Power battery full-period management method and system

Abstract

The invention provides a power battery full-period management method and system, and the method comprises the steps: enabling an edge end to obtain the real-time temperature of a plurality of regions on each pole piece in each single battery in an actual battery module, obtaining the temperature data of each pole piece, and obtaining the operation data of the actual battery module; the edge end preprocesses the temperature data and the operation data to obtain the preprocessed temperature data and the preprocessed operation data; the edge end sends the preliminary analysis data, the preprocessed temperature data and the preprocessed operation data to the cloud end; performing dynamic embedded simulation by the digital twin model, and outputting dynamic operation data; a parameter prediction result of the actual battery module is obtained through prediction of the data prediction model according to the dynamic operation data and / or the preprocessed operation data; and determining a processing strategy of the actual battery module by the decision model according to the preliminary analysis data and / or the parameter prediction result. And effective management of the whole life cycle of the battery module is realized.

Description

Technical Field

[0001] The present application relates to the technical field of management of power battery modules of automobiles, and more specifically, to a method and system for full-cycle management of power batteries. Background Art

[0002] Power battery modules are key components of electric vehicles. Overcharging or over-discharging shortens the lifespan of each battery cell and can lead to safety issues that endanger the vehicle and its passengers. Therefore, full lifecycle management of power battery modules is essential to protect them and prevent safety incidents.

[0003] Currently, the full life cycle management of power batteries faces severe challenges. For example, regarding the problem of micro-short circuits, tiny short circuits caused by metal dendrites piercing the diaphragm or manufacturing defects initially manifest as only a 0.1mV voltage fluctuation or a 0.5°C temperature rise, which is difficult for traditional battery management systems (BMS) to effectively identify.

[0004] In summary, how to effectively manage the entire life cycle of power batteries is a problem that needs to be solved. Summary of the Invention

[0005] The purpose of this application is to provide a power battery full cycle management method and system to address the deficiencies in the above-mentioned prior art, so as to improve the effectiveness of power battery full cycle management.

[0006] To achieve the above objectives, the technical solutions adopted in the embodiments of the present application are as follows: In a first aspect, an embodiment of the present application provides a power battery full-cycle management method, which is applied to a power battery full-cycle management system. The power battery full-cycle management system includes at least: an edge terminal and a cloud; the cloud terminal is deployed with a pre-trained digital twin model, a data prediction model, and a decision model; the method includes: The edge end obtains the real-time temperature of multiple areas on each electrode in each single battery in the actual battery module, obtains the temperature data of each electrode, and obtains the operating data of the actual battery module; The edge end preprocesses each of the temperature data and each of the operation data to obtain preprocessed temperature data and preprocessed operation data; The edge performs preliminary data analysis based on the preprocessed temperature data and the preprocessed operation data to obtain preliminary analysis data, and sends the preliminary analysis data, the preprocessed temperature data, and the preprocessed operation data to the cloud; The cloud inputs the pre-processed temperature data and the pre-processed operation data into the digital twin model, and the digital twin model performs dynamic embedding simulation and outputs dynamic operation data; The data prediction model predicts the parameter prediction results of the actual battery module based on the dynamic operation data and / or the preprocessed operation data; the decision model determines the processing strategy of the actual battery module based on the preliminary analysis data and / or the parameter prediction results.

[0007] Optionally, performing preliminary data analysis based on the preprocessed temperature data and the preprocessed operation data to obtain preliminary analysis data includes: Performing feature extraction on the voltage signal in each of the pre-processed operating data by wavelet transform to obtain the micro-short circuit characteristics of the actual battery module; Meshing the actual battery module to obtain a plurality of grids, each grid corresponding to an area of a pole piece in the actual battery module; Determine the temperature data of each grid based on the temperature data of each electrode; According to the temperature data of each grid, a three-dimensional temperature field is established; The micro-short circuit characteristics and the three-dimensional temperature field are used as the preliminary analysis data.

[0008] Optionally, the data prediction model includes: an aging prediction sub-model; The method of obtaining the parameter prediction result of the actual battery module by the data prediction model based on the dynamic operation data and / or the pre-processed operation data includes: The aging prediction sub-model performs aging prediction based on the historical cycle data of the actual battery module and the voltage, current and impedance in the preprocessed operating data, predicts the capacity decay rate and internal resistance growth rate of the actual battery module, and sends the capacity decay rate and internal resistance growth rate to the decision model.

[0009] Optionally, the data prediction model includes a remaining useful life prediction sub-model; The method of predicting the actual battery module parameter prediction result based on the dynamic operation data and / or pre-processed operation data by the data prediction model includes: The remaining service life prediction sub-model predicts the remaining service life of the actual battery module based on the experimental temperature and experimental state of charge corresponding to the actual battery module and the actual temperature and actual state of charge of the actual battery module, obtains a remaining service life prediction result, and sends the remaining service life prediction result to the decision model.

[0010] Optionally, the data prediction model includes: a risk prediction sub-model; The method of obtaining the parameter prediction result of the actual battery module by the data prediction model based on the dynamic operation data and / or the pre-processed operation data includes: The risk prediction sub-model predicts the risk result of lithium dendrites piercing the diaphragm of the actual battery module based on the lithium ion concentration gradient data in the kinetic operation data, and sends the risk result to the decision model in the cloud.

