S4: Enhance multi-dimensional security situation awareness;

[0013] Step S5: Battery management system enhancement.

[0014] Further, in step S1, the heterogeneous data fusion acquisition is used to collect the raw data related to the battery state required for the analysis of the battery management system. Specifically, through heterogeneous data fusion acquisition, the raw dataset of battery state management is collected, and preliminary optimization and preprocessing are performed to obtain the enhanced dataset of the battery management system.

[0015] The heterogeneous data fusion acquisition specifically involves acquiring battery state data through the deployment of sensor arrays and dynamic feature acquisition of electrochemical impedance spectroscopy.

[0016] The original dataset for battery state management includes battery voltage data, current data, temperature data, electrochemical data, mechanical support data, solid electrolyte interface film evolution data, and multiphysics fusion data.

[0017] The preliminary optimization and preprocessing steps include data denoising, missing value handling, outlier correction, data standardization, time series data stabilization, and feature selection. By performing the preliminary optimization and preprocessing on the original battery state management dataset, an enhanced battery management system dataset is obtained.

[0018] The battery management system enhanced dataset includes optimized time-series data, derived feature data, multi-physics fusion data, fusion index data, evolutionary feature data, labeled data, and synchronous multidimensional data.

[0019] Further, in step S2, the dynamic digital twin modeling is used to construct a four-level digital twin model for battery state management. Specifically, based on the battery management system enhanced dataset, a four-level resolution twin model construction method is used to perform dynamic digital twin modeling to obtain a battery management enhanced digital twin model.

[0020] The battery management system augmentation dataset includes quantum chemical scale models, electrode particle scale models, single cell scale models, and system integration scale models.

[0021] The quantum chemical scale model is used to introduce a multi-scale coupling strategy for comprehensive modeling of electronic states and internal and external environmental parameters of the battery.

[0022] The electrode particle-scale model is used to introduce non-uniform television distribution correction to model the potential spatial variation within the electrode particles and to improve the traditional phase-field model.

[0023] The single-cell scale model is used to introduce the double-layer effect to improve the modeling of the influence of the double layer on the current distribution in the traditional motor model.

[0024] The system integration scale model is used to introduce a dynamic resistance correction method based on state estimation to improve the dynamic adjustment model of the equivalent resistance model.

[0025] The steps of using a four-level resolution twin model construction method to perform dynamic digital twin modeling and obtain an enhanced digital twin model for battery management include:

[0026] Step S21: Electronic structure modeling, used to construct a quantum chemical scale model, specifically by introducing a multi-scale coupling strategy to calculate the electronic structure and obtain the reference value of the electronic structure Hamiltonian;

[0027] Step S22: Particle electrochemical reaction modeling, used to construct an electrode particle-scale model, specifically by introducing a potential correction coefficient into the standard phase field model to model the particle electrochemical reaction and obtain reference values ​​for ion diffusion and reaction rate;

[0028] Step S23: Electrode current density modeling, used to construct a single cell scale model, specifically by introducing double layer effect correction into the standard porous motor model, performing electrode current density modeling, and obtaining an improved electrode current density model.

[0029] Step S24: Battery equivalent circuit modeling, used to construct a system integration scale model. Specifically, in the standard equivalent circuit model, the resistance is treated as a dynamic variable and dynamic correction is performed to obtain the battery equivalent circuit model. The calculation formula is as follows:

[0030] ;

[0031] In the formula, V OC This is the open-circuit voltage value, I is the battery current value, and R is the open-circuit voltage value. eq V(t) is the equivalent resistance value, and V(t) is the battery operating voltage value.

[0032] It is a battery equivalent circuit modeling model, which is based on the equivalent resistance value R. eq Dynamic resistance correction is performed, and a system integration scale model is constructed. eq (t) is the equivalent resistance value considered as a dynamic variable, and R1 is the base resistance, used to represent the resistance value of the battery under normal operating conditions. This is the dynamic resistance change term, SOC(t) is the battery's state of charge, and SOH(t) is the battery's state of health. This refers to the battery temperature status;

[0033] Step S25: Four-level integration, used to construct a four-level resolution twin model, specifically by integrating the electronic structure modeling, the particle electrochemical reaction modeling, the electrode current density modeling, and the battery equivalent circuit modeling to obtain a battery management enhanced digital twin model.

[0034] Further, in step S3, the dynamic charge-discharge enhancement is used to dynamically enhance battery charge-discharge management. Specifically, based on the battery management system enhancement dataset and the battery management enhancement digital twin model, a reinforcement learning method combining dual-delay deep deterministic policy gradient and reward function updates is used to perform dynamic charge-discharge enhancement, resulting in a dynamic charge-discharge enhancement reference policy. This includes the following steps:

[0035] Step S31: Construct a dual-delay reinforcement learning framework. Specifically, construct a standard reinforcement learning environment parameter model, adopt a deep deterministic gradient strategy, combine the value function and policy function to set up the basic framework of reinforcement learning, and update the parameters of the value network and policy network by constructing a dual-delay update strategy.

