Retired power battery sorting method, equipment, medium and product

Abstract

The invention discloses an out-of-service power battery sorting method and device, a medium and a product, and relates to the field of out-of-service power battery sorting. The method comprises the steps that charging voltage data of out-of-service power batteries to be sorted in a constant-current charging stage are extracted; preprocessing the charging voltage data; the preprocessing comprises data dimension reduction and normalization processing; according to the pre-processed charging voltage data, obtaining a voltage data image by adopting a Gramb angle difference field method; according to the voltage data image, adopting a retired power battery sorting model to obtain a sorting result; the retired power battery sorting model is constructed based on an offset window converter model. According to the invention, the efficiency and accuracy of large-scale retired battery sorting can be improved.

Description

Technical Field

[0001] The present application relates to the field of sorting of retired power batteries, and particularly to a method, device, medium and product for sorting retired power batteries. Background Art

[0002] With the rapid growth in the number of electric vehicles, how to properly address energy and environmental issues has become a major pain point in today's society. By 2030, the share of electric vehicles in global sales is expected to reach 22%. As one of the key components of electric vehicles, power batteries will no longer meet the requirements of new energy vehicles due to a capacity drop to 80% and will face a retirement wave. It is estimated that by 2025, the quantity of retired power batteries in China will reach 820,000 tons. The retired power batteries of electric vehicles still have a high residual value, and their direct scrapping will result in huge waste of resources.

[0003] That is, although retired batteries cannot meet the requirements of electric vehicles in terms of endurance due to capacity decline, they still meet the requirements of energy storage systems with lower power requirements, low-speed electric vehicles, etc. Recycling with a second life can improve the utilization rate of power batteries and reduce the cost of popularizing electric vehicles. Recycling with a second life is an important means for recycling retired batteries and improving the utilization rate of power batteries. Sorting and recombining retired batteries according to parameters such as capacity and internal resistance for recycling with a second life in energy storage systems with lower power requirements, low-speed vehicles, etc. When the battery performance further deteriorates, the retired batteries are then disassembled and recycled. How to achieve rapid sorting of retired batteries to reduce the economic cost and time cost of recycling with a second life is the key to recycling retired batteries with a second life.

[0004] For large-scale retired batteries, their performance parameters such as capacity and internal resistance at the time of retirement are highly inconsistent and cannot be directly recombined for use. The inconsistency of battery packs is an inevitable problem during the use of power batteries, and the inconsistency problem of retired batteries is more serious. In order to improve the utilization rate of battery capacity and avoid potential safety hazards caused by inconsistency, rapid sorting of retired batteries before recycling with a second life is an important link that cannot be avoided. In order to improve the energy efficiency of retired batteries in recycling with a second life, avoid performance degradation and safety accidents that may be caused by parameter dispersion problems, and extend the cycle life of the system, it is necessary to sort large-scale retired batteries and re-group the batteries with good consistency according to the sorting results for recycling with a second life.

[0005] The existing sorting methods for retired power batteries mainly fall into three categories: direct measurement methods, model-based methods, and data-driven methods. The direct measurement method refers to sorting using the battery measurement data directly obtained during the battery test, and directly using various factors such as capacity, internal resistance, and pulse voltage to evaluate the consistency of retired batteries. Although the accuracy is very high, the efficiency is low and it is time-consuming; the model-based method aims to predict the health state of the battery through a physical model. In the electrochemical model, partial differential equations are used to describe the thermodynamic and kinetic processes inside the battery, and then the health indicators are achieved by extrapolating the model parameters. The model-based method requires higher model accuracy and the implementation process is more complex; the data-driven method uses the battery data information to obtain the implicit information inside the battery, such as using the incremental capacity curve to estimate the remaining capacity of the old battery, but the feature engineering is complex and the generality is insufficient. The current methods all have the problems of low efficiency and long time consumption, and cannot meet the rapid sorting requirements of large-scale retired batteries in the actual cascade utilization process.