[0011] Optionally, the data prediction model includes: a stress prediction sub-model; The method of obtaining the parameter prediction result of the actual battery module by the data prediction model based on the dynamic operation data and / or the pre-processed operation data includes: The stress prediction sub-model predicts the stress distribution prediction result of the actual battery module according to the dynamic stress data in the dynamic operation data, and sends the stress distribution prediction result to the decision model.

[0012] Optionally, the data prediction model includes: a battery health level prediction sub-model; The method of obtaining the parameter prediction result of the actual battery module by the data prediction model based on the dynamic operation data and / or the pre-processed operation data includes: The battery health level prediction sub-model predicts the health level of the actual battery module based on the multi-dimensional feature parameters in the pre-processed operating data, and sends the health level prediction result to the decision model.

[0013] Optionally, determining the actual battery module processing strategy based on the preliminary analysis data and / or the parameter prediction result includes: The decision model dynamically adjusts the actual charging cut-off voltage of the battery module according to the capacity attenuation rate and the internal resistance growth rate in the parameter prediction result; The decision model dynamically adjusts the fast charging current of the actual battery module according to the stress distribution prediction result in the parameter prediction result; The decision model dynamically adjusts the actual temperature of the battery module according to the risk result of lithium dendrites piercing the diaphragm in the parameter prediction result; The decision model determines the actual usage status of the battery module according to the remaining service life prediction result in the parameter prediction result, and controls the use of the actual battery module according to the usage status; The decision model determines a cascade utilization matching result of the battery module according to a health level prediction result in the parameter prediction result, and controls the actual utilization result of the battery module according to the cascade utilization matching result; The decision model determines the safety status of the actual battery module according to the micro-short circuit characteristics in the preliminary analysis data, and determines whether to send micro-short circuit warning information according to the safety status.

[0014] Optionally, the performing dynamic embedding simulation by the digital twin model includes: The digital twin model corrects the stress data of the digital twin model according to the dynamic operation temperature in the dynamic operation data to obtain dynamic operation stress data.

[0015] In a second aspect, an embodiment of the present application further provides a power battery full cycle management system, which includes at least: an edge terminal, a cloud, and an application terminal; the cloud is deployed with a pre-trained digital twin model, a data prediction model, and a decision model; The cloud is used to execute the method steps executed by the cloud in the first aspect, and the edge is used to execute the method steps executed by the edge in the first aspect.

[0016] In a second aspect, an embodiment of the present application further provides a power battery full cycle management device, the device comprising: In a third aspect, an embodiment of the present application further provides an electronic device comprising: a processor, a storage medium and a bus, wherein the storage medium stores program instructions executable by the processor. When the application is running, the processor communicates with the storage medium through the bus, and the processor executes the program instructions to perform the steps of the power battery full-cycle management method described in the first aspect above.

[0017] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and the computer program is read and executes the steps of the power battery full-cycle management method described in the first aspect above.

[0018] The beneficial effects of this application are: The present application provides a power battery full-cycle management method and system. This method uses an edge device to obtain real-time temperatures of multiple regions on each electrode in each single cell in an actual battery module, obtain temperature data for each electrode, and obtain operational data for the actual battery module. Compared to the prior art method of only obtaining the temperature of the entire battery module, this embodiment obtains multiple real-time temperatures of each electrode, making subsequent calculations based on the obtained real-time temperatures more accurate. The edge device preprocesses each temperature data and each operational data to obtain preprocessed temperature data and preprocessed operational data. The edge device performs preliminary data analysis based on the preprocessed temperature data and preprocessed operational data to obtain preliminary analysis data, and sends the preliminary analysis data, preprocessed temperature data, and preprocessed operational data to the cloud. Data preprocessing and preliminary analysis performed on the edge device can make data processing faster and more timely. A data prediction model predicts parameter prediction results for the actual battery module based on the dynamic operational data and / or preprocessed operational data, and a decision model determines a processing decision for the actual battery module based on the preliminary analysis data and / or parameter prediction results. By building a digital twin model that integrates electrochemical mechanisms and pre-processed operating data, accurate mapping of microscopic states can be achieved. Moreover, through the collaboration of edge computing and cloud-based decision-making models, millisecond-level computing and real-time decision-making can be achieved, completing cross-temporal and spatial data analysis and strategic decision-making, significantly improving the management efficiency of the battery module throughout its life cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A schematic diagram of the architecture of a power battery full-cycle management system provided in an embodiment of the present application; Figure 2 A flowchart of a power battery full-cycle management method provided in an embodiment of the present application; Figure 3 A flow chart of a second method for managing a power battery throughout its life cycle provided in an embodiment of the present application; Figure 4 A schematic diagram of the workflow of a battery full cycle management system provided in an embodiment of the present application; Figure 5 A schematic diagram of a closed-loop management process for full lifecycle degradation characteristics provided in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of illustration and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts can be implemented out of sequence, and steps without logical context can be reversed or implemented simultaneously. In addition, those skilled in the art, under the guidance of the contents of this application, can add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0022] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0023] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0024] Figure 1 This is a schematic diagram of the architecture of a power battery full cycle management system provided in an embodiment of the present application, such as Figure 1 As shown, the power battery full-cycle management system may include: edge end, cloud end and application terminal, wherein the edge end is deployed with a pre-trained digital twin model, data prediction model and decision model.