[0036] Step S32: Dynamic charge and discharge enhancement reward improvement, specifically by designing a multi-dimensional reward function to comprehensively integrate the battery's lifespan, efficiency, and safety aspects, to improve the dynamic charge and discharge enhancement reward, thereby obtaining a dynamic charge and discharge enhancement reward function, and then performing reinforcement learning training based on the dynamic charge and discharge enhancement reward function;

[0037] The calculation formula for the multidimensional reward function is as follows:

[0038] ;

[0039] In the formula, R w It is a multidimensional reward function, used as the reward function for the dual-delay reinforcement learning framework described in step S31. It is a weighted factor for battery life. It is a battery life gain term. It is the weight of charge and discharge efficiency. It is the charge / discharge efficiency item. S is the safety factor weight, and S is the safety factor term;

[0040] Step S33: Iterative update of dynamic charging and discharging strategy, specifically by applying the multidimensional reward function to the dual-delay reinforcement learning framework to iteratively update the dynamic charging and discharging strategy, and obtain iteratively updated dynamic charging and discharging strategy data;

[0041] Step S34: Dynamic charge-discharge enhancement, specifically, involves iterative training of the dynamic charge-discharge enhancement model through the dual-delay reinforcement learning framework, the dynamic charge-discharge enhancement reward improvement, and the dynamic charge-discharge strategy iteration update, to obtain the dynamic charge-discharge enhancement model. DH And by using the dynamic charge-discharge enhancement model Model DH Based on the battery management system enhancement dataset and the battery management enhancement digital twin model, dynamic charge and discharge enhancement is performed to obtain a dynamic charge and discharge enhancement reference strategy.

[0042] The dynamic charge-discharge enhancement reference strategy specifically includes battery health status data, charge-discharge control data, and optimized charge-discharge data.

[0043] Further, in step S4, the multi-dimensional safety situation awareness enhancement is used to enhance the battery safety situation awareness management in multiple dimensions. Specifically, based on the battery management system enhanced dataset and the battery management enhanced digital twin model, a spatiotemporal graph convolutional network combining the battery microstructure evolution map is used to enhance the multi-dimensional safety situation awareness and obtain multi-dimensional safety situation awareness reference data, including the following steps:

[0044] Step S41: Construct a spatiotemporal graph convolutional network to capture the spatiotemporal dependencies and basic features in the battery microstructure. Specifically, this involves constructing battery microstructure graph data and building a standard spatiotemporal graph convolutional network as the foundational network for enhancing multidimensional security situation awareness.

[0045] Step S42: Construct a battery microstructure evolution map to model the evolution process of the battery's internal structure and support dynamic prediction. Specifically, construct a basic model of battery microstructure state evolution and predict the evolution of battery micro-deformation, lithium plating, and solid electrolyte interface film growth based on the basic model of battery microstructure state evolution.

[0046] Step S43: Early fault propagation modeling, used to detect microstructure changes and implement early warning prediction of battery faults. Specifically, early fault propagation modeling is performed through the coupling effect of current, temperature and state of charge in the battery, and early warning prediction of battery faults is performed through the early fault propagation modeling.

[0047] Step S44: Training the multi-dimensional security situation awareness enhancement model. Specifically, this involves training the multi-dimensional security situation awareness enhancement model using the spatiotemporal graph convolutional basic network, the battery microstructure evolution map, and the early fault propagation modeling, to obtain the multi-dimensional security situation awareness enhancement model. EH ;

[0048] Step S45: Enhance multi-dimensional security situation awareness, specifically by using the aforementioned multi-dimensional security situation awareness enhancement model. EH Based on the battery management system enhanced dataset and the battery management enhanced digital twin model, multi-dimensional security situation awareness enhancement is performed to obtain multi-dimensional security situation awareness reference data.

[0049] The multi-dimensional safety situation awareness reference data specifically includes battery prediction parameters, fault propagation prediction rate, and battery safety status change prediction reference data.

[0050] Further, in step S5, the battery management system enhancement is used to comprehensively enhance the battery management system by combining battery charge and discharge management and safety situation management. Specifically, it involves constructing the basic mathematical model required for the battery management system enhancement through dynamic digital twin modeling, and comprehensively enhancing the battery management system through the dynamic charge and discharge enhancement and the multi-dimensional safety situation awareness enhancement, thereby obtaining a dynamic charge and discharge enhancement reference strategy and multi-dimensional safety situation awareness reference data. Finally, by combining the dynamic charge and discharge enhancement reference strategy and the multi-dimensional safety situation awareness reference data, comprehensive enhancement reference data for the battery management system is obtained.

[0051] The present invention provides an artificial intelligence-based battery management system enhancement system, which includes a quantum sensing module, a digital twin module, a dynamic optimization module, a safety situation awareness module, and a battery management system enhancement module.

[0052] The quantum sensing module is used for heterogeneous data fusion acquisition. Through heterogeneous data fusion acquisition, an enhanced dataset of the battery management system is obtained, and the enhanced dataset of the battery management system is sent to the digital twin module, the dynamic optimization module, and the security situation awareness module.

[0053] The digital twin module is used for dynamic digital twin modeling. Through dynamic digital twin modeling, an enhanced digital twin model for battery management is obtained, and the enhanced digital twin model for battery management is applied to the dynamic optimization module and the security situation awareness module.

[0054] The dynamic optimization module is used for dynamic charge and discharge enhancement. Through dynamic charge and discharge enhancement, a dynamic charge and discharge enhancement reference strategy is obtained, and the dynamic charge and discharge enhancement reference strategy is sent to the battery management system enhancement module.

[0055] The security situation awareness module is used for multi-dimensional security situation awareness enhancement. Through multi-dimensional security situation awareness enhancement, multi-dimensional security situation awareness reference data is obtained, and the multi-dimensional security situation awareness reference data is sent to the battery management system enhancement module.

[0056] The battery management system enhancement module is used to enhance the battery management system and obtain comprehensive enhancement reference data for the battery management system through the enhancement.

[0057] The beneficial effects achieved by the present invention using the above solution are as follows:

[0058] (1) In the existing battery management system, the traditional system mainly monitors voltage, temperature and current and performs charge and discharge control to achieve equalization management and safety optimization. However, the existing technical solutions are not accurate enough in predicting battery aging. At the same time, the adaptability to dynamic operating conditions is poor and it is difficult to cope with the changing battery working environment. The charge and discharge strategies are also not efficient enough, which leads to the technical problem that the existing battery management system needs to be enhanced and optimized in terms of intelligence. This solution creatively adopts a comprehensive battery management system management enhancement method that combines dynamic digital twin modeling optimization, dynamic charge and discharge strategy optimization and multi-dimensional safety situation perception optimization. Through the enhancement and optimization of multiple dimensions, the dynamics, prediction accuracy and overall strategy adjustment efficiency of the battery management system are improved.