[0006] Existing patents have also proposed various different methods to meet the rapid sorting requirements of large-scale retired batteries during the actual cascade utilization process. For example, Scheme (1): The patent with the publication number CN111420898A discloses a method for sorting retired batteries, which includes: S1. Disassembling the retired battery pack and screening out qualified single batteries; S2. Obtaining the charge and discharge data of each of the qualified single batteries; S3. Obtaining a charge and discharge curve through the charge and discharge data and calculating the dynamic bending distance between the charge and discharge curve and a fixed reference curve; S4. Classifying each of the qualified single batteries into different groups based on the dynamic bending distance. Scheme (1) solves the problem that consistency cannot be taken into account in the prior art. Scheme (2): The patent with the publication number CN112808629A discloses a method, device, storage medium and electronic device for sorting retired batteries. The method for sorting retired batteries includes: Detecting and recording the DC resistance of the battery module; Calculating the average DC resistance of the battery module; Determining a judgment criterion based on the average DC resistance; Screening the battery module according to the judgment criterion. The method for sorting retired batteries in Scheme (2) improves the consistency of the battery module by screening the battery module based on the average DC resistance. Scheme (3): The patent with the publication number CN117654928A discloses a method, device and equipment for sorting retired batteries based on time distance sequence. The method includes: Performing a charge and discharge pulse test on multiple retired batteries to be sorted for a first preset time to obtain multiple first voltage curves, and performing a charge and discharge pulse test on new batteries of the same model for a second preset time to obtain a second voltage curve; Calculating the time distance sequence within the first preset time based on the first voltage curve and the second voltage curve; Predicting the voltage curves within a third preset time of multiple retired batteries to be sorted based on the time distance sequence to obtain multiple third voltage curves within the third preset time; Calculating the average Euclidean deviation based on the third voltage curve and sorting the retired batteries through the average Euclidean deviation. Scheme (3) predicts the voltage curve of the retired battery through pulse test and sorts the retired battery according to the predicted voltage curve, improving the sorting efficiency of the battery.

[0007] Regarding Scheme (1), the charge and discharge curve is obtained by using the charge and discharge data of retired batteries, and the dynamic bending distance between the charge and discharge curve and the fixed reference curve is calculated; and based on the dynamic bending distance, each of the qualified single batteries is sorted into different groups to adapt to different cascade usage scenarios. However, this method relies on the charge and discharge data of the battery, and the charge and discharge data requires charge and discharge tests on retired batteries, which takes a long time and has low efficiency, and cannot be applied in the actual large-scale retired battery sorting scenario.

[0008] For solution (2), the DC resistance of the battery module is detected and recorded to calculate the average DC resistance of the battery module, then the judgment criterion is determined based on the average DC resistance, and finally the battery module is screened according to the judgment criterion. Since the DC internal resistance detection may be affected by various factors, including temperature, state, etc., and this method requires corresponding detection equipment and technical support, including precise resistance measurement instruments, data processing software, etc., which will increase the sorting cost and complexity of retired batteries.

[0009] For solution (3), the voltage curve of the battery is obtained by using the pulse test method, and then the retired batteries are sorted by calculating the average Euclidean deviation of different curves. Although this method can improve the sorting efficiency by predicting the voltage curve of the battery through pulse testing, implementing this method requires precise testing equipment and advanced data processing technical support, including pulse testing equipment, etc., and the extracted features need to be processed, which increases the economic and time costs of sorting and reduces the applicability of the method.

[0010] In summary, in the face of the growing trend of retired battery recycling, the existing methods all face the problems that the detection time of the retired battery sorting and consistency evaluation methods is long, the efficiency is low, or the indicators are single and the effect is poor, which increases the cost of cascade utilization and also restricts the development of the electric vehicle industry, further resulting in low universality of the retired power battery sorting method. In the future, with the advent of the retired battery wave and the increase in the number of retired batteries, there is an urgent need to provide an efficient and time-saving retired power battery sorting method to efficiently and correctly complete the sorting work of retired batteries, thereby improving the utilization rate of environmental resources. Summary of the Invention

[0011] The purpose of this application is to provide a method, device, medium and product for sorting retired power batteries, which can improve the efficiency and accuracy of sorting large-scale retired batteries.

