Battery health management system and method

Abstract

The invention discloses a battery health management system and method, and belongs to the technical field of battery management, the system comprises an implantable sensor network, an edge and cloud collaborative computing platform, a multi-modal prediction model and a dynamic optimization execution unit; the implantable sensor network integrates various sensors to collect multi-physical field parameters in a battery, and the energy collection unit supplies power through cross-shielding communication transmission. The edge and cloud collaboration platform processes data, trains a model and carries out block chain evidence storage; the multi-modal prediction model is fused with space-time double-flow Transform and causal reasoning, SOH and RUL are predicted, and thermal runaway is early warned; the dynamic optimization execution unit realizes efficient control through layered equalization and self-adaptive thermal management; the method comprises an initialization stage, an operation stage and a maintenance stage to form a closed loop. The method breaks through the traditional limitation, improves the safety, reliability and economy of the battery, and is suitable for electric vehicles, energy storage power stations and other scenes.

Description

Technical Field

[0001] This invention relates to the field of battery management, and more particularly to a system and method for battery health management. Background Technology

[0002] With the rapid development of the new energy industry, lithium-ion batteries, as the core component of energy storage and conversion, have been widely used in electric vehicles, energy storage power stations and other fields. The accurate assessment and safety protection of battery health status (SOH) and remaining life (RUL) directly affect the operational reliability, economy and safety of the equipment. Therefore, battery health management technology has become a core research topic in the industry.

[0003] Currently, while traditional battery management systems play a role in monitoring and protecting basic battery parameters, their ability to manage complex operating conditions remains significantly insufficient.

[0004] At the state perception level, existing systems mainly rely on macroscopic parameters such as external voltage, current, and surface temperature to indirectly infer the battery health status. They cannot directly obtain the physicochemical changes inside the cell. For example, micro-states such as the initiation of microcracks in electrode materials, the decay of electrolyte ion conductivity, and the decomposition of SEI film are difficult to capture. As a result, health assessment can only be based on empirical models, and the accuracy is limited by the mapping error between external parameters and internal states, which cannot provide essential basis for health management.

[0005] In terms of data processing and collaborative architecture, existing technologies mostly adopt single edge computing or centralized cloud computing modes: edge computing is limited by hardware computing power and it is difficult to train complex models; cloud computing requires uploading a large amount of raw data, which poses a risk of data privacy leakage and lacks an effective distributed data sharing mechanism. At the same time, traditional systems lack the ability to dynamically map the state of the battery throughout its entire life cycle, the real-time synchronization between digital models and physical batteries is poor, and the storage of key health data depends on centralized servers, which are easily tampered with or lost, and cannot meet the traceability requirements.

[0006] In terms of health prediction and early warning, existing prediction models are mostly based on single-dimensional data or simple machine learning algorithms, which are difficult to integrate multi-source heterogeneous information. For example, SOH prediction often uses empirical formulas and ignores the coupling effect of dynamic factors such as temperature and charge / discharge rate. Thermal runaway early warning only relies on surface temperature threshold and lacks causal analysis of early characteristics such as pressure and gas generation, resulting in an early warning time of less than 10 minutes, which is difficult to meet the safety protection requirements.

[0007] At the level of optimized control, traditional active balancing systems mostly adopt fixed-path energy transfer, without considering the dynamic SOC difference between battery packs, resulting in balancing efficiency of less than 90%. Thermal management strategies are mostly based on simple PID control, which can only maintain single-point temperature stability and cannot achieve precise control of the temperature difference across the entire battery pack. Furthermore, the balancing and thermal management strategies are independent of each other and do not form a synergistic optimization, resulting in limited improvement in battery cycle life.

[0008] Therefore, this application proposes a novel system and method for battery health management. Summary of the Invention

[0009] One objective of this invention is to propose a system and method for battery health management. This invention can overcome the limitations of traditional battery health management technologies in terms of state perception, data processing, prediction and early warning, and optimized control. By embedding multi-physics parameter acquisition, edge and cloud collaborative computing, multimodal intelligent prediction, and dynamic collaborative optimization, it can achieve accurate assessment of battery health status, early safety warning, and efficient management throughout the entire life cycle, significantly improving the safety, reliability, and economy of battery use.

