Lithium battery remaining service life prediction method and system and medium

Abstract

The invention relates to a lithium battery remaining service life prediction method and system and a medium. The method comprises the steps of data preprocessing, indirect health factor IHI extraction, fusion of the extracted indirect health factors IHI, and wavelet noise filtering of the fused indirect health factors IHI. Network optimization: dividing particle swarms in a particle swarm optimization (PSO), independently optimizing each sub-swarm, introducing sub-swarm merging and splitting to form a dynamic multi-swarm particle swarm optimization, coupling a forgetting gate and an input gate in an LSTM network structure, and inputting hyper-parameters needing to be manually determined in an improved CNN-LSTM network into the dynamic multi-swarm particle swarm optimization. Performing round optimization to finally determine hyper-parameters in the network; and predicting the remaining service life of the lithium battery. According to the method, more accurate parameter calculation is realized, and the problems of local optimum and the like are avoided.

Description

Technical Field

[0001] The present application relates to the field of predictive maintenance and health management of lithium batteries, and specifically to a method, system, and medium for predicting the remaining useful life of lithium batteries. Background Art

[0002] Lithium-ion batteries, as the core power source for applications such as electric vehicles, power grids, and consumer electronics, have become an indispensable component of modern society. The increasing number of charge and discharge cycles inevitably leads to a continuous decline in battery efficiency, ultimately impacting overall system performance and reliability. Establishing a reliable battery health prediction system is crucial to ensuring the safe and efficient operation of electronic devices and energy storage systems. Accurately estimating remaining useful life (RUL), a key indicator of battery health, supports users in establishing maintenance and replacement plans and making informed management decisions.

[0003] With the rise of machine learning and artificial intelligence technologies, data-driven battery life prediction and health assessment methods have flourished. Currently, academia has adopted a variety of models for battery RUL prediction, including decision trees (DTs), random forests (RFs), multi-layer perceptrons (MLPs), long short-term memory networks (LSTMs), recurrent neural networks (RNNs), and bidirectional LSTMs (Bi-LSTMs).

[0004] Among swarm intelligence optimization algorithms, the particle swarm optimization (PSO) has attracted widespread attention due to its simple structure and fast convergence speed. However, its core flaw is that particles tend to converge prematurely during the search process and fall into a local optimal solution. Summary of the Invention

[0005] The purpose of the embodiments of the present application is to provide a method, system and medium for predicting the remaining service life of a lithium battery, so as to achieve more accurate parameter calculation and avoid problems such as local optimum.

[0006] To achieve the above objectives, this application provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present application provides a method for predicting the remaining service life of a lithium battery, comprising the following specific steps:

[0008] Data preprocessing, extracting indirect health factors IHI, fusing the extracted indirect health factors IHI, and then performing wavelet denoising on the fused indirect health factors IHI;

[0009] Network optimization, the particle swarm in the particle swarm algorithm PSO is divided into Each subgroup is optimized independently and subgroup merging and splitting are introduced to form a dynamic multi-group particle swarm algorithm. The forget gate and input gate in the LSTM network structure are coupled so that the input value at the new moment can directly affect the information state of the forget gate; and two hole connections are added so that the forget gate and output gate can accept the memory cell state at the previous moment. As input, the hyperparameters that need to be manually determined in the improved CNN-LSTM network are input into the dynamic multi-swarm particle swarm algorithm, and a round of optimization is performed to finally determine the hyperparameters in the network;

[0010] To predict the remaining service life of lithium batteries, the pre-processed data is input into the network in the network optimization step for training to obtain a pre-trained model. The real data of lithium battery capacity decline and the obtained pre-trained model are input into the network in the network optimization step for model fine-tuning training to obtain the final model for capacity prediction.

[0011] The extraction of the indirect health factor IHI includes extracting an indirect health factor that can reflect the battery capacity degradation trend from the lithium-ion battery charge and discharge data. The extraction results are: the minimum voltage slope under different cycles, the time when the charging temperature reaches the peak, the discharge voltage at equal time intervals, the discharge temperature difference at equal time intervals, the constant voltage rise charging time, the internal resistance curve, the constant discharge voltage duration, the time when the discharge voltage reaches the minimum value, and the time when the discharge temperature reaches the peak.

