A method and device for gradient screening of retired power batteries and a readable medium

Abstract

The application discloses a kind of retired power battery gradient screening method, device and readable medium, by obtaining the charging voltage data of retired power battery, by piecewise aggregation approximation algorithm, the charging voltage data is processed, and dimension reduction data is obtained;Gram angle field conversion is carried out to dimension reduction data, and polar coordinate data is obtained;Polar coordinate data is converted into coded image using gram angle field;The coded image is input into the battery classification model trained, and the battery capacity category is obtained, and the battery classification model uses ConvNeXt model.The method can retain feature information while reducing the amount of calculation by dimension reduction and normalization processing of charging voltage data.And the time series data after dimension reduction is converted into GADF image using gram angle field, and then the GADF image is classified and screened using ConvNeXt model, with high screening accuracy and good accuracy, and certain applicability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of battery screening, in particular to a retired power battery gradient screening method and device and readable medium. BACKGROUND

[0002] The global environment is in a new period of major changes in energy structure, promoting energy greenization and reducing environmental pollution caused by energy consumption has become an important research topic. In recent years, new energy vehicles have high energy density and low self-discharge advantages, driven by various policies, complete industrial support, stable supply and price of upstream raw materials, and their sales have shown a high growth trend in countries around the world. However, the "climbing" growth of new energy vehicle sales has brought a series of subsequent problems. When new energy vehicles are at the end of their life cycle, the safe and effective recycling of sorted batteries becomes a top priority.

[0003] Due to the need for complete supporting facilities for battery decomposition and other operations, improper recycling and disposal of mainstream batteries such as ternary lithium batteries and lithium iron phosphate batteries can cause long-term serious harm to the environment. Generally, the screening of waste batteries uses battery capacity as the main basis, and the screening methods can be divided into three types:

[0004] (1) Obtain the capacity parameter by fully charging and discharging the battery. Although this method can obtain relatively accurate data, it is very inefficient and time-consuming.

[0005] (2) Calculate the battery health status based on the physical model of the battery health status algorithm, for example, by establishing an equivalent circuit model to predict the battery health status. However, this method often has high requirements for model accuracy and the implementation process is complex.

[0006] (3) Establish a data model through machine learning to realize the gradient screening of the battery, for example, use SVM to screen based on the battery discharge curve, but the efficiency is low. Or by optimizing the important features of the charging curve, using traditional clustering algorithms to quickly screen batteries, but traditional clustering algorithms have some defects. For example, when the randomly determined cluster center and the true cluster center differ greatly, it will cause the class center error to be large, and then affect the subsequent classification results. The existing data-driven method needs a large amount of battery data to repeatedly train the model, so it is not suitable for small sample battery data sets. SUMMARY

[0007] In view of the above technical problems, the purpose of the embodiments of the present application is to propose a retired power battery gradient screening method, device and readable medium to solve the technical problems mentioned in the background section.

[0008] In a first aspect, the present application provides a gradient screening method for retired power batteries, comprising the following steps:

[0009] S1, obtaining the charging voltage data of the retired power batteries, and performing dimension reduction processing on the charging voltage data by using a piecewise aggregation approximation algorithm to obtain dimension reduction data;

[0010] S2, performing Gram angle field conversion on the dimension reduction data to obtain polar coordinate data;

[0011] S3, converting the polar coordinate data into an encoded image by using Gram angle field conversion;

[0012] S4, inputting the encoded image into a trained battery classification model to obtain a battery capacity category, and the battery classification model adopts a ConvNeXt model.

[0013] Preferably, in step S1, the retired power batteries are subjected to constant current constant voltage charging and constant current discharging, and when the battery voltage reaches the charging and discharging cutoff voltage, the charging and discharging process is stopped to obtain charging and discharging data, and the charging voltage data is selected from the charging and discharging data.

[0014] Preferably, the charging voltage data is preprocessed to obtain voltage time series data Q=q1, q2, …, qm of length m. m After a dimension reduction process with a compression ratio of n=m / k, voltage time series data Q’=q’1, q’2, …, q’k of length k after dimension reduction is obtained. k q’k=Qk-q’1, q’2, …, q’k-1. i The following formula is used for calculation:

[0015]

[0016] The voltage time series data after dimension reduction corresponds to k time stamps t.

