Battery capacity recession mode quantification method and device, electronic equipment and storage medium

Abstract

The invention provides a battery capacity recession mode quantification method and device, electronic equipment and a storage medium. The method comprises the following steps: constructing a two-dimensional matrix based on pseudo open-circuit voltage curves of an aged lithium battery and a fresh lithium battery; inputting the two-dimensional matrix into a deep learning model to output a capacity degradation mode of the lithium battery, wherein the capacity degradation mode is a capacity degradation path of the lithium battery on an electrode layer; and determining a deep learning model of the second type of lithium battery according to the difference between the first type of lithium battery and the second type of lithium battery in the charging curve shape. According to the method, the two-dimensional data matrix is constructed through the pseudo open-circuit voltage curves of the aged lithium battery and the fresh lithium battery, the deep learning model is input to output the degradation path of the lithium battery on the electrode layer, and the deep learning model component of the cross-type lithium battery is optimized in combination with the transfer learning strategy; accurate quantitative analysis of the capacity recession mode of the electrode layer of the battery with different chemical systems is realized.

Description

Technical Field

[0001] This application relates to the field of battery technology, specifically to a method, apparatus, electronic device, and storage medium for quantifying battery capacity degradation modes. Background Technology

[0002] Lithium-ion batteries are a type of battery that uses lithium metal or lithium alloys as positive / negative electrode materials and a non-aqueous electrolyte solution. As a key energy component widely used in electronic devices, electric vehicles, and other fields, accurate assessment of battery health status is crucial for the safe and reliable operation of products using lithium-ion batteries. Existing technologies typically focus on battery life prediction and SOX (SOX State of X, referring to the battery's multi-dimensional state parameters) based on physical models and data-driven methods. SOX includes, but is not limited to, predictions of SOC (State of Charge), SOH (State of Health, which can be assessed through indicators such as capacity decay and internal resistance increase), and SOP (State of Power). However, battery aging is not only manifested in capacity decay and increased internal resistance but also involves electrode-level side reactions that are less studied in existing technologies. Furthermore, current battery degradation diagnosis methods generally suffer from low computational efficiency, dependence on specific experimental conditions, and difficulty in determining battery degradation mechanisms. Summary of the Invention

[0003] The main objective of this application is to propose a method, apparatus, electronic device, and storage medium for quantifying battery capacity degradation modes, aiming to improve the accuracy of quantifying lithium battery capacity degradation modes.

[0004] This application provides a method for quantifying battery capacity degradation modes, including: acquiring a first pseudo-open-circuit voltage curve and a second pseudo-open-circuit voltage curve of a first type of lithium battery; constructing a two-dimensional data matrix based on the difference between the first pseudo-open-circuit voltage curve and the second pseudo-open-circuit voltage curve; wherein the first pseudo-open-circuit voltage curve corresponds to an aged first type of lithium battery, and the second pseudo-open-circuit voltage curve corresponds to a fresh first type of lithium battery; inputting the two-dimensional data matrix into a deep learning model to output the capacity degradation mode of the first type of lithium battery, wherein the capacity degradation mode is the capacity degradation path of the first type of lithium battery at the electrode level; determining a deep learning model of the second type of lithium battery based on the difference between the charging curve shape of the first type of lithium battery and the charging curve shape of the second type of lithium battery; wherein the deep learning model of the second type of lithium battery includes at least some components of the deep learning model of the first type of lithium battery.

[0005] In one embodiment, constructing a two-dimensional data matrix based on the first pseudo-open-circuit voltage curve and the second pseudo-open-circuit voltage curve includes:

[0006] The first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve are interpolated and aligned within the same voltage range;

[0007] Calculate the capacity difference curve between the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve at each voltage sampling point;

[0008] The capacity difference curve is transformed into a two-dimensional data matrix, wherein the rows of the two-dimensional data matrix correspond to the voltage sampling points, and the columns of the two-dimensional data matrix correspond to the capacity difference.

[0009] In one embodiment, the capacity degradation mode includes at least one of the following: lithium-ion loss mode, positive electrode active material loss mode, negative electrode active material loss mode, and internal resistance increase mode.

[0010] In one embodiment, the deep learning model includes a convolutional neural network module and a physical information neural network module. The convolutional neural network module is used to output the capacity degradation mode of the first type of lithium battery, and the physical information neural network module is used to adjust the convolutional neural network module according to the capacity degradation mode.

[0011] In one embodiment, the convolutional neural network module includes convolutional layers, pooling layers, dense layers, and fully connected layers. The step of inputting the two-dimensional data matrix into a deep learning model to output the capacity degradation mode of the first type of lithium battery includes:

[0012] The two-dimensional data matrix is ​​input into the convolutional neural network module;

[0013] Based on the convolutional layer, multi-scale local aging features of the first type of lithium battery are extracted from the two-dimensional data matrix;

[0014] The global aging characteristics of the first type of lithium battery are determined by combining the multi-scale local features with the pooling layer.

[0015] The multi-scale local aging features and the global aging features are fused through the dense layer to obtain the fused aging features;

[0016] The fusion aging characteristics are mapped to the capacity degradation mode of the first type of lithium battery based on the fully connected layer.

[0017] In one embodiment, the step of inputting the two-dimensional data matrix into a deep learning model to output the capacity degradation mode of the first type of lithium battery further includes:

[0018] The two-dimensional data matrix is ​​input into the physical information neural network module to output the capacity degradation reference mode of the first type of lithium battery;

[0019] If the capacity decay reference mode is inconsistent with the capacity decay mode or has a large deviation, the convolutional neural network module is adjusted.