[0012] The fusion of the extracted indirect health factor IHI is specifically performed by using a standardization formula to standardize the feature matrix composed of IHI. The calculation formula is as follows:

[0013] Select participants for analysis The principal components form the following matrix

[0014] ,

[0015] Where, Indicates the first Rank The elements of the columns, the matrix contains The elements indicate that the matrix is ​​composed of Has The elements of a feature.

[0016] ,

[0017] in Represents the original data; represents the mean of the original data; Representative Rank Normalized data for columns; Representative column mean; For the Column standard deviation,

[0018] Then perform eigenvalue decomposition on the correlation coefficient matrix and rotate the data to project it onto the principal component axis. Finally, calculate the cumulative contribution rate of the principal component. The calculation formula is as follows:

[0019] ,

[0020] in For the The contribution rate of the principal components; For the front Cumulative contribution rate of principal components; 、 Respectively The first and eigenvalues, take the top one with a cumulative contribution rate exceeding 0.99 The principal components are used as fused IHI.

[0021] The wavelet denoising of the fused indirect health factor IHI is specifically performed by using two threshold denoising processes. The first denoising threshold is determined by the Sqtwolog method, and the formula is as follows:

[0022] ,

[0023] Where, The threshold used for Sqtwolog threshold filtering.

[0024] The second denoising threshold is determined by the Minimax method, and its formula is as follows:

[0025] ,

[0026] in, is the threshold used for Minimax threshold filtering, and N is the cycle life of the lithium battery.

[0027] Specifically, the network optimization is as follows:

[0028] ,

[0029] in, , For particles At the moment speed and position; , For particles At the moment speed and position, For particles The best historical position, is the global optimal position of the particle, , , are the inertia weight and acceleration coefficient, , for Random numbers within

[0030] The particle swarm in PSO is divided into Each subgroup is optimized independently and subgroup merging and splitting are introduced. The formula is as follows:

[0031] ,

[0032] For the The subgroup Particles in The speed of time, For the The subgroup Particles in The speed of time, For the The subgroup Particles in At the time position, each subgroup only needs to maintain its own That's it,

[0033] If two subgroups , of and If the following formula is satisfied, it will be merged into a subgroup, and the optimal position of the new subgroup is ,

[0034] ,

[0035] in, is the distance threshold; if the diversity of the subgroup If the number of particles is lower than the threshold, the population is split into two subpopulations and particles are randomly assigned. The formula is as follows:

[0036] ,

[0037] in, is the average distance between particles, is the number of subgroups, For the The subgroup particles, is the subgroup center.

[0038] In a second aspect, an embodiment of the present application provides a lithium battery remaining service life prediction system, comprising:

[0039] The data preprocessing module extracts the indirect health factor IHI, fuses the extracted indirect health factor IHI, and then performs wavelet denoising on the fused indirect health factor IHI;

[0040] The network optimization module divides the particle swarm in the particle swarm algorithm PSO into Each subgroup is optimized independently and subgroup merging and splitting are introduced to form a dynamic multi-group particle swarm algorithm. The forget gate and input gate in the LSTM network structure are coupled so that the input value at the new moment can directly affect the information state of the forget gate; and two hole connections are added so that the forget gate and output gate can accept the memory cell state at the previous moment. As input, the hyperparameters that need to be manually determined in the improved CNN-LSTM network are input into the dynamic multi-swarm particle swarm algorithm, and a round of optimization is performed to finally determine the hyperparameters in the network;

[0041] The lithium battery remaining service life prediction module inputs the pre-processed data into the network in the network optimization step for training, obtaining a pre-trained model. The real data of lithium battery capacity degradation and the obtained pre-trained model are then input into the network in the network optimization step for model fine-tuning training, resulting in the final model for capacity prediction.

[0042] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores program code, and when the program code is executed by a processor, the steps of the method for predicting the remaining service life of a lithium battery as described above are implemented.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] ① Fully explore the hidden information in the lithium battery capacity degradation data and extract it in the form of indirect health index (IHI). Compared with the traditional direct use of lithium battery capacity degradation curve, it contains more valuable information and can better reflect the capacity degradation trend;

[0045] ② Split the large group in the particle swarm optimization (PSO) into multiple subgroups and add splitting and fusion between subgroups to achieve more accurate parameter calculation and avoid problems such as local optimality;

[0046] ③In CNN-LSTM, the LSTM network parameters are improved to enhance its long-term memory capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly introduced as follows. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those of ordinary skill in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0048] Figure 1 The method flowchart of the present application.