[0025] like Figure 1 As shown, the edge end can be communicated with the application terminal and the cloud end respectively, the cloud end can also be distributed to communicate with the application terminal and the edge end, and the application terminal can also be communicated with the edge end and the cloud end respectively.

[0026] The application terminal may be, for example, a mobile phone, tablet computer, laptop computer, PDA, desktop computer, or other terminal device with computing and display capabilities. Specifically, the application terminal may be an application program in the terminal device, such as a mobile phone application (APP) or a computer application system.

[0027] Figure 2 This is a flow chart of a power battery full cycle management method provided in an embodiment of the present application, which is applied to the power battery full cycle management system mentioned above. Figure 2 As shown, the method may include: S101. The edge end obtains the real-time temperature of multiple areas on each electrode in each single battery in the actual battery module, obtains the temperature data of each electrode, and obtains the operating data of the actual battery module.

[0028] Specifically, optical fiber temperature sensors can be pre-deployed on multiple areas of each electrode on the actual battery module. Each optical fiber temperature sensor can then collect the real-time temperature of each area on each electrode of the actual battery module during the charge and discharge process, and the collected real-time temperature of each electrode is sent to the edge end. The edge end can then obtain the real-time temperature of multiple areas on each electrode in the actual battery module and obtain the temperature data of each electrode, that is, the temperature data of each electrode includes multiple real-time temperatures. The operating data of the battery module may include the voltage, current, internal resistance, porous electrode, and strain data of the actual battery module during the charge and discharge process. Specifically, a high-frequency impedance spectrum detection module can also be provided on the actual battery module, through which the internal resistance of each single cell in the actual battery module can be collected, and the collected internal resistances can be sent to the edge end.

[0029] S102 : The edge end preprocesses each temperature data and each operation data to obtain preprocessed temperature data and preprocessed operation data.

[0030] Specifically, the edge can time-align asynchronous data based on a high-precision clock source to ensure data consistency. For example, this can align the current obtained from current sampling and the voltage obtained from voltage sampling. This can also address issues such as packet loss. Memory-mapped files or lightweight networks can be used at the edge to implement low-latency data buffering to mitigate transient network jitter and ensure more accurate received data.

[0031] Optionally, after receiving the temperature and operating data at the edge, the data can be cleaned. Specifically, the temperature data obtained by high-frequency sampling can be reduced in dimension using a sliding window averaging method. The Sliding Window and Bottom-up (SWAB) algorithm can be used to retain key inflection points in the voltage curve, such as the point where constant current charging switches to constant voltage. Delta encoding and Zstandard compression can be used on the impedance spectrum data to achieve lossless compression.

[0032] S103. The edge performs preliminary data analysis based on the preprocessed temperature data and the preprocessed operation data to obtain preliminary analysis data, and sends the preliminary analysis data, the preprocessed temperature data, and the preprocessed operation data to the cloud.

[0033] The preliminary analysis may include, for example, anomaly detection, data feature extraction, and other analytical operations. Temperature analysis may be performed based on preprocessed temperature data, or based on the distribution of voltage, current, internal resistance, and strain data in preprocessed operational data. Preliminary analysis may also be performed in combination with temperature and operational data. Preliminary analysis data may include micro-short circuit characteristics of actual battery modules, preliminary temperature analysis results, and the relationship between temperature and stress.

[0034] Optionally, after obtaining the preliminary analysis data of the actual battery module, the edge end can send the preliminary analysis data obtained after the preliminary analysis and the preprocessed temperature data and preprocessed operation data obtained after preprocessing to the cloud.

[0035] S104. The cloud inputs the preprocessed temperature data and the preprocessed operation data into the digital twin model, and the digital twin model performs dynamic embedding simulation and outputs dynamic operation data.

[0036] Optionally, the digital twin model can be, for example, a pseudo-two-dimensional (P2D) model of the actual battery module. Specifically, the cloud can input the voltage and current from the preprocessed operating data into the digital twin model. Based on the input preprocessed operating data, the digital twin model can use the Butler-Volmer equation to accurately describe the insertion / extraction kinetics of lithium ions in the positive and negative electrode active materials, thereby outputting kinetic operating data. This kinetic operating data may include: kinetic impedance, lithium ion concentration gradient data, kinetic temperature, and kinetic stress data.

[0037] S105. The data prediction model predicts the parameter prediction results of the actual battery module based on the dynamic operation data and / or the pre-processed operation data, and the decision model determines the processing decision of the actual battery module based on the preliminary analysis data and / or the parameter prediction results.

[0038] Optionally, after the digital twin model performs a dynamic process and outputs dynamic operation data, the edge end can send the output dynamic data to the data prediction model in the cloud. The data prediction model can make predictions based on the received dynamic operation data and / or preprocessed operation data, thereby obtaining parameter prediction results of the actual battery module, wherein the parameter prediction results may include: remaining service life prediction results, risk prediction results of lithium dendrites piercing the diaphragm, stress prediction results, battery health level prediction results, etc.