[0059] (2) In the existing dynamic digital twin modeling process, there is a problem that traditional battery digital twin modeling has less integration with the internal and external conditions of the battery, making it difficult to fully reflect the changes in the battery's state and health status, resulting in the modeling at both the macro and micro levels being unable to reflect the dynamics of the battery. This solution creatively adopts a four-level digital twin integrated model for dynamic digital twin modeling. By introducing multi-scale coupling in electronic structure calculation, introducing non-uniform potential distribution correction in phase field model, and introducing double layer effect and dynamic equivalent resistance calculation methods from the perspective of single cell and system integration, the dynamics of the battery digital twin model are improved, thus forming an adaptive and dynamically adjustable complete digital twin system, which is helpful for subsequent battery health management, charge and discharge optimization and fault prediction.

[0060] (3) In view of the fact that the existing dynamic charge and discharge enhancement methods are insufficient in terms of intelligence and dynamism of the battery charge and discharge adjustment strategy combined with dynamic digital modeling, and that this further leads to the technical problem that the battery working environment changes faster than the battery management adjustment strategy, this solution creatively combines a reinforcement learning method of dual-delay deep deterministic strategy gradient and reward function update to perform dynamic charge and discharge enhancement. By combining the dual-delay update strategy with refined reward function improvement, the intelligence and dynamism of the battery charge and discharge adjustment strategy are improved from the bottom up. This not only helps to extend battery life and improve charge and discharge efficiency, but also ensures that the battery works within a safe range, thereby improving the overall performance and reliability of the battery system and the battery management system.

[0061] (4) In view of the existing multidimensional safety situation awareness enhancement methods, most existing battery safety management systems rely on traditional prediction methods, such as life prediction based on data during battery charging and discharging. However, many traditional methods cannot effectively capture changes at the micro level of the battery and early signals of faults. These systems lack the ability to dynamically and intelligently perceive and analyze micro changes throughout the battery's entire life cycle. This solution creatively adopts a spatiotemporal graph convolutional network that combines the battery microstructure evolution map to enhance multidimensional safety situation awareness. Through dynamic monitoring and multidimensional comprehensive analysis, it provides more refined, comprehensive, and dynamic safety monitoring and management capabilities. Attached Figure Description

[0062] Figure 1 A flowchart illustrating the AI-based battery management system enhancement method provided by this invention;

[0063] Figure 2 A schematic diagram of the AI-based battery management system enhancement system provided by the present invention;

[0064] Figure 3A flowchart illustrating the process of dynamic digital twin modeling in step S2;

[0065] Figure 4 A schematic diagram of the process for enhancing dynamic charge and discharge in step S3;

[0066] Figure 5 This is a flowchart illustrating the process of enhancing multi-dimensional security situational awareness in step S4.

[0067] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0068] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0069] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0070] Example 1, see Figure 1 The present invention provides an artificial intelligence-based battery management system enhancement method, which includes the following steps:

[0071] Step S1: Heterogeneous data fusion and acquisition;

[0072] Step S2: Dynamic digital twin modeling;

[0073] Step S3: Enhanced dynamic charge and discharge;

[0074] Step S4: Enhance multi-dimensional security situation awareness;

[0075] Step S5: Battery management system enhancement.

[0076] By performing the above operations, this solution addresses the technical problem in existing battery management systems. Traditional systems primarily monitor voltage, temperature, and current for charge and discharge control, achieving balanced management and safety optimization. However, existing technologies are not accurate enough in predicting battery aging, have poor adaptability to dynamic operating conditions, struggle to cope with changing battery working environments, and their charge and discharge strategies are not efficient enough. This leads to the need for enhancement and optimization of the intelligence of existing battery management systems. This solution creatively adopts a comprehensive battery management system enhancement method that combines dynamic digital twin modeling optimization, dynamic charge and discharge strategy optimization, and multi-dimensional safety situation awareness optimization. Through enhancement and optimization in multiple dimensions, it improves the dynamics, prediction accuracy, and overall strategy adjustment efficiency of the battery management system.

[0077] Example 2, see Figure 1 and Figure 2 In step S1, the heterogeneous data fusion acquisition is used to collect the raw data related to the battery state required for the analysis of the battery management system. Specifically, through heterogeneous data fusion acquisition, the raw dataset of battery state management is collected and pre-optimized and pre-processed to obtain the enhanced dataset of the battery management system.

[0078] The heterogeneous data fusion acquisition specifically involves acquiring battery state data through the deployment of sensor arrays and dynamic feature acquisition of electrochemical impedance spectroscopy.

[0079] The original dataset for battery state management includes battery voltage data, current data, temperature data, electrochemical data, mechanical support data, solid electrolyte interface film evolution data, and multiphysics fusion data.

[0080] The battery voltage data specifically includes battery terminal voltage data and battery internal voltage distribution sensing data;

[0081] The current data specifically includes charging current data, discharging current data, and current fluctuation data;

[0082] The temperature data specifically includes battery surface temperature data and battery internal sensor temperature data;

[0083] The electrochemical data specifically includes electrochemical impedance spectroscopy data and ion migration rate reference data;

[0084] The mechanical support data specifically includes battery expansion data and battery contraction data;

[0085] The solid electrolyte interface membrane evolution data specifically includes membrane thickness data and membrane composition analysis data;

[0086] The multiphysics fusion data specifically refers to thermo-mechanical-electro-chemical coupled integrated data, including multi-dimensional data fusion of thermal field, mechanical field, electric field and chemical reaction data;

[0087] The preliminary optimization and preprocessing steps include data denoising, missing value handling, outlier correction, data standardization, time series data stabilization, and feature selection. By performing the preliminary optimization and preprocessing on the original battery state management dataset, an enhanced battery management system dataset is obtained.