[0049] Figure 2 The LSTM network structure diagram.

[0050] Figure 3 The improved LSTM network structure diagram.

[0051] Figure 4 The prediction results of No. B0005 lithium battery under different prediction starting points.

[0052] Figure 5 The capacity prediction result evaluation index under different prediction starting points. DETAILED DESCRIPTION

[0053] The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0054] The term "comprising" or "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or equipment including the element.

[0055] The terms "first", "second", and the like are only used to distinguish one entity or operation from another entity or operation, and cannot be understood as indicating or implying relative importance, and cannot be understood as requiring or implying any such actual relationship or order between the entities or operations.

[0056] Referring to Figures 1 to 3 The present application provides a lithium battery remaining useful life prediction method based on dynamic multi-swarm particle swarm (DMS-PSO) and CNN-LSTM, mainly including the following steps.

[0057] Step 1: Data Preprocessing

[0058] 1.1 Extraction of Indirect Health Factor (IHI)

[0059] Indirect health factors that can reflect the battery capacity degradation trend are extracted from the charge and discharge data of lithium-ion batteries. The extraction results are: minimum voltage slope under different cycles, time when charging temperature reaches peak value, discharge voltage at equal time intervals, discharge temperature difference at equal time intervals, constant voltage rise charging time, internal resistance curve, constant discharge voltage duration, time when discharge voltage reaches minimum value and time when discharge temperature reaches peak value.

[0060] 1.2 Acquisition of fusion IHI

[0061] Use the standardization formula to standardize the feature matrix composed of IHI. The calculation formula is as follows:

[0062] ,

[0063] in Represents the original data; Representative Rank Normalized data for columns; Representative column mean; For the Column standard deviation.

[0064] Then use right The correlation coefficient matrix is ​​decomposed into eigenvalues ​​and the data is rotated and projected onto the principal component axis. Finally, the cumulative contribution rate of the principal component is calculated. The calculation formula is as follows:

[0065] ,

[0066] in For the The contribution rate of each principal component; For the front Cumulative contribution rate of principal components; 、 Respectively The first and eigenvalues.

[0067] The present invention takes the cumulative contribution rate exceeding 0.99 The principal components are used as fused IHI.

[0068] 1.3 Wavelet Noise Filtering

[0069] For the fusion IHI of the 1.2 output, twice threshold denoising is used for processing, the first denoising threshold is determined by the Sqtwolog method, and the formula is as follows:

[0070] ,

[0071] The second denoising threshold is determined by the Minimaxi method, and the formula is as follows:

[0072] ,

[0073] Wherein, is the threshold value used for Minimax threshold filtering, and N is the cycle life of the lithium battery.

[0074] Step 2: Network optimization

[0075] 2.1 Optimization of particle swarm algorithm (PSO)

[0076] The calculation formula of the particle swarm algorithm is as follows:

[0077] ,

[0078] Wherein, is the speed and position of the particle at time ; is the speed and position of the particle at time , is the historical optimal position of the particle , is the global optimal position of the particle, is the inertia weight and acceleration coefficient, is a random number in .

[0079] The particle swarm in the present application is divided into , and each subgroup is independently optimized and introduced into the subgroup merging and splitting, and the formula is as follows:

[0080] ,

[0081] is the speed of the particle in the th subgroup at time , is the speed of the particle in the th subgroup at time , is the speed of the particle in the th subgroup at time , is the speed of the particle in the th subgroup at time At the time position, each subgroup only needs to maintain its own That's it.

[0082] If two subgroups of and If the following formula is satisfied, it will be merged into a subgroup, and the optimal position of the new subgroup is .

[0083] ,

[0084] in, is the distance threshold; if the diversity of the subgroup If the number of particles is lower than the threshold, the population is split into two subpopulations and particles are randomly assigned. The formula is as follows:

[0085] ,

[0086] in, is the average distance between particles, is the number of subgroups, For the The subgroup particles, is the subgroup center.

[0087] 2.2 CNN-LSTM Network Structure Improvement

[0088] The present invention couples the forget gate and input gate in the LSTM network structure so that the input value at the new moment can directly affect the information state of the forget gate; and adds two hole connections so that the forget gate and output gate can accept the memory cell state at the previous moment. As input, it improves the network's long-term memory capability.