[0010] A battery health management system according to an embodiment of the present invention includes an implantable sensor network, an edge and cloud collaborative computing platform, a multimodal prediction model, and a dynamic optimization execution unit;

[0011] The implanted sensor network is used to collect multi-physics parameters inside the battery, including temperature, pressure, electrode deformation, and electrolyte ion conductivity.

[0012] The edge and cloud collaborative computing platform is communicatively connected to the implanted sensor network and is used to process the collected multi-physics parameters and construct a battery health status model.

[0013] The multimodal prediction model is deployed on the edge and cloud collaborative computing platform and is used to predict battery health indicators based on processed multiphysics parameters. The health indicators include State of Health (SOH) and Remaining Life (RUL).

[0014] The dynamic optimization execution unit is communicatively connected to the edge and cloud collaborative computing platform and is used to perform battery equalization control and thermal management adjustment based on predicted health indicators, forming a health management closed loop.

[0015] Furthermore, the implantable sensor network includes a miniaturized sensor array, a cross-shield communication module, and an energy harvesting unit;

[0016] A miniaturized sensor array, comprising a temperature and pressure integrated sensor, a distributed fiber Bragg grating strain sensor, and a micro / nano fluidic chip-type ion conductivity sensor, wherein the sensor uses a ceramic and polyimide composite encapsulation material and forms an electrolyte corrosion resistant coating through an atomic layer deposition process, wherein the coating has a temperature resistance range of -40℃ to 150℃ and a voltage resistance of ≥500V.

[0017] The cross-shielded communication module uses 2.4GHz carrier modulation technology and LoRa protocol to transmit data through the battery's metal casing, with a transmission delay of ≤10ms;

[0018] The energy harvesting unit generates an induced current to power the sensor by means of battery charging and discharging voltage fluctuations, and its energy conversion efficiency is ≥65%.

[0019] Furthermore, the center wavelength offset of the distributed fiber Bragg grating strain sensor satisfies the formula:

[0020] ;

[0021] in, The initial wavelength of the grating. The elastic coefficient is 1. The strain value of the electrode material. Thermo-optic coefficient, As the temperature change, the strain and temperature cross-sensitivity are separated by a temperature compensation algorithm, and the deformation measurement accuracy reaches ±2με.

[0022] Furthermore, the impedance spectrum of the micro / nanofluidic chip-type ion conductivity sensor satisfies the formula:

[0023] ;

[0024] in, For sensor impedance, For the resistance of the solution, It is a double-layer capacitor. For charge transfer resistor, For alternating current frequency, The imaginary unit represents ionic conductivity. pass Calculations are performed, where L is the flow channel length, A is the cross-sectional area, and the measurement error is <3%.

[0025] Furthermore, the edge and cloud collaborative computing platform includes edge nodes and a cloud-based intelligent hub;

[0026] The edge nodes are deployed with an anomaly detection model based on MobileNetv3 and federated learning nodes. The anomaly detection model filters invalid data in real time with a recognition rate of ≥99.2%. The federated learning nodes share desensitized aging features with adjacent battery packs and achieve differential privacy protection by adding Laplacian noise, with a model accuracy loss of ≤2%.

[0027] The cloud-based intelligent hub includes a digital twin engine and a blockchain evidence storage module. The blockchain evidence storage module uses the PBFT consensus mechanism to store health indicators and has a fault-tolerant node count. , where n is the total number of nodes.

[0028] Furthermore, the multimodal prediction model includes a spatiotemporal dual-stream Transformer module and a causal inference module, wherein the spatial flow characteristics of the spatiotemporal dual-stream Transformer module satisfy:

[0029] ;

[0030] in, Let i be the spatial flow feature vector of the i-th sensor node. For multilayer perceptrons, For feature splicing operations, Let i be the spatial coordinates of the i-th sensor node. Location encoding, Let be the strain characteristic of the i-th node. The temperature feature of the i-th node is... The pressure characteristic of the i-th node;

[0031] Time-flow self-attention computation satisfies:

[0032] ;

[0033] in, To output the feature matrix for self-attention, For querying the matrix, The key matrix, For value matrices, , , , To query the matrix weight parameters, Let be the temporal characteristic matrix at time t. These are the key matrix weight parameters. for The temporal feature matrix at time step 1. The weight parameters of the value matrix, For adjustable time steps, for and Dimensions This is the normalization function;

[0034] The causal reasoning module analyzes the causal relationship of multiphysics parameters based on the Do-Calculus algorithm, with a thermal runaway early warning time of ≥30 minutes and an accuracy of ≥98%.