[0020] In one embodiment, the battery capacity degradation mode quantification method further includes, before inputting the two-dimensional data matrix into the deep learning model:

[0021] Construct a convolutional neural network model;

[0022] A training dataset is synthesized based on an open-circuit voltage model. The training dataset includes multiple capacity difference samples labeled with capacity degradation modes. The open-circuit voltage model simulates battery voltage based on the positive and negative electrode potentials of the battery.

[0023] The convolutional neural network model is trained based on the training dataset and the loss function.

[0024] The trained convolutional neural network model is deployed to the convolutional neural network module of the deep learning model.

[0025] This application embodiment also provides a battery capacity degradation mode quantification device, including a data processing module, a deep learning module, and a transfer learning module; the data processing module is used to acquire a first pseudo-open circuit voltage curve and a second pseudo-open circuit voltage curve of a first type of lithium battery, and construct a two-dimensional data matrix based on the difference between the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve; wherein, the first pseudo-open circuit voltage curve corresponds to an aged first type of lithium battery, and the second pseudo-open circuit voltage curve corresponds to a fresh first type of lithium battery; the deep learning module is used to input the two-dimensional data matrix into a deep learning model to output the capacity degradation mode of the first type of lithium battery, wherein the capacity degradation mode is the capacity degradation path of the first type of lithium battery at the electrode level; the transfer learning module is used to determine a deep learning model of the second type of lithium battery based on the difference between the charging curve shape of the first type of lithium battery and the charging curve shape of the second type of lithium battery; wherein, the deep learning model of the second type of lithium battery includes at least some components of the deep learning model of the first type of lithium battery.

[0026] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method for quantifying battery capacity degradation modes.

[0027] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for quantifying battery capacity degradation modes.

[0028] This application provides a method, apparatus, electronic device, and storage medium for quantifying battery capacity degradation modes. It acquires a first pseudo-open-circuit voltage curve and a second pseudo-open-circuit voltage curve of a first type of lithium battery, and constructs a two-dimensional data matrix based on the difference between the two curves. The first pseudo-open-circuit voltage curve corresponds to an aged first-type lithium battery, and the second pseudo-open-circuit voltage curve corresponds to a fresh first-type lithium battery. The two-dimensional data matrix is ​​input into a deep learning model to output the capacity degradation mode of the first-type lithium battery, which is the capacity degradation path at the electrode level. Based on the difference between the charging curve shape of the first-type lithium battery and the charging curve shape of the second-type lithium battery, a deep learning model for the second-type lithium battery is determined. The deep learning model for the second-type lithium battery includes at least some components of the deep learning model for the first-type lithium battery. This application constructs a two-dimensional data matrix using the pseudo-open-circuit voltage curves of aged and fresh lithium batteries, inputs it into a deep learning model to output the degradation path at the electrode level, and combines a transfer learning strategy to optimize the deep learning model components for cross-type lithium batteries, achieving accurate quantitative analysis of capacity degradation modes at the electrode level for batteries with different chemical systems. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating the battery capacity degradation mode quantification method provided in the embodiments of this application;

[0030] Figure 2 This is a schematic diagram of the architecture of the deep learning model provided in the embodiments of this application;

[0031] Figure 3 This is a schematic diagram of the process of transferring a deep learning model to different types of lithium batteries according to an embodiment of this application;

[0032] Figure 4 This is a schematic diagram of the specific process of the battery capacity degradation mode quantification method provided in the embodiments of this application;

[0033] Figure 5 This is one of the schematic diagrams of the capacity decay mode output by the deep learning model provided in the embodiments of this application;

[0034] Figure 6 This is the second schematic diagram of the capacity decay mode output by the deep learning model provided in the embodiments of this application;

[0035] Figure 7This is one of the schematic diagrams showing the comparison between the actual value and the predicted value of the capacity decay mode provided in the embodiments of this application;

[0036] Figure 8 This is the second schematic diagram showing the comparison between the actual value and the predicted value of the capacity decay mode provided in the embodiments of this application;

[0037] Figure 9 This is the third schematic diagram showing the comparison between the actual value and the predicted value of the capacity decay mode provided in the embodiments of this application;

[0038] Figure 10 This is the fourth schematic diagram showing the comparison between the actual value and the predicted value of the capacity decay mode provided in the embodiments of this application;

[0039] Figure 11 This is one of the schematic diagrams of the capacity degradation mode output after transferring the deep learning model to different types of lithium batteries according to the embodiments of this application;

[0040] Figure 12 This is the second schematic diagram of the capacity degradation mode output after transferring the deep learning model to different types of lithium batteries according to the embodiments of this application;

[0041] Figure 13 This is a schematic diagram of the battery capacity degradation mode quantification device provided in the embodiments of this application;

[0042] Figure 14 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0043] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0044] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0045] The battery capacity degradation mode quantification method provided in this application embodiment can be applied to electronic devices and their software. The electronic device can be a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The software can be an application implementing the battery capacity degradation mode quantification method, but is not limited to the above forms.

[0046] The method for quantifying battery capacity degradation modes provided in this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] The performance degradation of lithium-ion batteries is a complex process resulting from the combined effects of changes in the microstructure of electrode materials and internal electrochemical side reactions. At the negative electrode, loss of active material (LAMNE) manifests as the cracking and exfoliation of graphite particles, reducing the effective reaction interface and compromising the integrity of the electrode structure. Simultaneously, under fast charging or low-temperature conditions, hindered lithium-ion intercalation leads to the precipitation of metallic lithium, which not only consumes active lithium but may also form dendrites, threatening battery safety. Dendrites refer to dendritic crystals formed during the solidification of liquid metals or alloys. Furthermore, the continuous reduction of the electrolyte on the negative electrode surface to form a solid electrolyte interphase (SEI) film, the excessive growth of which irreversibly depletes the lithium inventory (LLI). The decomposition of the SEI film under high-temperature or overcharge conditions further accelerates electrolyte degradation, creating a vicious cycle. These degradation mechanisms at the negative electrode level directly lead to a decrease in reversible battery capacity and coulombic efficiency. Coulomb efficiency refers to the ratio of the amount of lithium ions returning to the positive electrode during discharge to the amount of lithium ions leaving the positive electrode during charging in a complete cycle.