[0039] Optionally, after the data prediction model predicts different parameter prediction results, all parameter prediction results can be sent to the data decision model. The data decision model can determine the processing strategy of the actual battery module based on the preliminary analysis data and / or the parameter prediction results. After determining the processing strategy of the actual battery module, the cloud can send a control instruction corresponding to the processing strategy to the actual battery module based on the determined processing strategy, so that the actual battery module performs the action indicated by the control instruction based on the received control instruction, and can also return the execution result to the cloud. By controlling the actual battery module through the processing decision in this embodiment, the cycle service life of the actual battery module can be increased by 12% to 15%.

[0040] In this embodiment, the edge acquires the real-time temperature of multiple regions on each electrode in each single cell in the actual battery module, obtains temperature data for each electrode, and obtains operational data for the actual battery module. Compared to the prior art, which only acquires the temperature of the entire battery module, this embodiment acquires multiple real-time temperatures of each electrode, making subsequent calculations based on the acquired real-time temperatures more accurate. The edge preprocesses each temperature data and each operational data to obtain preprocessed temperature data and preprocessed operational data. The edge performs preliminary data analysis based on the preprocessed temperature data and preprocessed operational data to obtain preliminary analysis data, and sends the preliminary analysis data, preprocessed temperature data, and preprocessed operational data to the cloud. Preprocessing and preliminary analysis of data at the edge can make data processing faster and more timely. The data prediction model predicts the parameter prediction results of the actual battery module based on the dynamic operational data and / or preprocessed operational data, and the decision model determines the processing decision for the actual battery module based on the preliminary analysis data and / or parameter prediction results. By building a digital twin model that integrates electrochemical mechanisms and pre-processed operating data, accurate mapping of microscopic states can be achieved. Moreover, through the collaboration of edge computing and cloud-based decision-making models, millisecond-level computing and real-time decision-making can be achieved, completing cross-temporal and spatial data analysis and strategic decision-making, significantly improving the management efficiency of the battery module throughout its life cycle.

[0041] Figure 3A flow chart of the second power battery full cycle management method provided in the embodiment of the present application is shown as follows: Figure 3 As shown, in the above S103, the edge end performs preliminary data analysis based on the pre-processed temperature data and the pre-processed operation data to obtain preliminary analysis data, which may include: S201 , extracting features of the voltage signal in the pre-processed operating data through wavelet transform to obtain micro-short circuit features of the actual battery module.

[0042] Specifically, the edge end can use the wavelet transform method to extract features from the voltage signal of each single cell in the preprocessed operating data, and obtain the micro-short circuit characteristics of each single cell in the actual battery module, so as to achieve low-latency response and effective identification of the micro-short circuit characteristics. After obtaining the micro-short circuit characteristics, the analyzed micro-short circuit characteristics can be sent to the cloud, so that the cloud can make decisions on the battery module based on the micro-short circuit characteristics, thereby achieving rapid response and rapid warning of micro-short circuits.

[0043] S202 , dividing the actual battery module into grids to obtain a plurality of grids, each grid corresponding to an area of a pole piece in the actual battery module.

[0044] Specifically, the actual battery module can be gridded at the edge to obtain multiple grids, which can then form a three-dimensional grid. The multiple grids can include a grid corresponding to each region in each electrode. For example, if electrode 1 is divided into 9 regions, region 1 can correspond to grid 1 in multiple grids, region 2 can correspond to grid 2 in multiple grids, region 3 can correspond to grid 3 in multiple grids, region 4 can correspond to grid 4 in multiple grids, and so on. The position of the grid is the position of each region in the electrode.

[0045] S203 , determining the temperature data of each grid according to the temperature data of each electrode.

[0046] Optionally, a fiber optic temperature sensor can be deployed in each area of the pole piece, or in some areas of the pole piece. For areas where fiber optic temperature sensors are deployed, the fiber optic temperature sensors can collect the real-time temperature of the area in real time. The temperature data of each grid is the real-time temperature collected by the fiber optic sensor deployed in that grid.

[0047] S204: Establish a three-dimensional temperature field based on the temperature data of each grid.

[0048] Specifically, the temperature data of each grid may be filled into each grid. Since the multiple grids are three-dimensional grids, the obtained multiple grids containing the temperature data construct the three-dimensional temperature field of the actual battery module.

[0049] S205. Use the micro-short circuit characteristics and the three-dimensional temperature field as preliminary analysis data.

[0050] Optionally, the preliminary analysis data of the battery module may include micro-short circuit characteristics of each single battery in the actual battery module and the three-dimensional temperature field of the actual battery module.

[0051] Optionally, abnormal data monitoring can also be performed at the edge. For example, undervoltage thresholds, overvoltage thresholds, overcurrent thresholds, and thermal runaway thresholds can be set. The undervoltage threshold can identify voltages below the undervoltage threshold, the overvoltage threshold can identify voltages above the overvoltage threshold, the overcurrent threshold can identify currents above the overcurrent threshold, and the thermal runaway threshold can identify temperatures above the thermal runaway threshold. When a voltage below the undervoltage threshold is identified, an undervoltage alarm message can be issued; when a voltage above the overvoltage threshold is identified, an overvoltage alarm message can be issued; when a current above the overcurrent threshold is identified, an overcurrent alarm message can be issued; and when a temperature above the thermal runaway threshold is identified, a thermal runaway alarm message can be issued. This can achieve timely detection of abnormal information and improve the accuracy of abnormality detection in battery modules.