[0035] Furthermore, the state of health (SOH) is calculated using the following formula:

[0036] ;

[0037] in, , , The aging factor is... This is the charging and discharging current. For cell temperature, The rate of change of voltage. For the initial capacity, For the integral time infinitesimal;

[0038] The remaining lifespan RUL is determined by the formula. calculate, This is the retirement threshold.

[0039] Furthermore, the hierarchical active balancing system of the dynamic optimization execution unit includes bottom-level cell balancing and top-level inter-group balancing. The bottom level uses a Buck-Boost circuit to achieve energy transfer, resulting in high balancing efficiency. in ;

[0040] in, To balance the efficiency of battery cells. The output voltage of the circuit. For output current, To balance time, The circuit input voltage. For the input current, and ;

[0041] The top layer achieves inter-group energy dispatch through a single-inductor circuit, based on an objective function. Optimize the path, with the following constraints: The solution is obtained using an improved genetic algorithm.

[0042] in, For inter-group energy transfer current, For the resistance of the energy transfer path, For scheduling time, For the summation operation, This is an operation to find the minimum value.

[0043] Furthermore, the adaptive thermal management module of the dynamic optimization execution unit adopts model predictive control, and the temperature prediction model satisfies the formula:

[0044] ;

[0045] in The state matrix, For the control matrix, Here is the perturbation matrix. Let t be the actual temperature of the battery. Liquid cooling flow rate / heating power For disturbance terms;

[0046] The control objective function satisfies The constraints are By solving the problem using a sequential quadratic programming algorithm, and in conjunction with a paraffin and graphene composite phase change material, the battery temperature difference can be controlled within ±1.5℃.

[0047] in, To control the objective function value, n is the prediction time domain length. To predict the step size, This is the temperature deviation weighting coefficient. To control the weighting coefficients, The target temperature.

[0048] A method for battery health management includes the following steps:

[0049] S1. Initialization phase: Obtain the initial impedance spectrum of the battery through pulse testing, construct an individual feature library, synchronously generate a digital twin virtual battery and complete blockchain notarization;

[0050] S2. During the operation phase, the edge nodes collect multiphysics parameters every 100ms, which are then uploaded to the cloud after noise reduction. The cloud updates the health prediction model hourly based on the SOH formula and sends the optimized parameters for solving the objective function to the edge nodes.

[0051] S3. During the maintenance phase, when the predicted remaining lifespan (RUL) is ≤ 50 cycles, a personalized lifespan extension plan is generated. After battery retirement, the remaining value is assessed based on blockchain data to guide secondary utilization. The remaining lifespan (RUL) is determined through... calculate, This is the retirement threshold.

[0052] The beneficial effects of this invention are:

[0053] 1. In this invention, multiple physical field parameters such as internal battery temperature, pressure, electrode deformation, and electrolyte ion conductivity are directly collected through an implanted sensor network. Combined with a specific formula to quantify the sensing principle, the microscopic state of the battery is accurately captured, providing a more fundamental basis for health assessment. Existing technologies cannot achieve such comprehensive and accurate direct collection and quantitative analysis of multiple physical field parameters inside the battery.

[0054] 2. This invention adopts an edge and cloud collaborative computing architecture. The federated learning mechanism of edge nodes enables the sharing of aging features while protecting data privacy. The digital twin engine in the cloud is combined with blockchain notarization to dynamically map battery status and ensure that data is tamper-proof. This collaborative mode and data processing and notarization method are not available in existing technologies, which improves data utilization efficiency and security.

[0055] 3. The multimodal prediction model in this invention integrates the spatiotemporal dual-stream Transformer and the causal inference algorithm. It constructs the health indicator calculation and prediction logic through a specific formula. It can not only accurately predict SOH and RUL, but also provide early warning of thermal runaway more than 30 minutes in advance. The prediction accuracy and early warning timeliness far exceed those of traditional empirical models. Existing technologies lack such a multi-dimensional fusion and accurate quantitative prediction mechanism.