[0048] On the negative electrode side, the loss of active material in the positive electrode (LAMPE) originates from transition metal dissolution and lattice phase transitions. Acidic byproducts in the electrolyte corrode the positive electrode material, and dissolved metal ions migrate to the negative electrode, damaging the SEI film. Lattice reconstruction caused by high voltage or localized SOC inhomogeneity reduces lithium-ion diffusion kinetics. The cathode-electrolyte interphase (CEI) film formed by side reactions at the positive electrode-electrolyte interface increases charge transfer resistance, further deteriorating battery power performance. Positive electrode degradation not only reduces lithium intercalation sites, but its synergistic effect with negative electrode degradation also accelerates the performance degradation of the entire battery system. Macroscopic faults in lithium batteries, such as internal and external short circuits, overcharging and over-discharging, and electrolyte leakage, are external manifestations of microscopic degradation. Lithium dendrite growth may puncture the separator, causing internal short circuits; overcharging leads to positive electrode structural collapse; and over-discharging causes copper current collector dissolution. These faults can cause irreversible capacity loss or, in severe cases, thermal runaway.

[0049] To address the aforementioned challenges, embodiments of this application provide a method for quantifying battery capacity degradation modes. Please refer to... Figure 1 A method for quantifying battery capacity degradation modes provided in this application embodiment may include:

[0050] Step S101: Obtain the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve of the first type of lithium battery, and construct a two-dimensional data matrix based on the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve; wherein, the first pseudo-open circuit voltage curve corresponds to the aged first type of lithium battery, and the second pseudo-open circuit voltage curve corresponds to the fresh first type of lithium battery.

[0051] Optionally, the pseudo-open circuit voltage (pOCV) curve refers to a curve that approximates the true open circuit voltage, obtained by simulating the battery through low-rate constant current charge-discharge or resting voltage interpolation. The pOCV curve is used to characterize the relationship between battery voltage and capacity. The open circuit voltage of a battery refers to the potential difference between the positive and negative electrodes when the battery is in an open-circuit state (i.e., no current flows through the battery). "Fresh" refers to a battery that has not undergone charge-discharge cycles or has only experienced a very small number of cycles and whose electrochemical performance is in its initial design state. "Aging" refers to a battery that has undergone multiple charge-discharge cycles, resulting in a significant decrease in performance such as capacity and energy density due to mechanisms such as lithium-ion loss, degradation of active materials, or increased internal resistance.

[0052] Optionally, when obtaining the pseudo-open-circuit voltage curve, a low-rate constant-current charge-discharge protocol can be applied to the lithium battery, and voltage-capacity relationship data can be recorded simultaneously. First, a second pseudo-open-circuit voltage curve can be generated based on the charging data of the lithium battery's first cycle. Then, under the same test conditions, the charging data of the lithium battery's Nth cycle is obtained to generate a first pseudo-open-circuit voltage curve. For example, the value of N can be greater than or equal to 100 to ensure that the capacity of the target battery decreases significantly.

[0053] Optionally, when obtaining the pseudo-open-circuit voltage curve, lithium batteries from the same batch can be divided into two groups: one group consists of aged batteries that have undergone multiple cycles, and the other group consists of unused fresh batteries. Under preset test conditions, low-rate constant-current charge-discharge cycles are performed on both the aged and fresh batteries, and the charge-discharge voltage curves throughout the process are recorded. Then, a first pseudo-open-circuit voltage curve is generated based on the charge-discharge voltage curve of the aged battery group, and a second pseudo-open-circuit voltage curve is generated based on the charge-discharge voltage curve of the fresh batteries.

[0054] Optionally, the pseudo-open circuit voltage curve is a voltage-capacity curve. The capacity difference curve between the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve is transformed into a two-dimensional data matrix of voltage-capacity for the first type of lithium battery. In other words, the pseudo-open circuit voltage difference between aged and fresh batteries is transformed into spatial distribution characteristics, thereby realizing a visual characterization of battery capacity degradation.

[0055] In one embodiment, a two-dimensional data matrix is ​​constructed based on the first pseudo-open-circuit voltage curve and the second pseudo-open-circuit voltage curve, including:

[0056] The first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve are interpolated and aligned within the same voltage range.

[0057] Calculate the capacity difference curve between the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve at each voltage sampling point;

[0058] The capacity difference curve is transformed into a two-dimensional data matrix, where the rows of the two-dimensional data matrix correspond to voltage sampling points and the columns of the two-dimensional data matrix correspond to capacity differences.

[0059] Optionally, by interpolating and aligning the first and second pseudo-open-circuit voltage curves within the same voltage range, curve offsets caused by sampling errors or differences in measurement conditions can be eliminated, ensuring accurate correspondence between the two curves in the voltage dimension. Based on this, the capacity difference between the two curves at each voltage sampling point is calculated to obtain the capacity difference curve for the first type of lithium battery. This quantifies the capacity variation characteristics of the lithium battery in different voltage ranges, thereby revealing the impact of the battery's internal capacity degradation mode on the voltage response. Next, the capacity difference curve is transformed into a two-dimensional data matrix, where rows correspond to voltage sampling points and columns correspond to capacity differences. This not only achieves structured storage of multi-dimensional data but also provides data format support for subsequent matrix-based feature extraction.