[0012] To achieve the above purpose, this application provides the following solutions:

[0013] In the first aspect, this application provides a method for sorting retired power batteries, and the method for sorting retired power batteries includes:

[0014] Extract the charging voltage data of the retired power battery to be sorted during the constant current charging stage;

[0015] Preprocess the charging voltage data; the preprocessing includes: data dimensionality reduction and normalization processing;

[0016] According to the preprocessed charging voltage data, use the Gram angular difference field method to obtain the voltage data image;

[0017] According to the voltage data image, a sorting result is obtained by using a sorting model for retired power batteries; the sorting model for retired power batteries is constructed based on a shifted window transformer model.

[0018] Optionally, after extracting the charging voltage data of the retired power battery to be sorted in the constant current charging stage, the following further includes:

[0019] Retain the charging voltage data greater than the voltage threshold.

[0020] Optionally, the voltage threshold is 3.9V.

[0021] Optionally, the preprocessing of the charging voltage data specifically includes:

[0022] Data dimensionality reduction is performed on the charging voltage data by using the method of approximate segment aggregation; the method of approximate segment aggregation is to divide the charging voltage data into multiple segments according to set rules;

[0023] Normalize the charging voltage data within each segment.

[0024] Optionally, before obtaining the sorting result by using the sorting model for retired power batteries according to the voltage data image, the following further includes:

[0025] Construct a sorting model for retired power batteries based on a shifted window transformer model;

[0026] Obtain retired power battery data from the battery public dataset;

[0027] Construct a sorting dataset for retired power batteries according to the retired power battery data and the corresponding retired power battery capacity;

[0028] Use the sorting dataset for retired power batteries to train the sorting model for retired power batteries.

[0029] Optionally, the shifted window transformer model adopts a window mechanism and an attention mechanism.

[0030] In a second aspect, the present application provides a sorting device for retired power batteries, and the sorting device for retired power batteries includes:

[0031] A charging voltage data extraction module, configured to extract the charging voltage data of the retired power battery to be sorted in the constant current charging stage;

[0032] A preprocessing module, configured to preprocess the charging voltage data; the preprocessing includes: data dimensionality reduction and normalization processing;

[0033] A voltage data image determination module, configured to obtain a voltage data image by using the Gram angular difference field method according to the preprocessed charging voltage data;

[0034] A sorting result determination module, configured to obtain a sorting result according to a voltage data image by using a sorting model for retired power batteries, where the sorting model for retired power batteries is constructed based on a shifted window transformer model.

[0035] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the sorting method for retired power batteries.

[0036] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the sorting method for retired power batteries is implemented.

[0037] In a fifth aspect, the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the sorting method for retired power batteries is implemented.

[0038] According to the specific embodiments provided by the present application, the present application has the following technical effects:

[0039] The present application provides a sorting method, device, medium, and product for retired power batteries. The Gramian Angular Difference Fields (GADF) method is used to convert the preprocessed charging voltage data into a voltage data image. GADF converts time series data into a two-dimensional image without losing the original data information and extracts useful feature information from it. And a sorting model for retired power batteries constructed based on a shifted window transformer model (Swin-Transformer) is used for sorting. The present application significantly shortens the time required for battery sorting and improves the sorting efficiency on the premise of ensuring the accuracy of the sorting result, so that retired batteries can be quickly applied to actual different echelon scenarios. Description of the Drawings

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0041] Figure 1 It is a schematic flowchart of a sorting method for retired power batteries in an embodiment of the present application;

[0042] Figure 2 Schematic diagram of the principle of a method for sorting retired power batteries in an embodiment of the present application;

[0043] Figure 3 Schematic diagram of the GADF conversion process;