[0088] The battery management system enhanced dataset includes optimized time-series data, derived feature data, multi-physics fusion data, fusion index data, evolutionary feature data, labeled data, and synchronous multidimensional data;

[0089] The optimized timing data specifically refers to optimized battery voltage, current, and temperature data;

[0090] The derived feature data specifically refers to the optimized charge / discharge rate, thermal change rate, and internal resistance change rate feature data.

[0091] The multiphysics fusion data specifically refers to the fused thermal, mechanical, electrical, and chemical data, which are used to reflect the complex behavior of the battery under different operating conditions.

[0092] The fusion index data specifically refers to the optimized electrochemical tenant spectrum data and physical data;

[0093] The evolutionary characteristic data specifically refers to the optimized membrane thickness data and membrane composition analysis data;

[0094] The tagged data specifically refers to the data after the battery status is tagged, and the tags include normal, overcharged, over-discharged and aged;

[0095] The synchronized multidimensional data specifically refers to the battery voltage, current, and temperature data after the synchronization timestamp.

[0096] Example 3, see Figure 1 , Figure 2 and Figure 3 This embodiment is based on the above embodiment. In step S2, the dynamic digital twin modeling is used to construct a four-level digital twin model for battery state management. Specifically, based on the battery management system enhanced dataset, a four-level resolution twin model construction method is used to perform dynamic digital twin modeling to obtain a battery management enhanced digital twin model.

[0097] The battery management system augmentation dataset includes quantum chemical scale models, electrode particle scale models, single cell scale models, and system integration scale models.

[0098] The quantum chemical scale model is used to introduce a multi-scale coupling strategy for comprehensive modeling of electronic states and internal and external environmental parameters of the battery.

[0099] The electrode particle-scale model is used to introduce non-uniform television distribution correction to model the potential spatial variation within the electrode particles and to improve the traditional phase-field model.

[0100] The single-cell scale model is used to introduce the double-layer effect to improve the modeling of the influence of the double layer on the current distribution in the traditional motor model.

[0101] The system integration scale model is used to introduce a dynamic resistance correction method based on state estimation to improve the dynamic adjustment model of the equivalent resistance model.

[0102] The steps of using a four-level resolution twin model construction method to perform dynamic digital twin modeling and obtain an enhanced digital twin model for battery management include:

[0103] Step S21: Electronic structure modeling, used to construct a quantum chemical scale model. Specifically, this involves calculating the electronic structure by introducing a multi-scale coupling strategy to obtain the reference value of the electronic structure Hamiltonian. The calculation formula is as follows:

[0104] ;

[0105] In the formula, H is the Hamiltonian, used to describe the total energy of the system. It is the wave function, used to represent the quantum state and electronic state of the system; E is the total energy of the system; T is the kinetic energy, used to represent the kinetic energy state of the particles; V is the wave function. ext (r) is the potential energy influenced by the external electric field, V elec (r) is the electron potential within the electrode, and r is the potential energy parameter, V. therm (T) is the temperature potential energy, used to represent the thermal state inside the battery;

[0106] Step S22: Particle electrochemical reaction modeling, used to construct an electrode particle-scale model. Specifically, a potential correction coefficient is introduced into the standard phase-field model to model the particle electrochemical reaction, obtaining reference values ​​for ion diffusion and reaction rates. The calculation formula is as follows:

[0107] ;

[0108] In the formula, C is the ion concentration, and t is the time variable. The overall approach uses partial derivatives of ion concentrations to represent reference values ​​for ion diffusion and reaction rates. D is the diffusion coefficient, and R is the reaction rate. It is the Laplace operator for the spatial distribution of ion concentration diffusion. It is the gradient operator. It is a potential correction factor used to represent the influence of the electric field in the electrode material. It is the electric potential within the electrode particles. The whole is the potential gradient term, and R0 is the reaction rate parameter;

[0109] Step S23: Electrode current density modeling, used to construct a single-cell scale model. Specifically, double-layer effect correction is introduced into the standard porous electromechanical model to perform electrode current density modeling, resulting in an improved electrode current density model. The calculation formula is as follows:

[0110] ;

[0111] In the formula, j is the current density value. It is a conductivity parameter. It is the electric potential within the electrode particles. The overall term is the potential gradient, where F is the Faraday constant and R is the gas constant. It is temperature, and C is ion concentration. It is the double-layer effect coefficient. It refers to the ion concentration of the electric double layer. It is the potential of the electric double layer;

[0112] Step S24: Battery equivalent circuit modeling, used to construct a system integration scale model. Specifically, in the standard equivalent circuit model, the resistance is treated as a dynamic variable and dynamic correction is performed to obtain the battery equivalent circuit model. The calculation formula is as follows:

[0113] ;

[0114] In the formula, V OC This is the open-circuit voltage value, I is the battery current value, and R is the open-circuit voltage value. eq V(t) is the equivalent resistance value, and V(t) is the battery operating voltage value.

[0115] It is a battery equivalent circuit modeling model, which is based on the equivalent resistance value R. eq Dynamic resistance correction is performed, and a system integration scale model is constructed. eq (t) is the equivalent resistance value considered as a dynamic variable, and R1 is the base resistance, used to represent the resistance value of the battery under normal operating conditions. This is the dynamic resistance change term, SOC(t) is the battery's state of charge, and SOH(t) is the battery's state of health. This refers to the battery temperature status;

[0116] Step S25: Four-level integration, used to construct a four-level resolution twin model, specifically by integrating the electronic structure modeling, the particle electrochemical reaction modeling, the electrode current density modeling, and the battery equivalent circuit modeling to obtain a battery management enhanced digital twin model.