[0060] Step S102: Input the two-dimensional data matrix into the deep learning model to output the capacity degradation mode of the first type of lithium battery. The capacity degradation mode is the capacity degradation path of the first type of lithium battery at the electrode level.

[0061] Capacity degradation modes describe the capacity degradation paths of lithium-ion batteries at the electrode level, including at least one of the following: lithium-ion loss mode, positive electrode active material loss mode, negative electrode active material loss mode, and internal resistance increase mode. The lithium-ion loss mode refers to the path of continuous capacity decay caused by the irreversible consumption of cyclic lithium ions due to mechanisms such as electrolyte decomposition, continuous SEI film growth, or lithium dendrite shedding. The positive electrode active material loss mode refers to the path of capacity loss caused by a decrease in the lithium-ion insertion / extraction ability of the positive electrode material due to failure mechanisms such as structural distortion, transition metal dissolution, or particle fracture. The negative electrode active material loss mode refers to the path of capacity decay caused by the consumption of active lithium by SEI film thickening, structural pulverization due to graphite interlayer stress expansion, or lithium deposition (metallic lithium) leading to side reactions, thereby reducing the reversible lithium-ion insertion / extraction efficiency. The internal resistance increase mode refers to the path of capacity degradation caused by a continuous increase in battery internal resistance due to factors such as electrode material aging, electrolyte drying, increased SEI film impedance, or intensified interfacial polarization.

[0062] Different capacity degradation modes lead to shifts, slope changes, or feature point offsets in the pOCV curve. Therefore, by observing the morphological changes of the pOCV curve, the internal capacity degradation mechanism of the battery can be identified and quantified. The deep learning model in this application embodiment, given an input two-dimensional data matrix constructed based on the pseudo-open-circuit voltage curve of a first type of lithium battery, can predict different capacity degradation modes of the output lithium battery and their corresponding capacity change curves.

[0063] In one embodiment, the deep learning model includes a convolutional neural network module and a physical information neural network module. The convolutional neural network module is used to output the capacity degradation mode of the first type of lithium battery, and the physical information neural network module is used to adjust the convolutional neural network module according to the capacity degradation mode.

[0064] A Convolutional Neural Network (CNN) is a deep learning architecture used to process grid-based topological data (such as images and matrices). A Physics Informed Neural Network (PINN) is a neural network that encodes physical laws (such as differential equations) as constraints in a loss function.

[0065] Specifically, a two-dimensional data matrix constructed based on the pseudo-open-circuit voltage curve of a first-type lithium battery is input into a CNN module. The CNN module extracts the spatial features of the two-dimensional data matrix through multi-layer convolutional kernels and pooling operations, outputting the capacity degradation mode of the first-type lithium battery at the positive and negative electrode material levels. Subsequently, this capacity degradation mode is input into a PINN module, which uses the input capacity degradation mode to correct the parameters of the CNN. Thus, the deep learning model in this embodiment overcomes the dependence of traditional deep learning models on large-scale labeled data through the collaborative mechanism of the CNN module and the PINN module, while avoiding the limitations of pure physical models in modeling complex degradation processes.

[0066] In one embodiment, the convolutional neural network module includes convolutional layers, pooling layers, dense layers, and fully connected layers. A two-dimensional data matrix is ​​input into a deep learning model to output the capacity degradation mode of a first type of lithium battery, including:

[0067] Input the two-dimensional data matrix into the convolutional neural network module;

[0068] Multi-scale local aging features of the first type of lithium battery are extracted from a two-dimensional data matrix based on convolutional layers.

[0069] The global aging characteristics of the first type of lithium battery are determined by combining multi-scale local features through the pooling layer;

[0070] By fusing multi-scale local aging features and global aging features in a dense layer, a fused aging feature is obtained.

[0071] The fully connected layer maps the fusion aging characteristics to the capacity degradation mode of the first type of lithium battery.

[0072] Optionally, a two-dimensional data matrix is ​​input into a convolutional neural network module to identify the first type of lithium battery capacity degradation mode. The convolutional neural network module consists of two parts: a feature extraction part and a feature-output mapping part. The feature extraction part consists of convolutional layers and pooling layers, while the feature-output mapping part consists of dense layers and fully connected layers. The convolutional layers use various kernel sizes (e.g., three types) to extract multi-scale local aging features from the two-dimensional data matrix, capturing the degradation characteristics of the lithium battery during the microscopic electrochemical process. The pooling layers (e.g., max-pooling and average-pooling layers) extract global aging features from the two-dimensional data matrix through dimensionality reduction. The voltage response patterns of different degradation mechanisms can be distinguished through multi-scale local aging features and global aging features. Then, two dense layers nonlinearly fuse the multi-scale local aging features and global aging features to generate a fused aging feature vector, further strengthening the correlation between features. Finally, a fully connected layer maps the fused aging features to four preset capacity degradation modes, achieving capacity degradation mode identification at the lithium battery electrode level.

[0073] In one embodiment, inputting a two-dimensional data matrix into a deep learning model to output the capacity degradation mode of a first type of lithium battery further includes:

[0074] The two-dimensional data matrix is ​​input into the physical information neural network module to output a capacity degradation reference pattern for the first type of lithium battery.

[0075] If the capacity decay reference pattern is inconsistent with the capacity decay pattern or there is a large deviation, then the convolutional neural network module should be adjusted.