[0089] 2.3 CNN-LSTM Network Parameter Optimization

[0090] The hyperparameters that need to be manually determined in the improved CNN-LSTM network are input into the dynamic multi-swarm particle swarm algorithm (DMS-PSO), and rounds of optimization are performed to finally determine the hyperparameters in the network.

[0091] Step 3: Prediction of remaining service life of lithium batteries

[0092] 3.1 Pre-training

[0093] Input the preprocessed data obtained in step 1 into the network in 2.3 for training to obtain a pretrained model.

[0094] 3.2 Formal Training

[0095] The real data of lithium battery capacity decline and the pre-trained model obtained in 3.1 are input into the network in 2.3 for model fine-tuning training to obtain the final model for capacity prediction.

[0096] This network uses the NASA dataset and the B0005 lithium battery data for training. It predicts the remaining life of batteries B0006, B0007, and B0018, using 60, 70, and 80 cycles as the prediction starting point, respectively. Comparative tests are conducted using three benchmark networks: GRU, RNN, and LSTM. Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Square Error (RMSE) are used as evaluation indicators. The formulas are as follows:

[0097]

[0098]

[0099]

[0100] In the formula, the number of charge and discharge cycles of lithium batteries is given by Given, It is the actual value of lithium battery capacity; is the predicted value of the lithium battery capacity of the network in this article; 、 and The smaller the value of , the smaller the prediction error, that is, the better the prediction effect.

[0101] The final prediction results are as follows Figure 4 and Figure 5 .

[0102] Finally, here is a comparison table of prediction parameters of different networks at the same starting point:

[0103]

[0104] It can be seen that no matter how the prediction starting point is set, the prediction effect of this method is optimal.

[0105] The present application provides a system for predicting the remaining useful life of a lithium battery, comprising:

[0106] The data preprocessing module extracts the indirect health factor IHI, fuses the extracted indirect health factor IHI, and then performs wavelet denoising on the fused indirect health factor IHI;

[0107] The network optimization module divides the particle swarm in the particle swarm algorithm PSO into Each subgroup is optimized independently and subgroup merging and splitting are introduced to form a dynamic multi-group particle swarm algorithm. The forget gate and input gate in the LSTM network structure are coupled so that the input value at the new moment can directly affect the information state of the forget gate; and two hole connections are added so that the forget gate and output gate can accept the memory cell state at the previous moment. As input, the hyperparameters that need to be manually determined in the improved CNN-LSTM network are input into the dynamic multi-swarm particle swarm algorithm, and a round of optimization is performed to finally determine the hyperparameters in the network;

[0108] The lithium battery remaining service life prediction module inputs the pre-processed data into the network in the network optimization step for training, obtaining a pre-trained model. The real data of lithium battery capacity degradation and the obtained pre-trained model are then input into the network in the network optimization step for model fine-tuning training, resulting in the final model for capacity prediction.

[0109] An embodiment of the present application provides a computer-readable storage medium storing program code. When the program code is executed by a processor, the steps of the method for predicting the remaining useful life of a lithium battery as described above are implemented.

[0110] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0111] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0112] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0113] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0114] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0115] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0116] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology to store information. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0117] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for predicting the remaining service life of a lithium battery, characterized in that: The specific steps include: Data preprocessing, extracting indirect health factors IHI, fusing the extracted indirect health factors IHI, and then performing wavelet denoising on the fused indirect health factors IHI; Network optimization, the particle swarm in the particle swarm algorithm PSO is divided into Each subgroup is optimized independently and subgroup merging and splitting are introduced to form a dynamic multi-group particle swarm algorithm. The forget gate and input gate in the LSTM network structure are coupled so that the input value at the new moment can directly affect the information state of the forget gate; and two hole connections are added so that the forget gate and output gate can accept the memory cell state at the previous moment. As input, the hyperparameters that need to be manually determined in the improved CNN-LSTM network are input into the dynamic multi-swarm particle swarm algorithm, and a round of optimization is performed to finally determine the hyperparameters in the network; To predict the remaining service life of lithium batteries, the pre-processed data is input into the network in the network optimization step for training to obtain a pre-trained model. The real data of lithium battery capacity decline and the obtained pre-trained model are input into the network in the network optimization step for model fine-tuning training to obtain the final model for capacity prediction.