[0044] Figure 4 Schematic diagram of the Swin Transformer model structure;

[0045] Figure 5 Schematic diagram of the comparison result with the prior art. Specific implementation manners

[0046] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0047] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0048] In an exemplary embodiment, as Figure 1 and Figure 2 shown, a method for sorting retired power batteries is provided, and the method includes the following S101 to S104. Among them:

[0049] S101, extracting the charging voltage data of the retired power battery to be sorted in the constant current charging stage;

[0050] To improve the sorting efficiency and retain key information, the charging voltage data greater than the voltage threshold is retained; the voltage threshold is 3.9V.

[0051] S102, preprocessing the charging voltage data; the preprocessing includes: data dimensionality reduction and normalization processing;

[0052] S102 specifically includes:

[0053] S21, performing data dimensionality reduction on the charging voltage data by using the method of approximate segment aggregation; the method of approximate segment aggregation is to divide the charging voltage data into multiple segments according to the set rules; the main function of data dimensionality reduction is to reduce the complexity of the data while retaining key feature information. By data dimensionality reduction, the training efficiency of the retired power battery sorting model is improved, the calculation cost is reduced, and the overfitting risk is reduced.

[0054] S22, normalizing the charging voltage data within each segment.

[0055] Approximate segment aggregation performs approximate segment aggregation data dimensionality reduction on the constant current charging segment data above 3.9V, and divides the voltage time series into equal-length line segments. Then, the average value of the segment is used to represent the values of all voltage data in that segment. The calculation formula is as follows:

[0056]

[0057] In the formula, n is the length of the time series, d with a downward subscript is the length after data dimensionality reduction, and x i is the average value after annotation;

[0058] S103. According to the preprocessed charging voltage data, use the Gram Angle Difference Field method to obtain the voltage data image;

[0059] The Gram Angle Difference Field method is a feature extraction method for time series data analysis. It is usually used to process multi-dimensional time series data. The basic idea of GADF is to convert time series data into two-dimensional images, and then extract features from these images to reveal the patterns and structures in the data. Given a data set X containing k voltage sample points, each voltage sample point can be represented as a d-dimensional vector, that is, X = [X1, X2,..., X n , where X i ∈R d . The polar coordinates of the voltage segment are The Gram matrix G is defined as follows:

[0060]

[0061] In the formula, I is the unit vector, and X and X are the vectors before and after sequence scaling. is the angle between two phasors. The voltage segment time series at each time step is regarded as a dimensional space, and the time correlation matrix G is calculated by defining the inner product.

[0062] The GADF method converts time series data into two-dimensional images without losing the original data information, and extracts useful feature information from them for subsequent data training, classification, prediction and other tasks. Strongly non-linear data can enhance the sparsity of the data processed by GADF, eliminate multi-mode redundant information, weaken the non-linearity of the data, and reduce noise. This method has certain advantages in processing multi-dimensional time series data and can effectively reveal the complex structures and dynamic change laws in the data. Figure 3 The process of converting the voltage timing signal into a GADF image is given.

[0063] S104. Based on the voltage data image, use the retired power battery sorting model to obtain the sorting result; the retired power battery sorting model is constructed based on the shifted window transformer model.

[0064] Swin Transformer is an image classification model based on the Transformer structure. It first divides the input image into a series of image patches of fixed size and converts each image patch into a vector representation. Then multiple Transformer encoder layers are stacked together, and each encoder layer contains a multi-head self-attention mechanism and a feed-forward neural network. The advantage of the Swin Transformer model lies in its efficient modeling ability for large-size images. Compared with traditional convolutional neural networks, Swin Transformer has better scalability and generalization ability, especially suitable for scenarios that need to process high-resolution images. In image classification tasks, this model has achieved satisfactory performance. Generally speaking, Swin Transformer is an innovative image classification model. By introducing the window mechanism and the attention mechanism, it realizes the efficient modeling and processing of large-size images. The design of this model enables effective capture of global context information and performs outstandingly in image classification tasks, with a very high accuracy rate. By adjusting the hyperparameters of the model, such as the number of iterations, time window and other parameters, the predicted accuracy rate of the final sorting result is as high as 97.67%.