[0052] In this embodiment, a three-dimensional temperature field of the actual battery module is constructed by the edge end based on the received pre-processed temperature data. Compared with the prior art in which subsequent data processing is only based on the temperature of the entire battery module, the data processing based on the three-dimensional temperature field in this embodiment is more accurate, and a preliminary data analysis is performed on the pre-processed operating data to obtain micro-short circuit characteristics, so that the extraction of micro-short circuit characteristics can be achieved through millisecond-level processing and analysis, thereby more effectively performing micro-short circuit warnings, and then effectively managing the actual battery module in micro-short circuits.

[0053] Optionally, the above-mentioned S105, obtaining the actual battery module parameter prediction result by the data prediction model based on the dynamic operation data and / or the pre-processed operation data, may include: Optionally, the data prediction model may include: an aging prediction sub-model.

[0054] Specifically, the aging prediction sub-model can perform aging prediction based on the historical cycle data of the actual battery module and the voltage, current, and impedance in the pre-processed operating data, predict the actual battery module capacity decay rate and internal resistance growth rate, and send the capacity decay rate and internal resistance growth rate to the decision model. Among them, the historical cycle data of the actual battery module may include, for example: the number of charge and discharge times, capacity decay curve, and internal resistance growth trend. The error of the predicted capacity decay rate is less than or equal to 2.8%, and the error of the internal resistance growth rate is less than or equal to 3.5%.

[0055] In this embodiment, by fusing the pre-processed real-time operation data and historical cycle data to perform aging prediction, the predicted capacity attenuation rate and internal resistance growth rate of the actual battery module can be made more accurate.

[0056] Optionally, the above-mentioned S105, obtaining the actual battery module parameter prediction result by the data prediction model based on the dynamic operation data and / or the pre-processed operation data, may include: Optionally, the data prediction model may also include a remaining useful life prediction sub-model.

[0057] Optionally, the remaining service life prediction sub-model predicts the remaining service life of the actual battery module based on the experimental temperature and experimental state of charge corresponding to the actual battery module and the actual temperature and actual state of charge of the actual battery module, obtains a remaining service life prediction result, and sends the remaining service life prediction result to the decision model in the cloud. The remaining service life of the actual battery module can be predicted at the initial stage of the cycle, where the initial stage of the cycle is, for example, a stage with less than 200 cycles. The prediction method in this embodiment can achieve an error of less than or equal to 3% in the remaining service life prediction.

[0058] Among them, conducting experiments on the actual battery module in a laboratory environment can obtain the mapping relationship between the experimental temperature and the experimental state of charge, and the mapping relationship between the actual temperature and the actual state of charge can also be obtained during the actual vehicle operation of the actual battery module. The remaining service life prediction sub-model can make predictions based on the mapping relationship between the experimental temperature and the experimental state of charge and the mapping relationship between the actual temperature and the actual state of charge to obtain the service life prediction result of the actual battery module.

[0059] In this embodiment, when predicting the service life of the actual battery module, the coupled influence of temperature and state of charge is taken into consideration, so that the predicted service life of the actual battery module is more accurate.

[0060] Optionally, the above-mentioned S105, obtaining the actual battery module parameter prediction result by the data prediction model based on the dynamic operation data and / or the pre-processed operation data, may include: Among them, the data prediction model can also include a risk prediction sub-model.

[0061] Specifically, the risk prediction sub-model predicts the risk results of lithium dendrites piercing the diaphragm of the actual battery module based on the lithium ion concentration gradient data in the dynamic operation data, and sends the predicted risk results to the decision module in the cloud. Since the change in the lithium ion concentration gradient data causes the temperature change of the actual battery module, if the temperature of the actual battery module is too high, it will cause lithium dendrites to pierce the diaphragm, thereby causing heat accumulation, where heat accumulation is thermal runaway. Therefore, by predicting the risk results of lithium dendrites piercing the diaphragm, the thermal runaway of the actual battery module can be controlled. The thermal runaway warning response speed can be reduced to less than 50ms.

[0062] Optionally, the above-mentioned S105, obtaining the actual battery module parameter prediction result by the data prediction model based on the dynamic operation data and / or the pre-processed operation data, may include: The data prediction model includes: stress prediction sub-model; The stress prediction sub-model can predict the stress distribution prediction results of the battery module based on the dynamic stress data in the dynamic operation data, and send the stress distribution prediction results to the decision model.

[0063] Optionally, the above-mentioned S105, obtaining the actual battery module parameter prediction result by the data prediction model based on the dynamic operation data and / or the pre-processed operation data, may include: Optionally, the data prediction model may further include: a battery health level prediction sub-model.