[0117] By performing the above operations, this solution addresses the technical problem in existing dynamic digital twin modeling processes: traditional battery digital twin modeling rarely incorporates the internal and external conditions of the battery, making it difficult to fully reflect changes in battery state and health, thus hindering the modeling at both macroscopic and microscopic levels from reflecting the battery's dynamics. This solution creatively employs a four-level integrated digital twin model for dynamic digital twin modeling. Furthermore, it enhances the dynamics of the battery digital twin model by introducing multi-scale coupling in electronic structure calculations, non-uniform potential distribution correction in the phase-field model, and methods for calculating double-layer effects and dynamic equivalent resistance from the perspectives of single-cell and system integration. This results in an adaptive, dynamically adjustable complete digital twin system, facilitating subsequent battery health management, charge / discharge optimization, and fault prediction.

[0118] Example 4, see Figure 1 , Figure 2 and Figure 4 This embodiment is based on the above embodiment. In step S3, the dynamic charge-discharge enhancement is used to dynamically enhance the battery charge-discharge management. Specifically, based on the battery management system enhancement dataset and the battery management enhancement digital twin model, a reinforcement learning method combining dual-delay deep deterministic policy gradient and reward function update is used to perform dynamic charge-discharge enhancement to obtain a dynamic charge-discharge enhancement reference policy, including the following steps:

[0119] Step S31: Construct a dual-delay reinforcement learning framework. Specifically, construct a standard reinforcement learning environment parameter model, adopt a deep deterministic gradient strategy, combine the value function and policy function to set up the basic framework of reinforcement learning, and update the parameters of the value network and policy network by constructing a dual-delay update strategy.

[0120] Step S32: Dynamic charge and discharge enhancement reward improvement, specifically by designing a multi-dimensional reward function to comprehensively integrate the battery's lifespan, efficiency, and safety aspects, to improve the dynamic charge and discharge enhancement reward, thereby obtaining a dynamic charge and discharge enhancement reward function, and then performing reinforcement learning training based on the dynamic charge and discharge enhancement reward function;

[0121] The calculation formula for the multidimensional reward function is as follows:

[0122] ;

[0123] In the formula, Rw It is a multidimensional reward function, used as the reward function for the dual-delay reinforcement learning framework described in step S31. It is a weighted factor based on battery life. It is a battery life gain term. It is the weight of charge and discharge efficiency. It is the charge / discharge efficiency item. S is the safety factor weight, and S is the safety factor term;

[0124] The battery life gain term represents the change in battery health under the current charge / discharge operation, and its calculation formula is as follows:

[0125] ;

[0126] In the formula, This is the battery life gain term, and SOH(t) is the battery's state of health. opt (t) represents the battery health state under the optimal charge / discharge strategy;

[0127] The charge / discharge efficiency term represents the capacity utilization rate during the charge / discharge process, and its calculation formula is as follows:

[0128] ;

[0129] In the formula, It is the charge / discharge efficiency term, P out It is the power output when the battery discharges, P in This is the power input during battery charging;

[0130] The safety factor term represents the safety of the battery's current charge and discharge operation, and its calculation formula is as follows:

[0131] ;

[0132] In the formula, S is the safety factor term, and exp(·) is a function with the natural base. It is the temperature weighting of safety sensitivity, T current It is the temperature at the current moment, T safe It is the safe temperature of the battery. It is the safety sensitivity charging state weight, SOC current This is the current battery charging status, SOC. safe This is the safe charging state of the battery;

[0133] Step S33: Iterative update of dynamic charging and discharging strategy, specifically by applying the multidimensional reward function to the dual-delay reinforcement learning framework to iteratively update the dynamic charging and discharging strategy, and obtain iteratively updated dynamic charging and discharging strategy data;

[0134] Step S34: Dynamic charge-discharge enhancement, specifically, involves iterative training of the dynamic charge-discharge enhancement model through the dual-delay reinforcement learning framework, the dynamic charge-discharge enhancement reward improvement, and the dynamic charge-discharge strategy iteration update, to obtain the dynamic charge-discharge enhancement model. DH And by using the dynamic charge-discharge enhancement model Model DH Based on the battery management system enhancement dataset and the battery management enhancement digital twin model, dynamic charge and discharge enhancement is performed to obtain a dynamic charge and discharge enhancement reference strategy.

[0135] The dynamic charge-discharge enhancement reference strategy specifically includes battery health status data, charge-discharge control data, and optimized charge-discharge data.

[0136] By performing the above operations, this solution addresses the shortcomings of existing dynamic charge-discharge enhancement methods. These methods, when combined with dynamic digital modeling, lack sufficient intelligence and dynamism in battery charge-discharge adjustment strategies. Furthermore, this leads to the technical problem that changes in the battery's operating environment outpace the battery management adjustment strategy. This solution creatively combines a reinforcement learning method with dual-delay deep deterministic policy gradient and reward function updates for dynamic charge-discharge enhancement. Through a dual-delay update strategy combined with refined reward function improvements, the intelligence and dynamism of the battery charge-discharge adjustment strategy are enhanced from the bottom up. This not only helps extend battery life and improve charge-discharge efficiency but also ensures the battery operates within safe limits, thereby improving the overall performance and reliability of the battery system and battery management system.

[0137] Example 5, see Figure 1 , Figure 2 and Figure 5 This embodiment is based on the above embodiment. In step S4, the multi-dimensional safety situation awareness enhancement is used to enhance the battery safety situation awareness management in multiple dimensions. Specifically, based on the battery management system enhancement dataset and the battery management enhancement digital twin model, a spatiotemporal graph convolutional network combining the battery microstructure evolution map is used to enhance the multi-dimensional safety situation awareness and obtain multi-dimensional safety situation awareness reference data. This includes the following steps:

[0138] Step S41: Construct a spatiotemporal graph convolutional network to capture the spatiotemporal dependencies and basic features in the battery microstructure. Specifically, this involves constructing battery microstructure graph data and building a standard spatiotemporal graph convolutional network as the foundational network for enhancing multidimensional security situation awareness.