[0056] 4. The hierarchical active balancing system of the dynamic optimization execution unit in this invention adopts an improved genetic algorithm to optimize the energy path, combined with the MPC control strategy of adaptive thermal management, and achieves efficient balancing and precise temperature control through quantitative formulas, which significantly extends the battery cycle life by more than 25%. Existing technologies cannot achieve such efficient and precise dynamic optimization control. Attached Figure Description

[0057] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0058] Figure 1 This is a schematic diagram of the overall system framework of the battery health management system and method proposed in this invention;

[0059] Figure 2 This is a schematic diagram of the process flow of a battery health management system and method proposed in this invention. Detailed Implementation

[0060] To make the technical means and objectives and effects of the present invention easier to understand, the embodiments of the present invention will be described in detail below with reference to specific illustrations.

[0061] Example 1

[0062] like Figure 1As shown, this embodiment discloses a battery health management system that achieves precise health management throughout the entire battery lifecycle through deep integration of embedded sensing, edge and cloud collaborative computing, multimodal prediction, and dynamic optimization execution.

[0063] The system specifically includes an implanted sensor network, an edge and cloud collaborative computing platform, a multimodal prediction model, and a dynamic optimization execution unit. The structure and working principle of each part are as follows:

[0064] Implantable sensor networks are integrated inside battery cells to collect multi-physics parameters during battery operation in real time. Their core components include a miniaturized sensor array, a cross-shielded communication module, and an energy harvesting unit. The implantable sensor network acts as the system's "sensory nerves," and the arrangement of its miniaturized sensor array is as follows:

[0065] The integrated temperature and pressure sensor is surface-mounted near the battery cell tabs, where temperature and pressure changes best reflect the reaction intensity of the active materials in the battery cell. The sensor's sensing element is a silicon-doped piezoresistor. The temperature measurement range is -40℃ to 150℃, and the pressure measurement range is 0 to 500 kPa. The internal ADC module converts the analog signal into a 16-bit digital signal, ensuring a temperature accuracy of ±0.2℃ and a pressure accuracy of ±0.1 kPa.

[0066] The distributed fiber Bragg grating strain sensor is embedded in a serpentine path between the electrode plates and the current collector, using a fiber grating with a grating length of 10 mm and an initial wavelength of... =1550nm, elastic coefficient =0.22, thermo-optical coefficient The wavelength shift is monitored in real time using an optical demodulator. Then according to the formula Calculate electrode deformation To eliminate the interference of temperature on deformation measurement, another strain-free grating on the same optical fiber is used as a temperature reference. Temperature compensation is achieved through differential calculation, ultimately enabling the deformation measurement accuracy to reach ±2. It can capture the minute strain caused by the volume expansion of electrode materials during charge-discharge cycles.

[0067] The micro / nanofluidic chip-type ion conductivity sensor is integrated near the electrolyte filling port of the battery cell. The chip is fabricated using PDMS material through soft photolithography, and the flow channel length is... =5mm, cross-sectional area =0.1mm², platinum electrodes are placed at both ends of the flow channel, and an AC signal of 10Hz-1MHz is applied. The impedance spectrum is measured by an impedance analyzer. The solution resistance is: The ionic conductivity can be calculated by obtaining the real intercept of the impedance spectrum in the high-frequency range. The sensor has a measurement error of less than 3% and can effectively reflect changes in the migration ability of lithium ions in the electrolyte. For example, when the SEI film breaks down and the electrolyte decomposes, the ion conductivity will decrease by more than 5% within 1 hour.

[0068] The cross-shielded communication module adopts a system-on-a-chip design and operates in the 2.4GHz ISM band. To penetrate the shielding of the battery's metal casing, it uses frequency hopping spread spectrum technology with a channel switching rate of 100 hops / second, a transmit power controlled within 10dBm, and a receive sensitivity of -148dBm. By attaching a flexible FPCB antenna to the inside of the battery casing, it achieves wireless transmission of sensor data. Actual testing showed that even with a battery pack stack thickness of 30cm, the transmission delay could still be controlled within ≤10ms, and the packet loss rate was <0.1%.