[0064] The battery health level prediction sub-model predicts the health level of the actual battery module based on the multi-dimensional feature parameters of the operating data, and sends the health level prediction result to the decision model. Among them, the multi-dimensional feature parameters may include, for example, multi-dimensional feature parameters such as the open circuit voltage, DC internal resistance, and current of the actual battery module. The multi-dimensional feature parameters are input into the battery health level prediction sub-model. The health level prediction sub-model can predict the health level of the actual battery module based on the multi-dimensional feature parameters to obtain the health level prediction result of the actual battery module. Among them, the battery health level prediction model is a model based on a deep belief network. The health level of the actual battery module predicted by the multi-dimensional feature parameters is more accurate, and the accurate health level prediction result based on the actual battery module makes the cascade utilization matching efficiency of the actual battery module higher. Through this embodiment, the accuracy of battery health status prediction is significantly improved, so that the error of the health level prediction accuracy is less than or equal to 2%.

[0065] Optionally, determining the actual processing decision of the battery module according to the preliminary analysis data and / or parameter prediction results in S105 may include: Specifically, the decision model dynamically adjusts the actual battery module's charge cutoff voltage based on the capacity decay rate and internal resistance growth rate in the parameter prediction results. Specifically, the decision model can utilize an improved multi-objective particle swarm optimization algorithm, using the capacity decay rate and internal resistance growth rate as optimization targets, to dynamically adjust the actual battery module's charge cutoff voltage and maximum charge / discharge rate. Specifically, if the capacity decay rate or internal resistance growth rate is high, the actual battery module's charge cutoff voltage is reduced.

[0066] Specifically, the decision model dynamically adjusts the fast charge current of the actual battery module based on the stress distribution prediction result in the parameter prediction result, so that the current accuracy is ±1A. For example, when the stress distribution prediction result exceeds the stress distribution threshold, the fast charge current is linearly reduced to the current threshold, so that the lithium dendrite growth rate is reduced. The method in this embodiment can suppress the lithium dendrite growth rate by 40%.

[0067] The decision model dynamically adjusts the actual battery module temperature based on the parameter prediction results of the risk of lithium dendrites piercing the separator. For example, if the risk of lithium dendrites piercing the separator exceeds the risk threshold, the decision model controls the actual battery module temperature within a preset temperature range.

[0068] The decision model determines the actual battery module usage status based on the remaining service life prediction results in the parameter prediction results, and controls the use of the actual battery module based on the usage status. For example, if the actual battery module usage status is overused, the actual battery module can be set as a waste battery module.

[0069] The decision model determines the battery module's cascade utilization matching results based on the health level prediction results in the parameter prediction results. The actual battery module utilization results are then managed and controlled based on the cascade utilization matching results. For example, if the actual battery module's health level is high, it can be used as the core battery module. This increases the battery reconfiguration matching efficiency to 90%.

[0070] The decision model determines the safety status of the actual battery module based on the micro-short circuit characteristics in the preliminary analysis data, and decides whether to send a warning message based on the safety status. If the micro-short circuit characteristics are abnormal, the actual battery module is determined to be in a dangerous state, and a micro-short circuit warning message can be issued.

[0071] Optionally, the digital twin model performing dynamic embedded simulation and outputting dynamic operation data in S104 may include: the digital twin model correcting stress data of the digital twin model based on the dynamic operation temperature in the dynamic operation data to obtain dynamic operation stress data, and using the dynamic operation stress data as one of the data in the dynamic operation data. Specifically, the stress data may be corrected using an Arrhenius correction formula to quantify the temperature to obtain the dynamic operation stress data.

[0072] Alternatively, an improved unscented Kalman filter (UKF) algorithm can be used to identify model parameters online, combined with a particle swarm optimization (PSO) algorithm to dynamically correct the solid-phase diffusion coefficient conductivity (error ≤ 3%) and the interface reaction impedance (error ≤ 5%).

[0073] An embodiment of the present application also provides a power battery full-cycle management system, which includes at least: an edge terminal, a cloud terminal, and an application terminal; a pre-trained digital twin model, a data prediction model, and a decision-making model are deployed on the cloud terminal.

[0074] The cloud is used to execute the method steps executed by the cloud in the aforementioned specific implementation manner, and the edge is used to execute the method steps executed by the edge in the aforementioned specific implementation manner.

[0075] Figure 4 A schematic diagram of the workflow of a battery full cycle management system provided in an embodiment of the present application, from Figure 4 The collaborative work between the edge and cloud in this application can be seen in the following: S301: Deploy the edge. S302: Run a lightweight LSTM network in the computing layer of the edge and use it to perform real-time tasks. Examples of these real-time tasks include constructing a three-dimensional temperature field and extracting micro-short features as described in the aforementioned specific implementation. The resolution of the constructed three-dimensional temperature field is 0.1°C, and the sensitivity of micro-short feature extraction is 0.05mV. S303: The data relay layer at the edge processes the data. Specifically, LoRa+5G dual-mode communication is used, with AES encryption and data desensitization processing to achieve secure aggregation of multi-node data, minimizing data transmission packet loss rates to less than 0.01%. S304: Determine whether the data is securely aggregated. If so, the data is sent to the cloud; if not, it is returned to the data relay layer for re-aggregation processing. S305: The cloud analyzes and predicts the data, obtaining parameter prediction results and processing strategies. S306: The parameter prediction results are sent to the application terminal for display.