[0076] Optionally, the two-dimensional data matrix is ​​simultaneously input into both the Physical Information Neural Network (PINN) module and the Convolutional Neural Network (CNN) module. The PINN module calculates a capacity degradation reference pattern based on battery electrochemistry principles, providing a benchmark that conforms to physical laws for the model. The CNN module outputs the current capacity degradation pattern through multi-scale feature extraction and nonlinear fusion. If the outputs of the two modules are inconsistent or show significant deviations, an adaptive adjustment mechanism is triggered: the parameters of the CNN module are optimized using a backpropagation algorithm, such as adjusting the number and size of the convolutional kernels, so that the output of the CNN module approximates the physical constraint results of the PINN module. In this way, dual verification and dynamic calibration of the first type of lithium battery capacity degradation pattern are achieved, improving the robustness and generalization ability of the deep learning model in recognizing lithium battery capacity degradation patterns.

[0077] Please refer to Figure 2This paper presents a deep learning model that integrates a Convolutional Neural Network (CNN) module and a Physical Information Neural Network (PINN) module to predict the capacity degradation mode of lithium-ion batteries. Input data is sequentially processed through convolutional layers, pooling layers, and dense layers for feature extraction and fusion. The final output, presented in a fully connected layer, shows the capacity degradation mode of the lithium-ion battery, including lithium active material loss (LLI), positive electrode active material loss (LAMdePE), negative electrode active material loss (LAMdeNE), and internal resistance increase (RI), enabling multi-dimensional quantitative analysis of the battery aging path. Furthermore, the CNN and PINN modules are combined, and the CNN model output is optimized based on the PNN module.

[0078] In one embodiment, the battery capacity degradation mode quantification method further includes, before inputting the two-dimensional data matrix into the deep learning model:

[0079] Construct a convolutional neural network model;

[0080] The training dataset is synthesized based on the open-circuit voltage model. The training dataset includes multiple capacity difference samples labeled with capacity degradation modes. The open-circuit voltage model simulates the battery voltage based on the positive and negative electrode potentials of the battery.

[0081] Train a convolutional neural network model based on the training dataset and loss function;

[0082] Deploy the trained convolutional neural network model into the convolutional neural network module of the deep learning model.

[0083] Optionally, before inputting the two-dimensional data matrix into the deep learning model, an open-circuit voltage model is first constructed based on the positive and negative electrode potentials of the battery. This model simulates the voltage response of the battery under different capacity degradation modes, generating capacity difference samples containing labels for multiple degradation modes as a training dataset. This eliminates the need for time-consuming degradation experiments required for training a CNN. Therefore, when training a deep learning model using synthetic data, the experimental cost and training time are extremely low, and the development time for diagnostic methods based on deep learning models is relatively short. Furthermore, the training dataset can contain almost all potential aging paths, representing over 1 billion possibilities. This allows a general diagnostic method based on deep learning models to quantify the degradation modes of aging batteries under various conditions. Next, a convolutional neural network model is trained using the training dataset to capture multi-scale local and global aging features in the capacity difference curve and output accurate capacity degradation modes. Finally, the trained convolutional neural network model is deployed into the convolutional neural network module of the deep learning model. Thus, by constructing a convolutional neural network model and combining it with a physically driven synthetic training dataset, the accuracy and robustness of the battery capacity degradation mode quantification method are significantly improved.

[0084] Step S103: Based on the difference between the charging curve shape of the first type of lithium battery and the charging curve shape of the second type of lithium battery, determine the deep learning model of the second type of lithium battery; wherein the deep learning model of the second type of lithium battery includes at least some components of the deep learning model of the first type of lithium battery.

[0085] Optionally, the deep learning model for the second type of lithium battery can be adjusted by analyzing the differences in charging curve shapes between the first and second types of lithium batteries, such as voltage slope and the duration of the constant voltage phase. The first and second types of lithium batteries can be NMC (LiNiCoMnO2, lithium nickel cobalt manganese oxide), NCA (LiNiCoAlO2, lithium nickel cobalt aluminum oxide), or LFP (LiFePO4, lithium iron phosphate), etc.

[0086] Optionally, the deep learning model for the first type of lithium battery includes convolutional layers, pooling layers, a first dense layer, a second dense layer, and a fully connected layer. Based on the similarity between the charging curves of the first type of lithium battery and the charging curves of the second type of lithium battery, some components of the deep learning model for the first type of lithium battery are transferred to the deep learning model for the second type of lithium battery. Transfer learning refers to the ability to transfer knowledge from a large model trained on a specific task to a new, similar task. For example, since the charging curves of the first type of lithium battery and the second type of lithium battery have a high similarity, the convolutional layers, pooling layers, first dense layer, and fully connected layer of the deep learning model for the first type of lithium battery can be retained in the deep learning model for the second type of lithium battery; only the second dense layer needs to be retrained and fine-tuned. For example, the charging curves of the first type of lithium battery and the second type of lithium battery have low similarity. Therefore, the pooling layer, first dense layer, and fully connected layer of the deep learning model for the first type of lithium battery can be retained in the deep learning model for the second type of lithium battery, while the convolutional layers and second dense layer in the deep learning model for the second type of lithium battery are retrained and fine-tuned. In this way, even with limited available data, transfer learning can be used to fine-tune the available data using knowledge generated by the pre-trained model, without needing to train the model from scratch. This improves model training efficiency and enables accurate quantitative analysis of capacity degradation patterns at the electrode level of batteries with different chemical systems.

[0087] Alternatively, please refer to Figure 3This paper demonstrates the training and fine-tuning process of deep learning models. First, the initial structure of the NMC deep learning model is built using the feature extraction and output mapping modules. Next, the NMC deep learning model is pre-trained based on the NMC dataset, with parameter optimization performed sequentially through convolutional layers and two dense layers. After the NMC deep learning model is trained, fine-tuning is performed on the NCA and LFP deep learning models respectively. In transfer learning, considering that the charging curve shapes of NCA and NMC chemistry are almost identical, the NCA deep learning model does not need to retrain the feature extraction part; it directly uses the parameters trained on the source domain dataset. That is, for NCA, only the second dense layer of the feature-output mapping part needs to be retrained. Conversely, the charging curve shapes of LFP and NMC chemistry differ significantly. Therefore, the LFP deep learning model needs to retrain the two layers that contact the input and mapping parts, namely the convolutional layer and the second dense layer; the first dense layer can be left untrained.