[0069] The energy harvesting unit consists of a miniature coil, a rectifier bridge, and a supercapacitor. The coil is wound around the battery cell casing. When the battery charges and discharges, the current changes, and the coil induces an alternating voltage. After being converted into a DC voltage by the rectifier bridge, it powers the 3.3V sensor. The energy conversion efficiency is ≥65%. At a 1C charge / discharge rate, it can output a stable current of 5mA. With the help of a 1mF supercapacitor, it ensures that the sensor can continue to work for more than 4 hours when the battery is idle.

[0070] The edge and cloud collaborative computing platform undertakes data processing and model building functions, and is divided into two parts: edge nodes and cloud intelligent hub.

[0071] Edge nodes are deployed on an anomaly detection model based on MobileNetv3. The model input is time-series data of temperature, pressure, deformation, and ionic conductivity collected by sensors. Features are extracted through depthwise separable convolutional layers, and anomaly probabilities are output through a Sigmoid activation function. Data with a probability > 0.5 is considered invalid. After training on 100,000 sets of labeled data, the invalid data identification rate reached 99.2%.

[0072] Simultaneously, the edge nodes integrate federated learning nodes, establish encrypted communication with the edge nodes of adjacent battery packs, and share the de-identified aging characteristics. The workflow of the federated learning nodes is as follows:

[0073] Edge nodes standardize local aging features, encrypt them using a homomorphic encryption algorithm, and exchange parameters with the edge nodes of the 10 adjacent battery packs. The aggregation server updates model parameters using a federated averaging algorithm. To protect data privacy, Laplacian noise is added before feature upload, and the noise scale is dynamically adjusted according to feature sensitivity. Tests show that the model accuracy loss is ≤2%.

[0074] The cloud-based intelligent hub adopts a GPU cluster architecture, and its digital twin engine is based on COMSOL Multiphysics to build a three-dimensional electrochemical virtual model of the battery.

[0075] Specifically, the construction process of a digital twin engine is as follows:

[0076] First, the three-dimensional structure of the battery is obtained through CT scanning. An electrochemical model containing the positive electrode, negative electrode, separator, and electrolyte is established in COMSOL. The initial values ​​of key parameters in the model are obtained from literature. Then, multiphysics parameters uploaded by edge nodes are received every hour. The model parameters are dynamically calibrated using a particle swarm optimization algorithm to ensure that the error between the simulated voltage and temperature values ​​of the virtual battery and the measured values ​​of the physical battery is less than 2%, thus achieving dynamic mapping.

[0077] The blockchain evidence storage module is built on the Hyperledger Fabric framework, comprising 3 sorting nodes and 10 peer nodes. It employs the PBFT consensus mechanism, with a block generation interval of 10 seconds. Each block contains 50 health indicator records, and the data is stored after being hashed using SHA-256 to ensure immutability. Its fault tolerance meets [specific requirements]. ,in When =10, =3, which can resist malicious attacks from 3 nodes.

[0078] The multimodal prediction model is deployed on an edge and cloud collaborative computing platform to predict health status by fusing multi-physics data.

[0079] The model includes a spatiotemporal dual-stream Transformer module and a causal inference module:

[0080] The specific structure of the spatiotemporal dual-stream Transformer module is as follows: the spatial stream contains four convolutional attention modules. Each module first performs a 3×3 convolution on the 32×32 sensor spatial distribution matrix, and then calculates the correlation weights between nodes through a self-attention mechanism, outputting:

[0081] ;

[0082] in Position encoding for sensor coordinates, These represent the strain, temperature, and pressure characteristics of the i-th node, respectively.

[0083] The time-series data employs a 6-layer Transformer encoder, with input consisting of time-series data from the past hour. After positional encoding, the data is processed through a self-attention mechanism. Processing time-series data, where , Let be the characteristic matrix at time t. With an adjustable time step of 5-30 seconds, it captures the long-term trend of battery aging.

[0084] The causal reasoning module, based on the Do-Calculus algorithm, constructs a causal graph containing 12 nodes and 30 causal edges. It identifies key triggers through intervention analysis. When a causal chain of "temperature > 60℃, pressure increase > 5 kPa / min, and ion conductivity decrease > 10%" is detected, a thermal runaway warning is triggered. After 100 simulated thermal runaway experiments, the warning time is ≥ 30 minutes and the accuracy is ≥ 98%.