[0139] Step S42: Construct a battery microstructure evolution map to model the evolution process of the battery's internal structure and support dynamic prediction. Specifically, construct a basic model of battery microstructure state evolution and predict the evolution of battery micro-deformation, lithium plating, and solid electrolyte interface film growth based on the basic model of battery microstructure state evolution.

[0140] The calculation formula for the basic model of the battery microstructure state evolution is as follows:

[0141] ;

[0142] In the formula, S(t) represents the state of the battery's microstructure at the current time t, indicating the evolution of microscopic deformation, lithium plating, and solid electrolyte interfacial film growth. It is the change in the battery's microstructure over time t;

[0143] Step S43: Early fault propagation modeling, used to detect microstructure changes and implement early warning prediction of battery faults. Specifically, early fault propagation modeling is performed through the coupling effect of current, temperature and state of charge in the battery, and early warning prediction of battery faults is performed through the early fault propagation modeling.

[0144] The calculation formula for the early fault propagation model is as follows:

[0145] ;

[0146] In the formula, Fp(t) is the fault propagation rate parameter, w1 is the battery microstructure evolution weight, and m Li w1 represents the rate of evolution of the battery's microstructure, and w2 represents the weighting of temperature changes. w3 is the rate of temperature change, and w3 is the battery state-of-charge weight. It is the rate of change of the battery's state of charge;

[0147] Step S44: Training the multi-dimensional security situation awareness enhancement model. Specifically, this involves training the multi-dimensional security situation awareness enhancement model using the spatiotemporal graph convolutional basic network, the battery microstructure evolution map, and the early fault propagation modeling, to obtain the multi-dimensional security situation awareness enhancement model. EH ;

[0148] Step S45: Enhance multi-dimensional security situation awareness, specifically by using the aforementioned multi-dimensional security situation awareness enhancement model. EH Based on the battery management system enhanced dataset and the battery management enhanced digital twin model, multi-dimensional security situation awareness enhancement is performed to obtain multi-dimensional security situation awareness reference data.

[0149] The multi-dimensional safety situation awareness reference data specifically includes battery prediction parameters, fault propagation prediction rate, and battery safety status change prediction reference data.

[0150] By performing the above operations, this solution addresses the technical problem that existing multidimensional safety situation awareness enhancement methods rely heavily on traditional prediction methods, such as lifespan prediction based on data from the battery charging and discharging process. However, many traditional methods cannot effectively capture early signals of changes and faults at the microscopic level of the battery. These systems lack the ability to dynamically and intelligently perceive and analyze microscopic changes throughout the battery's entire life cycle. This solution creatively employs a spatiotemporal graph convolutional network that combines the battery's microstructure evolution map to enhance multidimensional safety situation awareness. Through dynamic monitoring and multidimensional comprehensive analysis, it provides more refined, comprehensive, and dynamic safety monitoring and management capabilities.

[0151] Example 6, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S5, the battery management system enhancement is used to comprehensively enhance the battery management system by combining battery charge and discharge management and safety situation management. Specifically, it constructs the basic mathematical model required for the battery management system enhancement through dynamic digital twin modeling, and comprehensively enhances the battery management system through the dynamic charge and discharge enhancement and the multi-dimensional safety situation awareness enhancement, thereby obtaining the dynamic charge and discharge enhancement reference strategy and the multi-dimensional safety situation awareness reference data. Finally, it obtains the comprehensive enhancement reference data of the battery management system by combining the dynamic charge and discharge enhancement reference strategy and the multi-dimensional safety situation awareness reference data.

[0152] Example 7, see Figure 1 and Figure 2 This embodiment is based on the above embodiments. The artificial intelligence-based battery management system enhancement system provided by the present invention includes a quantum sensing module, a digital twin module, a dynamic optimization module, a safety situation awareness module, and a battery management system enhancement module.

[0153] The quantum sensing module is used for heterogeneous data fusion acquisition. Through heterogeneous data fusion acquisition, an enhanced dataset of the battery management system is obtained, and the enhanced dataset of the battery management system is sent to the digital twin module, the dynamic optimization module, and the security situation awareness module.

[0154] The digital twin module is used for dynamic digital twin modeling. Through dynamic digital twin modeling, an enhanced digital twin model for battery management is obtained, and the enhanced digital twin model for battery management is applied to the dynamic optimization module and the security situation awareness module.

[0155] The dynamic optimization module is used for dynamic charge and discharge enhancement. Through dynamic charge and discharge enhancement, a dynamic charge and discharge enhancement reference strategy is obtained, and the dynamic charge and discharge enhancement reference strategy is sent to the battery management system enhancement module.

[0156] The security situation awareness module is used for multi-dimensional security situation awareness enhancement. Through multi-dimensional security situation awareness enhancement, multi-dimensional security situation awareness reference data is obtained, and the multi-dimensional security situation awareness reference data is sent to the battery management system enhancement module.

[0157] The battery management system enhancement module is used to enhance the battery management system and obtain comprehensive enhancement reference data for the battery management system through the enhancement.