[0088] Alternatively, please refer to Figure 4 This document illustrates the specific process of the battery capacity degradation mode quantification method according to an embodiment of this application. First, a training dataset is synthesized by combining pseudo-OCV curves (pseudo-open-circuit voltage curves) with corresponding aging labels. This dataset is then used as input to train a deep learning model, which outputs the degradation mode quantification results of the lithium battery. The deep learning model includes a CNN module and a PINN module. The degradation mode quantification results of the lithium battery include four types of degradation modes: lithium-ion loss (LLI), positive electrode active material loss (LAMdePE), negative electrode active material loss (LAMdeNE), and internal resistance increase (RI). Based on the lithium battery degradation mode quantification results output by the deep learning model, the degradation mechanism of the lithium battery is analyzed, revealing degradation mechanisms such as SEI film thickening, lithium plating, electrode saturation, and increased resistance. Furthermore, after the deep learning model is trained on the NMC (LiNiCoMnO2, lithium nickel cobalt manganese oxide) battery dataset, the NMC battery dataset can be transferred to the deep learning model of NCA (LiNiCoAlO2, lithium nickel cobalt aluminum oxide) battery or LFP (LiFePO4, lithium iron phosphate) battery for pre-training. Then, the deep learning model can be fine-tuned by combining the NCA battery dataset or LFP battery dataset to improve the universality of the deep learning model of NMC lithium battery.

[0089] Figures 5 to 6 The quantitative results of four capacity degradation modes in NMC batteries are shown. Figure 5 The image shows the model prediction results for three capacity degradation modes in NMC batteries: lithium battery loss (LLI) (orange dots), negative electrode active material loss (LAMdeNE) (blue triangles), and positive electrode active material loss (LAMdePE) (brown rectangles). Figure 6The image shows the model prediction results for the capacity degradation mode RI (green dots) caused by increased internal resistance in NMC batteries. The horizontal axis represents the number of charge-discharge cycles of the lithium battery, and the vertical axis represents the quantified percentage of the capacity degradation mode. Figure 5 and Figure 6 It is known that as the number of charge-discharge cycles of a lithium battery increases, the overall degradation of each path at the electrode level of the lithium battery leads to degradation, but the degree of degradation varies among different paths. When the number of cycles reaches 180, the percentage of internal resistance degradation reaches over 80%, the loss of positive electrode active material is approximately 22%, the loss of negative electrode active material is approximately 17%, and the loss of the lithium battery itself is approximately 12%.

[0090] Figures 7 to 10 The graph shows a comparison between the actual and predicted values ​​of four capacity degradation modes in NMC batteries. Figure 7 This is a comparison chart of LLI loss in lithium batteries. Figure 8 This is a comparison chart showing the loss of LAMdeNE in the negative electrode active material. Figure 9 This is a comparison chart showing the loss of LAMdePE in the positive electrode active material. Figure 10 A comparison chart showing the effect of increasing RI with increasing internal resistance. For example... Figures 7 to 10 As shown, the horizontal axis represents the range of true values, the vertical axis represents the range of predicted values, the black line represents the ideal predicted value, and the colored scatter points represent the predicted values. Ideally, all colored scatter points should fall on the diagonal of y = x (i.e., the predicted value is completely consistent with the true value). The closer the colored scatter points are to the diagonal, the better the model performance. As can be seen from the figure, most of the colored scatter points are concentrated near the diagonal, indicating that the model prediction results are highly consistent with the true values. That is, the battery capacity degradation mode quantification method of this application embodiment can accurately quantify the degradation mode of NMC batteries.

[0091] Figure 11 and Figure 12 The applications of transfer learning in LFP and NCA batteries are illustrated respectively. Figure 11 The predicted results for three capacity degradation modes of LFP cells are as follows. Figure 12 This section presents the predicted capacity degradation modes for NCA batteries. The horizontal axis represents the number of charge-discharge cycles of the lithium battery, and the vertical axis represents the quantified percentage of each capacity degradation mode. The three capacity degradation modes are the model predictions for lithium battery loss (LLI) (orange dots), negative electrode active material loss (LAMdeNE) (blue triangles), and positive electrode active material loss (LAMdePE) (brown rectangles). Figure 11 It can be seen that with the increase of the number of charge-discharge cycles of LFP lithium batteries, the overall capacity degradation occurs across all pathways at the electrode level, but the degree of degradation varies among different pathways. When the number of cycles reaches 100, the negative electrode active material loses approximately 13%, the positive electrode active material loses approximately 8%, and the lithium battery itself loses approximately 7%, with the values ​​of the three capacity degradation modes decreasing sequentially. Figure 12It can be seen that as the number of charge-discharge cycles of NCA lithium batteries increases, the overall capacity degradation occurs across all pathways at the electrode level, but the degree of degradation varies among different pathways. When the number of cycles reaches 100, the positive electrode active material loses approximately 15%, the negative electrode active material loses approximately 13%, and the lithium battery loses approximately 7%, with the values ​​of the three capacity degradation modes decreasing sequentially.