[0085] In the calculation of State of Health (SOH), the aging coefficient is calibrated through accelerated aging experiments: 100 batteries of the same specifications are cycled to SOH=0.8 at different temperatures and different rates, and the results under each condition are recorded. , , Data, specifically using formulas Calculation, where The aging coefficient is determined experimentally. This is the charging and discharging current. For cell temperature, The rate of change of voltage. This is the initial capacity;

[0086] The prediction process for Remaining Lifetime (RUL) is as follows: First, the State of Health (SOH) curve for the next 100 cycles is predicted using an LSTM network, and then based on... calculate, The retirement threshold is the number of cycles when the SOH first drops to the threshold after predicting the next n cycles. The retirement threshold can be set to 0.8.

[0087] The dynamic optimization execution unit performs active balancing control and adaptive thermal management adjustment based on the output of the multimodal prediction model.

[0088] The hierarchical active balancing system is divided into bottom-level cell balancing and top-level inter-group balancing:

[0089] The underlying layer uses a Buck-Boost circuit to achieve energy transfer between battery cells, and its balancing efficiency is high. in D is an adjustable duty cycle of 0-1, which is adjusted in real time through adaptive PI control to ensure that the cell voltage difference is ≤5mV;

[0090] The top layer achieves energy dispatch among battery packs through a single-inductor circuit, based on an objective function. Optimize the energy transfer path, with constraints for any battery pack i and j. An improved genetic algorithm is used to solve the problem.

[0091] The crossover probability of the algorithm ,in The variance of population fitness and the probability of mutation. ,in For individual fitness, To achieve the maximum fitness of the population, a mixture fitness function is used. It also optimizes energy consumption and balancing speed, and retains the top 5% of elite individuals in each generation to directly enter the next generation, improving the convergence speed by 40% compared to the standard genetic algorithm.

[0092] The adaptive thermal management module employs model predictive control, and its temperature prediction model satisfies... ,in The state matrix, For the control matrix, Liquid cooling flow rate / heating power For ambient temperature / charge / discharge power disturbances, the control objective function is: ,in, Here are the weighting coefficients, and the constraints are: The optimal control quantity is solved by using a sequential quadratic programming algorithm, and combined with the paraffin and graphene composite phase change material filled in the gaps between the batteries, the temperature difference of the battery pack is controlled within ±1.5℃.

[0093] Example 2

[0094] like Figure 2 As shown, this embodiment discloses a battery health management method based on the system of Embodiment 1, including an initialization phase, an operation phase, and a maintenance phase. The specific process is as follows:

[0095] S1. Initialization phase: When the battery is first activated, an AC pulse signal of 5Hz-1kHz is applied to the battery through a pulse test system to collect impedance responses at different frequencies and build an individual feature library containing parameters such as initial impedance spectrum, capacity, and internal resistance.

[0096] Meanwhile, the cloud-based intelligent hub generates a digital twin virtual battery based on this feature library. The geometric parameters of its three-dimensional model, such as electrode thickness and membrane pore size, are completely consistent with the physical battery. It also uploads initial parameters, such as production batch and activation date, to the blockchain evidence storage module to complete the first write to the distributed ledger.

[0097] S2. During the operation phase, the implanted sensor network continuously collects multiple physical field parameters such as temperature, pressure, electrode deformation, and electrolyte ion conductivity at 100ms intervals, and transmits them to the edge nodes via the cross-shielded communication module.

[0098] Edge nodes first filter invalid data using the MobileNetv3 anomaly detection model, then denoise the valid data, extract features, and upload it to the cloud-based intelligent hub. Every hour, the cloud calls the spatiotemporal dual-stream Transformer module, combining the real-time status of the digital twin virtual battery to update the SOH and RUL prediction results. It then uses an improved genetic algorithm and MPC algorithm to solve for the optimal equilibrium strategy and thermal management parameters, which are then distributed to the edge nodes.