[0158] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0159] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0160] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. An artificial intelligence-based method for enhancing a battery management system, characterized in that: The method includes the following steps: Step S1: Heterogeneous data fusion and acquisition to obtain the battery management system enhanced dataset; Step S2: Dynamic digital twin modeling. Based on the battery management system augmentation dataset, a four-level resolution twin model construction method is used to perform dynamic digital twin modeling to obtain the battery management augmentation digital twin model. The battery management system augmentation dataset includes a quantum chemical scale model, an electrode particle scale model, a single cell scale model, and a system integration scale model. The quantum chemical scale model is used to introduce a multi-scale coupling strategy for comprehensive modeling of electronic states and internal and external environmental parameters of the battery; the electrode particle scale model is used to introduce non-uniform electrical distribution correction to model the potential spatial variation within electrode particles and improve the traditional phase-field model; the single cell scale model is used to introduce the double-layer effect to improve the modeling of the influence of the double layer on the current distribution in the traditional motor model; the system integration scale model is used to introduce a dynamic resistance correction method based on state estimation to improve the dynamic adjustment modeling of the equivalent resistance model. Step S3: Dynamic charging and discharging enhancement. A reinforcement learning method combining dual-delay deep deterministic policy gradient and reward function update is used to perform dynamic charging and discharging enhancement to obtain a dynamic charging and discharging enhancement reference policy. By designing a multidimensional reward function, the reward function is updated. The calculation formula for the multidimensional reward function is as follows: ； In the formula, R w It is a multidimensional reward function, used as the reward function in a two-delay reinforcement learning framework. It is a weighted factor based on battery life. It is a battery life gain term. It is the weight of charge and discharge efficiency. It is the charge / discharge efficiency item. This refers to the safety factor weight, where S is the safety factor term; Step S4: Enhance multi-dimensional security situation awareness. A spatiotemporal graph convolutional network combining the battery microstructure evolution map is used to enhance multi-dimensional security situation awareness and obtain multi-dimensional security situation awareness reference data. Step S5: Battery management system enhancement, obtaining comprehensive enhancement reference data for the battery management system.

2. The method for enhancing a battery management system based on artificial intelligence according to claim 1, characterized in that: In step S1, the heterogeneous data fusion acquisition is used to collect raw data related to battery status required for battery management system analysis. Specifically, through heterogeneous data fusion acquisition, the raw dataset of battery status management is collected, and preliminary optimization and preprocessing are performed to obtain the enhanced dataset of battery management system. The heterogeneous data fusion acquisition specifically involves acquiring battery state data through the deployment of sensor arrays and dynamic feature acquisition of electrochemical impedance spectroscopy. The original dataset for battery state management includes battery voltage data, current data, temperature data, electrochemical data, mechanical support data, solid electrolyte interface film evolution data, and multiphysics fusion data. The preliminary optimization and preprocessing steps include data denoising, missing value handling, outlier correction, data standardization, time series data stabilization, and feature selection. By performing the preliminary optimization and preprocessing on the original battery state management dataset, an enhanced battery management system dataset is obtained. The battery management system enhanced dataset includes optimized time-series data, derived feature data, multi-physics fusion data, fusion index data, evolutionary feature data, labeled data, and synchronous multidimensional data.

3. The method for enhancing a battery management system based on artificial intelligence according to claim 2, characterized in that: In step S2, the dynamic digital twin modeling is used to construct a four-level digital twin model for battery state management. Specifically, based on the battery management system enhanced dataset, a four-level resolution twin model construction method is used to perform dynamic digital twin modeling to obtain a battery management enhanced digital twin model. The steps of using a four-level resolution twin model construction method to perform dynamic digital twin modeling and obtain an enhanced digital twin model for battery management include: Step S21: Electronic structure modeling, used to construct a quantum chemical scale model, specifically by introducing a multi-scale coupling strategy to calculate the electronic structure and obtain the reference value of the electronic structure Hamiltonian; Step S22: Particle electrochemical reaction modeling, used to construct an electrode particle-scale model, specifically by introducing a potential correction coefficient into the standard phase field model to model the particle electrochemical reaction and obtain reference values ​​for ion diffusion and reaction rate; Step S23: Electrode current density modeling, used to construct a single cell scale model, specifically by introducing double layer effect correction into the standard porous motor model, performing electrode current density modeling, and obtaining an improved electrode current density model. Step S24: Battery equivalent circuit modeling, used to construct a system integration scale model. Specifically, in the standard equivalent circuit model, the resistance is treated as a dynamic variable and dynamic correction is performed to obtain the battery equivalent circuit model. The calculation formula is as follows: ； In the formula, V OC This is the open-circuit voltage value, I is the battery current value, and R is the open-circuit voltage value. eq V(t) is the equivalent resistance value, and V(t) is the battery operating voltage value. It is a battery equivalent circuit modeling model, which is based on the equivalent resistance value R. eq Dynamic resistance correction is performed, and a system integration scale model is constructed. eq (t) is the equivalent resistance value considered as a dynamic variable, and R1 is the base resistance, used to represent the resistance value of the battery under normal operating conditions. This is the dynamic resistance change term, SOC(t) is the battery's state of charge, and SOH(t) is the battery's state of health. This refers to the battery temperature status; Step S25: Four-level integration, used to construct a four-level resolution twin model, specifically by integrating the electronic structure modeling, the particle electrochemical reaction modeling, the electrode current density modeling, and the battery equivalent circuit modeling to obtain a battery management enhanced digital twin model.