[0076] Optionally, the Unity3D engine can be used to build a dynamic model of the battery digital twin on the application terminal. The edge end can send the operating data of the digital twin model during the embedded simulation process to the application terminal, so that the lithium ion concentration field, stress distribution and other microscopic states of the actual battery module can be mapped in real time on the application terminal.

[0077] Figure 5 This is a schematic diagram of a closed-loop management process for degradation characteristics throughout the entire lifecycle, provided by an embodiment of this application. During the manufacturing phase, data traceability is implemented, and 72 key process parameters, such as electrode coating uniformity (thickness deviation ≤ 1μm) and electrolyte injection volume (accuracy ±0.1mL), are entered into the initial database of the digital twin model, creating a "digital gene map" for the battery cell. During the operation phase, adaptive control is implemented in the cloud, dynamically adjusting the fast charging strategy based on real-time stress distribution (spatial resolution 1mm). When local stress in the negative electrode is detected to exceed 2MPa, the charging current is reduced from 3C to 1.5C, reducing the lithium dendrite growth rate by 60%. Simultaneously, through collaborative optimization of the balancing strategy in the cloud, the cell voltage variation within the module is controlled within ±10mV. During the retirement phase, a retired battery grading network based on a deep belief network (DBN) is established in the cloud. It inputs operational data on 12 characteristic parameters, such as open-circuit voltage (OCV) relaxation curves and DC internal resistance (DCR), and outputs a battery health grade (SOH grading error ≤ 2%), increasing the efficiency of cascade utilization matching from the traditional 70% to 92%. This application deploys a dedicated AI computing unit (such as an NPU chip with a computing power of ≥4TOPS) at the edge, running a pruned and optimized lightweight LSTM network, enabling millisecond-level real-time processing. High-precision temperature field reconstruction: Based on a distributed fiber optic sensor array (spatial density ≥8 sensors / module), a spatiotemporal convolution algorithm is used to reconstruct the three-dimensional temperature field with a resolution of 0.1°C, capable of identifying abnormal temperature rise gradients of 0.3°C / s. Micro-short feature extraction: Using wavelet packet transform combined with an attention mechanism, 0.05mV micro-short circuit features (frequency band 10-100Hz) are extracted from a 100Hz sampled voltage signal, achieving a detection sensitivity five times higher than traditional methods.

[0078] Optionally, a real-time decision-making model can be implemented at the edge. The embedded system completes fault classification (normal / warning / alarm) within 10ms and triggers active protection strategies, such as dynamically adjusting the maximum charge and discharge rate (with an accuracy of ±0.1C) to nip the risk of thermal runaway in the bud. K-anonymization (k ≥ 10) and homomorphic encryption can be implemented at the edge, combined with AES-256 encrypted transmission, to reduce the risk of sensitive data leakage (such as battery serial number and location information) by 98%, achieving privacy enhancement. This application utilizes LoRa (communication range 3km @ 125kHz) + 5G (uplink rate 100Mbps) collaborative transmission, and a dynamic channel selection algorithm (packet loss compensation rate > 99.9%) to ensure connection reliability under complex operating conditions.

[0079] The cloud-based intelligent evolution in this application can build a battery swarm intelligent analysis engine based on the Spark distributed platform: through federated learning, it aggregates millions of battery data from different automakers (processing PB-level data daily), trains the digital twin degradation model, and improves the accuracy of thermal runaway warning to 99.6%; applies deep reinforcement learning (DRL) to optimize the charging strategy, implements nonlinear current control in the SOC range of 20%-80% (adjustment step size ±0.5A), and extends the battery cycle life by 15%.

[0080] The application terminal in this application can include a three-dimensional visualization interface, using the Unity3D engine to build a dynamic model of the battery digital twin, real-time mapping of microscopic states such as lithium ion concentration field (color gradient resolution 256 levels), stress distribution (vector arrow density ≥ 50 / cm²), and support synchronous refresh of SOC / SOH / SOF parameters (frequency 30Hz).

[0081] The power battery full-cycle management system in this application supports the OPC UA protocol and docking with the MES / ERP system, achieving full-link traceability from battery cell production parameters to vehicle operation data, and improving the efficiency of root cause analysis (RCA) by 40%.

[0082] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0083] In addition, the functional units in the various embodiments of the present application can be integrated into a single processing unit, each unit can exist physically separately, or two or more units can be integrated into a single unit. If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0084] The above is only a specific implementation method of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the protection scope of the present application.

Claims

1. A power battery full cycle management method, characterized in that: Applied to a power battery full cycle management system, the power battery full cycle management system includes at least: an edge end and a cloud; the cloud is deployed with a pre-trained digital twin model, a data prediction model, and a decision model; the method includes: The edge end obtains the real-time temperature of multiple areas on each electrode in each single battery in the actual battery module, obtains the temperature data of each electrode, and obtains the operating data of the actual battery module; The edge end preprocesses each of the temperature data and each of the operation data to obtain preprocessed temperature data and preprocessed operation data; The edge performs preliminary data analysis based on the preprocessed temperature data and the preprocessed operation data to obtain preliminary analysis data, and sends the preliminary analysis data, the preprocessed temperature data, and the preprocessed operation data to the cloud; The cloud inputs the pre-processed temperature data and the pre-processed operation data into the digital twin model, and the digital twin model performs dynamic embedding simulation and outputs dynamic operation data; The data prediction model predicts the parameter prediction results of the actual battery module based on the dynamic operation data and / or the preprocessed operation data; the decision model determines the processing strategy of the actual battery module based on the preliminary analysis data and / or the parameter prediction results.