[0065] The Swin Transformer model can integrate information between different windows through shifted windows, so that the generated image contains more comprehensive and real information.

[0066] The Swin Transformer model can be divided into four structures: Swin-T, Swin-S, Swin-B and Swin-L according to the number of stacking times and input depth in different stages of image processing. Based on the selected small sample data set, Swin-T structure is selected for classification in this application. The network structure of Swin-T is as Figure 4 shown in part a. The input GADF voltage data image is divided into non-overlapping patches and patch splitting modules of the same size, that is, every 4×4 adjacent pixels correspond to a patch. The linear embedding layer projects the tensor of this dimension onto any dimension C, that is, a linear transformation of the channel data of each pixel. Four stages construct feature maps of different sizes. Only stage 1 first passes through the linear embedding layer, while the next three stages first perform downsampling through the patch merging layer, and finally pass through the pooling layer and the fully connected layer to output the battery classification result.

[0067] The Swin Transformer block layer has two structures, as Figure 4As shown in part b). The difference between these two structures is that one uses the window multi-head self-attention (W-MSA) structure and the other uses the shifted window multi-head self-attention (SW-MSA) structure. Therefore, each stage is used in pairs. This application introduces the W-MSA module to reduce the computational complexity. Compared with the traditional MSA module, the W-MSA used in the Swin transformer can divide the feature map into separate windows of different sizes and perform the self-attention mechanism calculation separately within each window. Its calculation formula is as follows:

[0068]

[0069] where Q, K, and V are input matrices, representing the query matrix, key matrix, and value matrix respectively, and d k is the vector dimension. The role of the Attention formula is to weight the similarity between Q and K and use it to obtain the weighted sum of V corresponding to the input.

[0070] Figure 4 Part c) shows the SW-MSA structure for calculating self-attention shifted windows in the Swin transformer structure. The first layer (left) uses the traditional window partitioning scheme to calculate self-attention within each window. In the next layer l+1 (right), the window partition is shifted to generate new windows. In the new windows, the calculation of self-attention crosses the boundaries of the previous windows, establishing connections between them. The SW-MSA improves the W-MSA and enhances its ability to transfer information between windows. For example, several windows in the middle of the layer graph can exchange information between multiple windows in the layer, significantly improving the modeling ability and computational efficiency. To extend the generality of the Swin transform in image classification, this application applies the Swin transform to the image classification of GADF images converted from the original voltage distribution.

[0071] To verify the effectiveness of the GADF data imaging method and the Swin transformer model, we respectively conducted comparative analyses with other data imaging methods and clustering algorithms, such as Markov transformation (MTF), Naive Bayes Classifier, K-Nearest Neighbor algorithm (KNN), and Back Propagation Neural Network (BPNN), as Figure 5 shown, and the results show that the sorting method proposed in this application has a better sorting effect compared with the current mainstream methods. At the same time, since the data used in this application only includes voltage data above 3.9V in the constant current charging segment, the data requirement is very small, it is easy to obtain, and the sorting of retired batteries can be quickly completed, greatly improving the sorting efficiency and being applicable to the actual use scenario of large-scale retired battery sorting.

[0072] This application can complete the rapid sorting of 300 retired batteries, with an accuracy rate of over 97%. The accuracy rate, F1 score, recall rate, and precision rate of the sorting results by other patented methods are all inferior to those of the method provided by this application. Moreover, the sorting time of other methods is too long, which cannot meet the requirements of rapid and accurate sorting of a large number of retired batteries.

[0073] In an exemplary embodiment, as Figure 2 shown, this application obtains retired power battery data from the battery public dataset; constructs a retired power battery sorting dataset based on the retired power battery data and the corresponding retired power battery capacity; and uses the retired power battery sorting dataset to train the retired power battery sorting model.