4. The method for enhancing a battery management system based on artificial intelligence according to claim 3, characterized in that: In step S3, the dynamic charge-discharge enhancement is used to dynamically enhance battery charge-discharge management. Specifically, based on the battery management system enhancement dataset and the battery management enhancement digital twin model, a reinforcement learning method combining dual-delay deep deterministic policy gradient and reward function updates is used to perform dynamic charge-discharge enhancement, resulting in a dynamic charge-discharge enhancement reference policy. This includes the following steps: Step S31: Construct a dual-delay reinforcement learning framework. Specifically, construct a standard reinforcement learning environment parameter model, adopt a deep deterministic gradient strategy, combine the value function and policy function to set up the basic framework of reinforcement learning, and update the parameters of the value network and policy network by constructing a dual-delay update strategy. Step S32: Dynamic charge and discharge enhancement reward improvement, specifically by designing a multi-dimensional reward function to comprehensively integrate the battery's lifespan, efficiency, and safety aspects, to improve the dynamic charge and discharge enhancement reward, thereby obtaining a dynamic charge and discharge enhancement reward function, and then performing reinforcement learning training based on the dynamic charge and discharge enhancement reward function; Step S33: Iterative update of dynamic charging and discharging strategy, specifically by applying the multidimensional reward function to the dual-delay reinforcement learning framework to iteratively update the dynamic charging and discharging strategy, and obtain iteratively updated dynamic charging and discharging strategy data; Step S34: Dynamic charge-discharge enhancement, specifically, involves iterative training of the dynamic charge-discharge enhancement model through the dual-delay reinforcement learning framework, the dynamic charge-discharge enhancement reward improvement, and the dynamic charge-discharge strategy iteration update, to obtain the dynamic charge-discharge enhancement model. DH And by using the dynamic charge-discharge enhancement model Model DH Based on the battery management system enhancement dataset and the battery management enhancement digital twin model, dynamic charge and discharge enhancement is performed to obtain a dynamic charge and discharge enhancement reference strategy. The dynamic charge-discharge enhancement reference strategy specifically includes battery health status data, charge-discharge control data, and optimized charge-discharge data.

5. The method for enhancing a battery management system based on artificial intelligence according to claim 4, characterized in that: In step S4, the multi-dimensional safety situation awareness enhancement is used to enhance the battery safety situation awareness management in multiple dimensions. Specifically, based on the battery management system enhanced dataset and the battery management enhanced digital twin model, a spatiotemporal graph convolutional network combining the battery microstructure evolution map is used to enhance the multi-dimensional safety situation awareness and obtain multi-dimensional safety situation awareness reference data. This includes the following steps: Step S41: Construct a spatiotemporal graph convolutional network to capture the spatiotemporal dependencies and basic features in the battery microstructure. Specifically, this involves constructing battery microstructure graph data and building a standard spatiotemporal graph convolutional network as the foundational network for enhancing multidimensional security situation awareness. Step S42: Construct a battery microstructure evolution map to model the evolution process of the battery's internal structure and support dynamic prediction. Specifically, construct a basic model of battery microstructure state evolution and predict the evolution of battery micro-deformation, lithium plating, and solid electrolyte interface film growth based on the basic model of battery microstructure state evolution. Step S43: Early fault propagation modeling, used to detect microstructure changes and implement early warning prediction of battery faults. Specifically, early fault propagation modeling is performed through the coupling effect of current, temperature and state of charge in the battery, and early warning prediction of battery faults is performed through the early fault propagation modeling. Step S44: Training the multi-dimensional security situation awareness enhancement model. Specifically, this involves training the multi-dimensional security situation awareness enhancement model using the spatiotemporal graph convolutional basic network, the battery microstructure evolution map, and the early fault propagation modeling, to obtain the multi-dimensional security situation awareness enhancement model. EH ; Step S45: Enhance multi-dimensional security situation awareness, specifically by using the aforementioned multi-dimensional security situation awareness enhancement model. EH Based on the battery management system enhanced dataset and the battery management enhanced digital twin model, multi-dimensional security situation awareness enhancement is performed to obtain multi-dimensional security situation awareness reference data.

6. The method for enhancing a battery management system based on artificial intelligence according to claim 5, characterized in that: In step S4, the multi-dimensional safety situation awareness reference data specifically includes battery prediction parameters, fault propagation prediction rate, and battery safety state change prediction reference data.

7. The method for enhancing a battery management system based on artificial intelligence according to claim 6, characterized in that: In step S5, the battery management system enhancement is used to comprehensively enhance the management of the battery management system by combining battery charge and discharge management and safety situation management. Specifically, it involves constructing the basic mathematical model required for the battery management system enhancement through dynamic digital twin modeling, and performing comprehensive management enhancement of the battery management system through the dynamic charge and discharge enhancement and the multi-dimensional safety situation awareness enhancement to obtain dynamic charge and discharge enhancement reference strategy and multi-dimensional safety situation awareness reference data. Finally, by combining the dynamic charge and discharge enhancement reference strategy and the multi-dimensional safety situation awareness reference data, comprehensive enhancement reference data for the battery management system is obtained.

8. An AI-based battery management system enhancement system, used to implement the AI-based battery management system enhancement method as described in any one of claims 1-7, characterized in that: It includes a quantum sensing module, a digital twin module, a dynamic optimization module, a security situation awareness module, and a battery management system enhancement module.

9. The AI-based battery management system enhancement system according to claim 8, characterized in that: The quantum sensing module is used for heterogeneous data fusion acquisition. Through heterogeneous data fusion acquisition, an enhanced dataset of the battery management system is obtained, and the enhanced dataset of the battery management system is sent to the digital twin module, the dynamic optimization module, and the security situation awareness module. The digital twin module is used for dynamic digital twin modeling. Through dynamic digital twin modeling, an enhanced digital twin model for battery management is obtained, and the enhanced digital twin model for battery management is applied to the dynamic optimization module and the security situation awareness module. The dynamic optimization module is used for dynamic charge and discharge enhancement. Through dynamic charge and discharge enhancement, a dynamic charge and discharge enhancement reference strategy is obtained, and the dynamic charge and discharge enhancement reference strategy is sent to the battery management system enhancement module. The security situation awareness module is used for multi-dimensional security situation awareness enhancement. Through multi-dimensional security situation awareness enhancement, multi-dimensional security situation awareness reference data is obtained, and the multi-dimensional security situation awareness reference data is sent to the battery management system enhancement module. The battery management system enhancement module is used to enhance the battery management system and obtain comprehensive enhancement reference data for the battery management system through the enhancement.

Images