2. The power battery full cycle management method according to claim 1, characterized in that: The step of performing preliminary data analysis based on the pre-processed temperature data and the pre-processed operation data to obtain preliminary analysis data includes: Performing feature extraction on the voltage signal in each of the pre-processed operating data by wavelet transform to obtain the micro-short circuit characteristics of the actual battery module; Meshing the actual battery module to obtain a plurality of grids, each grid corresponding to an area of a pole piece in the actual battery module; Determine the temperature data of each grid based on the temperature data of each electrode; According to the temperature data of each grid, a three-dimensional temperature field is established; The micro-short circuit characteristics and the three-dimensional temperature field are used as the preliminary analysis data.

3. The power battery full cycle management method according to claim 1, characterized in that: The data prediction model includes: an aging prediction sub-model; The method of predicting the actual battery module parameter prediction result based on the dynamic operation data and / or pre-processed operation data by the data prediction model includes: The aging prediction sub-model performs aging prediction based on the historical cycle data of the actual battery module and the voltage, current and impedance in the preprocessed operating data, predicts the capacity decay rate and internal resistance growth rate of the actual battery module, and sends the capacity decay rate and internal resistance growth rate to the decision model.

4. The power battery full cycle management method according to claim 1, characterized in that: The data prediction model includes a remaining useful life prediction sub-model; The method of predicting the actual battery module parameter prediction result based on the dynamic operation data and / or pre-processed operation data by the data prediction model includes: The remaining service life prediction sub-model predicts the remaining service life of the actual battery module based on the experimental temperature and experimental state of charge corresponding to the actual battery module and the actual temperature and actual state of charge of the actual battery module, obtains a remaining service life prediction result, and sends the remaining service life prediction result to the decision model.

5. The power battery full cycle management method according to claim 1, characterized in that: The data prediction model includes: a risk prediction sub-model; The method of predicting the actual battery module parameter prediction result based on the dynamic operation data and / or pre-processed operation data by the data prediction model includes: The risk prediction sub-model predicts the risk result of lithium dendrites piercing the diaphragm of the actual battery module based on the lithium ion concentration gradient data in the kinetic operation data, and sends the risk result to the decision model in the cloud.

6. The power battery full cycle management method according to claim 1, characterized in that: The data prediction model includes: a stress prediction sub-model; The method of predicting the actual battery module parameter prediction result based on the dynamic operation data and / or pre-processed operation data by the data prediction model includes: The stress prediction sub-model predicts the stress distribution prediction result of the actual battery module according to the dynamic stress data in the dynamic operation data, and sends the stress distribution prediction result to the decision model.

7. The power battery full cycle management method according to claim 1, characterized in that: The data prediction model includes: a battery health level prediction sub-model; The method of predicting the actual battery module parameter prediction result based on the dynamic operation data and / or pre-processed operation data by the data prediction model includes: The battery health level prediction sub-model predicts the health level of the actual battery module based on the multi-dimensional feature parameters in the preprocessed operating data, obtains a health level prediction result, and sends the health level prediction result to the decision model.

8. The power battery full cycle management method according to any one of claims 3 to 7, characterized in that: Determining the actual battery module processing strategy based on the preliminary analysis data and / or the parameter prediction result includes: The decision model dynamically adjusts the actual charging cut-off voltage of the battery module according to the capacity attenuation rate and the internal resistance growth rate in the parameter prediction result; The decision model dynamically adjusts the fast charging current of the actual battery module according to the stress distribution prediction result in the parameter prediction result; The decision model dynamically adjusts the actual temperature of the battery module according to the risk result of lithium dendrites piercing the diaphragm in the parameter prediction result; The decision model determines the actual usage status of the battery module according to the remaining service life prediction result in the parameter prediction result, and controls the use of the actual battery module according to the usage status; The decision model determines a cascade utilization matching result of the battery module according to a health level prediction result in the parameter prediction result, and controls the actual utilization result of the battery module according to the cascade utilization matching result; The decision model determines the safety status of the actual battery module according to the micro-short circuit characteristics in the preliminary analysis data, and determines whether to send micro-short circuit warning information according to the safety status.

9. The power battery full cycle management method according to claim 1, characterized in that: The dynamic embedding simulation performed by the digital twin model includes: The digital twin model corrects the stress data of the digital twin model according to the dynamic operation temperature in the dynamic operation data to obtain dynamic operation stress data.

10. A power battery full cycle management system, characterized in that: The power battery full cycle management system includes at least: an edge terminal, a cloud terminal, and an application terminal; the cloud terminal is deployed with a pre-trained digital twin model, a data prediction model, and a decision model; The cloud is used to execute the method steps executed by the cloud according to any one of claims 1 to 9, and the edge is used to execute the method steps executed by the edge according to any one of claims 1 to 9.

Images