[0074] Based on the same inventive concept, the embodiment of this application also provides a retired power battery sorting device for implementing the above-mentioned retired power battery sorting method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more of the following embodiments of the retired power battery sorting device can refer to the limitations on the retired power battery sorting method in the above text, and will not be elaborated here.

[0075] In an exemplary embodiment, a retired power battery sorting device is provided, including:

[0076] A charging voltage data extraction module, configured to extract the charging voltage data of the retired power battery to be sorted during the constant current charging stage;

[0077] A preprocessing module, configured to preprocess the charging voltage data; the preprocessing includes: data dimensionality reduction and normalization processing;

[0078] A voltage data image determination module, configured to obtain a voltage data image by using the Gram angular difference field method according to the preprocessed charging voltage data;

[0079] The sorting result determination module is used to obtain a sorting result by using a retired power battery sorting model based on the voltage data image; the retired power battery sorting model is constructed based on the shifted window transformer model. In an exemplary embodiment, a computer device is provided, which can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. The computer program, when executed by the processor, implements a method for sorting retired power batteries.

[0080] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which, when executed by a processor, implements the steps in the above method embodiments.

[0081] In an exemplary embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps in the above method embodiments.

[0082] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0083] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memories (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0084] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0085] In the present application, all actions of obtaining signals, information, or data are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where the location is located and obtaining the authorization given by the owner of the corresponding device.

[0086] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0087] In this article, specific examples are used to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A method for sorting retired power batteries, characterized in that: The retired power battery sorting method comprises: Extract charging voltage data of the retired power batteries to be sorted during the constant current charging stage; Preprocessing the charging voltage data; the preprocessing includes: data dimension reduction and normalization processing; According to the pre-processed charging voltage data, the voltage data image is obtained by using the Gram angle difference field method; According to the voltage data image, a retired power battery sorting model is used to obtain a sorting result; the retired power battery sorting model is constructed based on an offset window converter model.

2. The method for sorting retired power batteries according to claim 1, characterized in that: The step of extracting charging voltage data of the retired power battery to be sorted during the constant current charging stage further includes: The charging voltage data that is greater than the voltage threshold is retained.

3. The method for sorting retired power batteries according to claim 2, characterized in that: The voltage threshold is 3.9V.

4. The method for sorting retired power batteries according to claim 1, characterized in that: The preprocessing of the charging voltage data specifically includes: The charging voltage data is subjected to data dimension reduction by using an approximate segment aggregation method; the approximate segment aggregation method is to divide the charging voltage data into multiple segments according to a set rule; The charging voltage data in each segment is normalized.

5. The method for sorting retired power batteries according to claim 1, characterized in that: The step of using a retired power battery sorting model to obtain a sorting result based on the voltage data image also includes: Construct a retired power battery sorting model based on the offset window converter model; Obtain retired power battery data from public battery datasets; Construct a retired power battery sorting data set based on retired power battery data and corresponding retired power battery capacity; The retired power battery sorting model is trained using the retired power battery sorting dataset.

6. The method for sorting retired power batteries according to claim 1, characterized in that: The offset window transformer model adopts a window mechanism and an attention mechanism.

7. A retired power battery sorting device, characterized in that: The retired power battery sorting equipment comprises: A charging voltage data extraction module is used to extract the charging voltage data of the retired power batteries to be sorted during the constant current charging stage; A preprocessing module, used for preprocessing the charging voltage data; the preprocessing includes: data dimension reduction and normalization processing; A voltage data image determination module is used to obtain a voltage data image by using a Gram angle difference field method according to the pre-processed charging voltage data; The sorting result determination module is used to obtain the sorting result according to the voltage data image by adopting the retired power battery sorting model; the retired power battery sorting model is constructed based on the offset window converter model.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for sorting retired power batteries according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the retired power battery sorting method described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the retired power battery sorting method described in any one of claims 1 to 6 is implemented.

Images