[0099] Edge nodes drive dynamic optimization of execution units based on the parameters sent down: cell balancing is achieved by adjusting the duty cycle of the Buck-Boost circuit, and temperature is adjusted by controlling the opening of the liquid cooling valve and the PTC heating power, forming a closed-loop control of "sensing-analysis-decision-execution".

[0100] S3. During the maintenance phase, when the system predicts that the RUL (Range Limit) is ≤ 50 cycles, a personalized life extension plan is automatically generated, including limiting the depth of charge and discharge and reducing the maximum charge and discharge rate, and maintenance reminders are pushed to the user terminal.

[0101] When the battery's state of harm (SOH) drops to the retirement threshold of 0.8, the system assesses its remaining value based on blockchain-stored full lifecycle data, such as cycle count, maximum temperature, and fault records.

[0102] For batteries with a SOH of 0.6-0.8, it is recommended to use them in energy storage power stations and other cascade utilization scenarios.

[0103] For batteries with SOH < 0.6, they should be directed to professional recycling organizations for material regeneration.

[0104] Meanwhile, the historical data stored on the blockchain can support traceability analysis, providing a basis for battery design optimization.

[0105] The above-mentioned system and method can achieve accurate perception, prediction and optimized control of battery health status, which can extend battery cycle life by more than 25% compared with traditional BMS system, improve the accuracy of thermal runaway early warning to more than 98%, and significantly improve the safety and economy of battery use.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A battery health management system, characterized in that, This includes implanted sensor networks, edge and cloud collaborative computing platforms, multimodal prediction models, and dynamically optimized execution units; The implanted sensor network is used to collect multi-physics parameters inside the battery, including temperature, pressure, electrode deformation, and electrolyte ion conductivity. The edge and cloud collaborative computing platform is communicatively connected to the implanted sensor network and is used to process the collected multi-physics parameters and construct a battery health status model. The multimodal prediction model is deployed on the edge and cloud collaborative computing platform and is used to predict battery health indicators based on processed multiphysics parameters. The health indicators include State of Health (SOH) and Remaining Life (RUL). The dynamic optimization execution unit is communicatively connected to the edge and cloud collaborative computing platform and is used to perform battery equalization control and thermal management adjustment based on predicted health indicators, forming a health management closed loop.

2. The battery health management system according to claim 1, characterized in that, The implantable sensor network includes a miniaturized sensor array, a cross-shielded communication module, and an energy harvesting unit. A miniaturized sensor array, comprising a temperature and pressure integrated sensor, a distributed fiber Bragg grating strain sensor, and a micro / nano fluidic chip-type ion conductivity sensor, wherein the sensor uses a ceramic and polyimide composite encapsulation material and forms an electrolyte corrosion resistant coating through an atomic layer deposition process, wherein the coating has a temperature resistance range of -40℃ to 150℃ and a voltage resistance of ≥500V. The cross-shielded communication module uses 2.4GHz carrier modulation technology and LoRa protocol to transmit data through the battery's metal casing, with a transmission delay of ≤10ms; The energy harvesting unit generates an induced current to power the sensor by means of battery charging and discharging voltage fluctuations, and its energy conversion efficiency is ≥65%.

3. The battery health management system according to claim 2, characterized in that, The center wavelength offset of the distributed fiber Bragg grating strain sensor satisfies the formula: ； in, The initial wavelength of the grating. The elastic coefficient is 1. The strain value of the electrode material. Thermo-optic coefficient, As the temperature change, the strain and temperature cross-sensitivity are separated by a temperature compensation algorithm, and the deformation measurement accuracy reaches ±2με.

4. The battery health management system according to claim 2, characterized in that, The impedance spectrum of the micro / nano fluidic chip-type ion conductivity sensor satisfies the following formula: ； in, For sensor impedance, For the resistance of the solution, It is a double-layer capacitor. For charge transfer resistor, For alternating current frequency, The imaginary unit represents ionic conductivity. pass Calculations are performed, where L is the flow channel length, A is the cross-sectional area, and the measurement error is <3%.