[0092] The battery capacity degradation mode quantification method of this application embodiment obtains a first pseudo-open circuit voltage curve and a second pseudo-open circuit voltage curve of a first type of lithium battery, and constructs a two-dimensional data matrix based on the difference between the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve; wherein, the first pseudo-open circuit voltage curve corresponds to an aged first type of lithium battery, and the second pseudo-open circuit voltage curve corresponds to a fresh first type of lithium battery; the two-dimensional data matrix is ​​input into a deep learning model to output the capacity degradation mode of the first type of lithium battery, which is the capacity degradation path of the first type of lithium battery at the electrode level; according to the difference between the charging curve shape of the first type of lithium battery and the charging curve shape of the second type of lithium battery, a deep learning model for the second type of lithium battery is determined; wherein, the deep learning model for the second type of lithium battery includes at least some components of the deep learning model for the first type of lithium battery. This application constructs a two-dimensional data matrix by using the pseudo-open circuit voltage curves of aged lithium batteries and fresh lithium batteries, inputs it into a deep learning model to output the degradation path of lithium batteries at the electrode level, and combines a transfer learning strategy to optimize the deep learning model components for cross-type lithium batteries, thereby achieving accurate quantitative analysis of the capacity degradation mode at the electrode level of batteries with different chemical systems.

[0093] Please see Figure 13 This application also provides a battery capacity degradation mode quantification device 700, which includes a data processing module 701, a deep learning module 702 and a transfer learning module 703.

[0094] The data processing module 701 is used to obtain the first pseudo open circuit voltage curve and the second pseudo open circuit voltage curve of the first type of lithium battery, and construct a two-dimensional data matrix based on the difference between the first pseudo open circuit voltage curve and the second pseudo open circuit voltage curve; wherein the first pseudo open circuit voltage curve corresponds to the aged first type of lithium battery, and the second pseudo open circuit voltage curve corresponds to the fresh first type of lithium battery.

[0095] The deep learning module 702 is used to input a two-dimensional data matrix into a deep learning model to output the capacity degradation mode of the first type of lithium battery. The capacity degradation mode is the capacity degradation path of the first type of lithium battery at the electrode level.

[0096] The transfer learning module 703 is used to determine a deep learning model for the second type of lithium battery based on the difference between the charging curve shape of the first type of lithium battery and the charging curve shape of the second type of lithium battery; wherein the deep learning model for the second type of lithium battery includes at least some components of the deep learning model for the first type of lithium battery.

[0097] In one embodiment, a two-dimensional data matrix is ​​constructed based on the first pseudo-open-circuit voltage curve and the second pseudo-open-circuit voltage curve, including:

[0098] The first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve are interpolated and aligned within the same voltage range.

[0099] Calculate the capacity difference curve between the first pseudo-open circuit voltage curve and the second pseudo-open circuit voltage curve at each voltage sampling point;

[0100] The capacity difference curve is transformed into a two-dimensional data matrix, where the rows of the two-dimensional data matrix correspond to voltage sampling points and the columns of the two-dimensional data matrix correspond to capacity differences.

[0101] In one embodiment, the capacity degradation mode includes at least one of the following: lithium-ion loss mode, positive electrode active material loss mode, negative electrode active material loss mode, and internal resistance increase mode.

[0102] In one embodiment, the deep learning model includes a convolutional neural network module and a physical information neural network module. The convolutional neural network module is used to output the capacity degradation mode of the first type of lithium battery, and the physical information neural network module is used to adjust the convolutional neural network module according to the capacity degradation mode.

[0103] In one embodiment, the convolutional neural network module includes convolutional layers, pooling layers, dense layers, and fully connected layers. A two-dimensional data matrix is ​​input into a deep learning model to output the capacity degradation mode of a first type of lithium battery, including:

[0104] Input the two-dimensional data matrix into the convolutional neural network module;

[0105] Multi-scale local aging features of the first type of lithium battery are extracted from a two-dimensional data matrix based on convolutional layers.

[0106] The global aging characteristics of the first type of lithium battery are determined by combining multi-scale local features through the pooling layer;

[0107] By fusing multi-scale local aging features and global aging features in a dense layer, a fused aging feature is obtained.

[0108] The fully connected layer maps the fusion aging characteristics to the capacity degradation mode of the first type of lithium battery.

[0109] In one embodiment, inputting a two-dimensional data matrix into a deep learning model to output the capacity degradation mode of a first type of lithium battery further includes:

[0110] The two-dimensional data matrix is ​​input into the physical information neural network module to output a capacity degradation reference pattern for the first type of lithium battery.

[0111] If the capacity decay reference pattern is inconsistent with the capacity decay pattern or there is a large deviation, then the convolutional neural network module should be adjusted.

[0112] In one embodiment, before inputting the two-dimensional data matrix into the deep learning model, the deep learning module 703 is further configured to include:

[0113] Construct a convolutional neural network model;

[0114] The training dataset is synthesized based on the open-circuit voltage model. The training dataset includes multiple capacity difference samples labeled with capacity degradation modes. The open-circuit voltage model simulates the battery voltage based on the positive and negative electrode potentials of the battery.

[0115] Train a convolutional neural network model based on the training dataset and loss function;

[0116] Deploy the trained convolutional neural network model into the convolutional neural network module of the deep learning model.

[0117] The battery capacity degradation mode quantification device provided in this application embodiment can implement all the steps of the above-described battery capacity degradation mode quantification method embodiment and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0118] This application also provides an electronic device, including a processor and a memory. The memory stores a program or instructions that can run on the processor. When the program or instructions are executed by the processor, they implement the various steps of the above-described battery capacity degradation mode quantification method embodiment and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0119] Figure 14 To illustrate the hardware structure of the electronic device according to the embodiments of this application, the electronic device includes:

[0120] The processor 801 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0121] The memory 802 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 802 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 802 and is called and executed by the processor 801 using the battery capacity degradation mode quantification method of the embodiments of this application.

[0122] The 803 input / output interface is used to implement information input and output.

[0123] The communication interface 804 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0124] Bus 805 transmits information between various components of the device (e.g., processor 801, memory 802, input / output interface 803, and communication interface 804);

[0125] The processor 801, memory 802, input / output interface 803, and communication interface 804 are connected to each other within the device via bus 805.