2. The method for predicting the remaining service life of a lithium battery according to claim 1, wherein: The extraction of the indirect health factor IHI includes extracting an indirect health factor that can reflect the battery capacity degradation trend from the lithium-ion battery charge and discharge data. The extraction results are: the minimum voltage slope under different cycles, the time when the charging temperature reaches the peak, the discharge voltage at equal time intervals, the discharge temperature difference at equal time intervals, the constant voltage rise charging time, the internal resistance curve, the constant discharge voltage duration, the time when the discharge voltage reaches the minimum value, and the time when the discharge temperature reaches the peak.

3. The method for predicting the remaining service life of a lithium battery according to claim 1, wherein: The fusion of the extracted indirect health factor IHI is specifically performed by using a standardization formula to standardize the feature matrix composed of IHI. The calculation formula is as follows: Select participants for analysis The principal components form the following matrix ， Where, Indicates the first Rank The elements of the columns, the matrix contains The elements indicate that the matrix is ​​composed of Has The elements of the characteristics, ， in Represents the original data; represents the mean of the original data; Representative Rank Normalized data for columns; Representative column mean; For the Column standard deviation, Then perform eigenvalue decomposition on the correlation coefficient matrix and rotate the data to project it onto the principal component axis. Finally, calculate the cumulative contribution rate of the principal component. The calculation formula is as follows: ， in For the The contribution rate of each principal component; For the front Cumulative contribution rate of principal components; 、 Respectively The first and eigenvalues, take the top one with a cumulative contribution rate exceeding 0.99 The principal components are used as fused IHI.

4. The method for predicting the remaining service life of a lithium battery according to claim 1, wherein: The wavelet denoising of the fused indirect health factor IHI is specifically performed by using two threshold denoising processes. The first denoising threshold is determined by the Sqtwolog method, and the formula is as follows: ， Where, is the threshold used for Sqtwolog threshold filtering, The second denoising threshold is determined by the Minimax method, and its formula is as follows: ， in, is the threshold used for Minimax threshold filtering, and N is the cycle life of the lithium battery.

5. The method for predicting the remaining service life of a lithium battery according to claim 1, wherein: Specifically, the network optimization is as follows: ， in, , For particles At the moment speed and position; , For particles At the moment speed and position, For particles The best historical position, is the global optimal position of the particle, , , are the inertia weight and acceleration coefficient, , for Random numbers within The particle swarm in PSO is divided into Each subgroup is optimized independently and subgroup merging and splitting are introduced. The formula is as follows: ， For the The subgroup Particles in The speed of time, For the The subgroup Particles in The speed of time, For the The subgroup Particles in At the time position, each subgroup only needs to maintain its own That's it, If two subgroups , of and If the following formula is satisfied, it will be merged into a subgroup, and the optimal position of the new subgroup is , ， in, is the distance threshold; if the diversity of the subgroup If the number of particles is lower than the threshold, the population is split into two subpopulations and particles are randomly assigned. The formula is as follows: ， in, is the average distance between particles, is the number of subgroups, For the The subgroup particles, is the subgroup center.

6. A lithium battery remaining service life prediction system, characterized in that: include, The data preprocessing module extracts the indirect health factor IHI, fuses the extracted indirect health factor IHI, and then performs wavelet denoising on the fused indirect health factor IHI; The network optimization module divides the particle swarm in the particle swarm algorithm PSO into Each subgroup is optimized independently and subgroup merging and splitting are introduced to form a dynamic multi-group particle swarm algorithm. The forget gate and input gate in the LSTM network structure are coupled so that the input value at the new moment can directly affect the information state of the forget gate; and two hole connections are added so that the forget gate and output gate can accept the output of the memory cell at the previous moment. , the hyperparameters that need to be manually determined in the improved CNN-LSTM network are input into the dynamic multi-group particle swarm algorithm, and round-trip optimization is performed to finally determine the hyperparameters in the network; The lithium battery remaining service life prediction module inputs the data preprocessing data into the network in the network optimization step for training to obtain a pre-trained model, and inputs the real data of lithium battery capacity decline and the obtained pre-trained model into the network in the network optimization step for model fine-tuning training to obtain the final model for capacity prediction.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program code, and when the program code is executed by a processor, the steps of the method for predicting the remaining service life of a lithium battery according to any one of claims 1 to 5 are implemented.

Images