5. The battery health management system according to claim 1, characterized in that, The edge and cloud collaborative computing platform includes edge nodes and a cloud-based intelligent hub. The edge nodes are deployed with an anomaly detection model based on MobileNetv3 and federated learning nodes. The anomaly detection model filters invalid data in real time with a recognition rate of ≥99.2%. The federated learning nodes share desensitized aging features with adjacent battery packs and achieve differential privacy protection by adding Laplacian noise, with a model accuracy loss of ≤2%. The cloud-based intelligent hub includes a digital twin engine and a blockchain evidence storage module. The blockchain evidence storage module uses the PBFT consensus mechanism to store health indicators and has a fault-tolerant node count. , where n is the total number of nodes.

6. The battery health management system according to claim 1, characterized in that, The multimodal prediction model includes a spatiotemporal dual-stream Transformer module and a causal inference module. The spatial flow characteristics of the spatiotemporal dual-stream Transformer module satisfy the following: ； in, Let i be the spatial flow feature vector of the i-th sensor node. For multilayer perceptrons, For feature splicing operations, Let i be the spatial coordinates of the i-th sensor node. Location encoding, Let be the strain characteristic of the i-th node. The temperature feature of the i-th node is... The pressure characteristic of the i-th node; Time-flow self-attention computation satisfies: ； in, To output the feature matrix for self-attention, For querying the matrix, The key matrix, For value matrices, , , , To query the matrix weight parameters, Let be the temporal characteristic matrix at time t. These are the key matrix weight parameters. for The temporal feature matrix at time step 1. The weight parameters of the value matrix, For adjustable time steps, for and Dimensions This is the normalization function; The causal reasoning module analyzes the causal relationship of multiphysics parameters based on the Do-Calculus algorithm, with a thermal runaway early warning time of ≥30 minutes and an accuracy of ≥98%.

7. The battery health management system according to claim 1, characterized in that, The state of health (SOH) is calculated using the following formula: ； in, , , The aging factor is... This is the charging and discharging current. For cell temperature, The rate of change of voltage. For the initial capacity, For the integral time infinitesimal; The remaining lifespan RUL is determined by the formula. calculate, This is the retirement threshold.

8. The battery health management system according to claim 1, characterized in that, The hierarchical active balancing system of the dynamic optimization execution unit includes bottom-level cell balancing and top-level inter-group balancing. The bottom level uses a Buck-Boost circuit to achieve energy transfer, resulting in high balancing efficiency. in ; in, To balance the efficiency of battery cells. The output voltage of the circuit. For output current, To balance time, The circuit input voltage. For the input current, and ; The top layer achieves inter-group energy dispatch through a single-inductor circuit, based on an objective function. Optimize the path, with the following constraints: The solution is obtained using an improved genetic algorithm. in, For inter-group energy transfer current, For the resistance of the energy transfer path, For scheduling time, For the summation operation, This is an operation to find the minimum value.

9. A battery health management system according to claim 8, characterized in that, The adaptive thermal management module of the dynamic optimization execution unit adopts model predictive control, and the temperature prediction model satisfies the formula: ； in The state matrix, For the control matrix, Here is the perturbation matrix. Let t be the actual temperature of the battery. Liquid cooling flow rate / heating power For disturbance terms; The control objective function satisfies The constraints are By solving the problem using a sequential quadratic programming algorithm, and in conjunction with a paraffin and graphene composite phase change material, the battery temperature difference can be controlled within ±1.5℃. in, To control the objective function value, n is the prediction time domain length. To predict the step size, This is the temperature deviation weighting coefficient. To control the weighting coefficients, The target temperature.

10. A method for battery health management, characterized in that, The battery health management system according to any one of claims 1-9 includes the following steps: S1. Initialization phase: Obtain the initial impedance spectrum of the battery through pulse testing, construct an individual feature library, synchronously generate a digital twin virtual battery and complete blockchain notarization; S2. During the operation phase, the edge nodes collect multiphysics parameters every 100ms, which are then uploaded to the cloud after noise reduction. The cloud updates the health prediction model hourly based on the SOH formula and sends the optimized parameters for solving the objective function to the edge nodes. S3. During the maintenance phase, when the predicted remaining lifespan (RUL) is ≤ 50 cycles, a personalized lifespan extension plan is generated. After battery retirement, the remaining value is assessed based on blockchain data to guide secondary utilization. The remaining lifespan (RUL) is determined through... calculate, This is the retirement threshold.

Images