[0126] The electronic device provided in this application embodiment can implement all the steps of the above-described battery capacity degradation mode quantification method embodiment and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0127] This application also provides a computer-readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various steps of the above-described battery capacity degradation mode quantification method embodiment and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0128] The processor is the processor in the electronic device described in the above embodiments. The computer-readable storage medium includes computer-readable storage media such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0129] This application also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various steps of the above-described battery capacity degradation mode quantification method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0130] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0131] This application provides a computer program product stored in a storage medium. The program product is executed by at least one processor to implement the various steps of the battery capacity degradation mode quantification method embodiment described above, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0132] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not delete other identical elements present in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0133] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0134] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method of quantifying a battery capacity fade pattern, the method comprising: determining a battery capacity fade pattern for a battery; and quantifying the battery capacity fade pattern. The method comprises: obtaining a first pseudo open-circuit voltage curve and a second pseudo open-circuit voltage curve of a first type of lithium battery, and constructing a two-dimensional data matrix based on the difference between the first pseudo open-circuit voltage curve and the second pseudo open-circuit voltage curve; wherein the first pseudo open-circuit voltage curve corresponds to the first type of lithium battery in an aged state, and the second pseudo open-circuit voltage curve corresponds to the first type of lithium battery in a fresh state; inputting the two-dimensional data matrix into a deep learning model to output a capacity degradation mode of the first type of lithium battery, wherein the capacity degradation mode is a capacity degradation path of the first type of lithium battery at the electrode level; determining a deep learning model of a second type of lithium battery according to the difference between the charging curve shape of the first type of lithium battery and the charging curve shape of the second type of lithium battery; wherein the deep learning model of the second type of lithium battery at least includes part of the components of the deep learning model of the first type of lithium battery.

2. The battery capacity fade pattern quantification method of claim 1, wherein, The method further comprises: interpolating and aligning the first pseudo open-circuit voltage curve and the second pseudo open-circuit voltage curve in the same voltage interval; calculating a capacity difference curve of the first pseudo open-circuit voltage curve and the second pseudo open-circuit voltage curve at each voltage sampling point; converting the capacity difference curve into a two-dimensional data matrix, wherein the rows of the two-dimensional data matrix correspond to the voltage sampling points, and the columns of the two-dimensional data matrix correspond to the capacity difference. 3.The battery capacity degradation pattern quantification method of claim 1, wherein, The capacity degradation mode comprises at least one of a lithium ion loss mode, a positive active material loss mode, a negative active material loss mode, and an internal resistance increase mode. 4.The battery capacity degradation pattern quantification method of claim 1, wherein, The deep learning model comprises a convolutional neural network module and a physical information neural network module, wherein the convolutional neural network module is used to output the capacity degradation mode of the first type of lithium battery, and the physical information neural network module is used to adjust the convolutional neural network module according to the capacity degradation mode.

5. The battery capacity fade pattern quantification method of claim 4, wherein, The convolutional neural network module comprises a convolutional layer, a pooling layer, a dense layer, and a fully connected layer, and the method further comprises: inputting the two-dimensional data matrix into the convolutional neural network module; extracting multi-scale local aging features of the first type of lithium battery from the two-dimensional data matrix based on the convolutional layer; determining global aging features of the first type of lithium battery by combining the multi-scale local features through the pooling layer; fusing the multi-scale local aging features and the global aging features through the dense layer to obtain fused aging features; mapping the fused aging features to the capacity degradation mode of the first type of lithium battery based on the fully connected layer.

6. The battery capacity fade pattern quantification method of claim 4, wherein, The method further comprises: inputting the two-dimensional data matrix into the physical information neural network module to output a capacity degradation reference mode of the first type of lithium battery. If the capacity degradation reference mode is inconsistent with the capacity degradation mode or has a large deviation, the convolutional neural network module is adjusted.

7. The battery capacity fade pattern quantification method of claim 1, wherein, Before the two-dimensional data matrix is input into the deep learning model, the battery capacity degradation mode quantification method further includes: a convolutional neural network model is constructed; a training data set is synthesized based on an open circuit voltage model, the training data set including a plurality of capacity difference samples with capacity degradation mode labels, and the open circuit voltage model simulating battery voltage based on positive and negative electrode potentials of the battery; the convolutional neural network model is trained based on the training data set and a loss function; the trained convolutional neural network model is deployed to a convolutional neural network module of the deep learning model.

8. A battery capacity fade pattern quantization apparatus, characterized by, The method includes a data processing module, a deep learning module, and a transfer learning module. The data processing module is configured to obtain a first pseudo open circuit voltage curve and a second pseudo open circuit voltage curve of a first type of lithium battery, and construct a two-dimensional data matrix based on a difference between the first pseudo open circuit voltage curve and the second pseudo open circuit voltage curve, wherein the first pseudo open circuit voltage curve corresponds to an aged first type of lithium battery, and the second pseudo open circuit voltage curve corresponds to a fresh first type of lithium battery. The deep learning module is configured to input the two-dimensional data matrix into a deep learning model to output a capacity degradation mode of the first type of lithium battery, wherein the capacity degradation mode is a capacity degradation path of the first type of lithium battery at an electrode level. The transfer learning module is configured to determine a deep learning model of a second type of lithium battery according to a difference between a charging curve shape of the first type of lithium battery and a charging curve shape of the second type of lithium battery, wherein the deep learning model of the second type of lithium battery at least includes part of components of the deep learning model of the first type of lithium battery.

9. An electronic device, comprising: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the battery capacity degradation mode quantification method when executing the computer program.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the battery capacity degradation mode quantification method.

Images