[0017] Preferably, step S2 specifically comprises:

[0018] The maximum-minimum scaler is used for normalization processing and scaling to the interval [-1, 1], and the formula is as follows:

[0019]

[0020] The scaled value is taken as the angle The time stamp is encoded as the radius r, and the data after normalization processing of the dimension reduction data is converted into polar coordinate data, and the formula is as follows:

[0021]

[0022] t1, t2, …, tk. iN is a constant factor generated after the regular polar coordinate system.

[0023] As a preference, the encoded image is a GADF image.

[0024] As a preference, the battery capacity category includes a first category, a second category and a third category, wherein the battery capacity of the first category is less than 2Ah, the battery capacity of the second category is between 2-2.4Ah, and the battery capacity of the third category is higher than 2.4Ah.

[0025] As a preference, the ConvNeXt model adopts a ConvNeXt-T network structure.

[0026] In a second aspect, the present application provides a retired power battery gradient screening device, comprising:

[0027] The dimension reduction module is configured to obtain the charging voltage data of the retired power battery, perform dimension reduction processing on the charging voltage data through a piecewise aggregation approximation algorithm, and obtain dimension reduction data;

[0028] The conversion module is configured to perform Gram angle field conversion on the dimension reduction data to obtain polar coordinate data.

[0029] The encoding module is configured to convert the polar coordinate data into an encoded image through Gram angle field conversion.

[0030] The classification module is configured to input the encoded image into a trained battery classification model to obtain a battery capacity category, and the battery classification model adopts a ConvNeXt model.

[0031] In a third aspect, the present application provides an electronic device, comprising one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any of the implementation manners of the first aspect.

[0032] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the method described in any of the implementation manners of the first aspect.

[0033] Compared with the prior art, the present application has the following beneficial effects:

[0034] (1) The retired power battery gradient screening method provided by the present application aims to the problem of quickly and accurately screening retired batteries, and the charging voltage time series data is processed through dimension reduction and normalization, which can not only retain feature information, but also reduce the amount of calculation.

[0035] (2) The gradient screening method for retired power batteries provided by the present application adopts Gram angle field transformation on the reduced dimension time series data to form a GADF image, and then adopts a ConvNeXt model to classify and screen the GADF image, with a screening accuracy of 95.26%, which fully verifies the accuracy and applicability of the method.

[0036] (3) The gradient screening method for retired power batteries provided by the present application can accurately and efficiently screen the battery capacity of the retired batteries, has a small amount of model operation data, high accuracy, and can effectively shorten the screening time. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0038] Figure 1 is an exemplary device architecture diagram to which an embodiment of the present application can be applied;

[0039] Figure 2 is a flowchart of the gradient screening method for retired power batteries of an embodiment of the present application;

[0040] Figure 3 is a Block structure diagram of the ConvNeXt model of the gradient screening method for retired power batteries of an embodiment of the present application;

[0041] Figure 4 is a schematic diagram of the gradient screening device for retired power batteries of an embodiment of the present application;

[0042] Figure 5 is a structure schematic diagram of a computer device suitable for realizing the electronic device of the embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0044] Figure 1 shows an exemplary device architecture 100 to which the gradient screening method for retired power batteries or the gradient screening device for retired power batteries of the embodiment of the present application can be applied.

[0045] As shown in Figure 1 The device architecture 100 can include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is a medium for providing a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 can include various connection types, such as wired, wireless communication links or fiber optic cables, etc.

[0046] A user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various applications can be installed on the terminal devices 101, 102, 103, such as data processing applications, file processing applications, etc.

[0047] The terminal devices 101, 102, 103 can be hardware or software. When the terminal devices 101, 102, 103 are hardware, they can be various electronic devices, including but not limited to smart phones, tablet computers, laptop computers and desktop computers, etc. When the terminal devices 101, 102, 103 are software, they can be installed in the above-mentioned electronic devices. They can be implemented as multiple software or software modules (such as software or software modules for providing distributed services) or as a single software or software module. No specific limitation is made herein.

[0048] The server 105 can be a server that provides various services, such as a background data processing server that processes files or data uploaded by the terminal devices 101, 102, 103. The background data processing server can process the obtained files or data to generate processing results.

[0049] It should be noted that the method for gradient screening of retired power batteries provided by the embodiments of the present application can be executed by the server 105 or the terminal devices 101, 102, 103, and correspondingly, the device for gradient screening of retired power batteries can be arranged in the server 105 or the terminal devices 101, 102, 103.

[0050] It should be understood that Figure 1 The number of terminal devices, networks and servers in the device architecture is only illustrative. According to the needs of implementation, there can be any number of terminal devices, networks and servers. In the case where the data to be processed does not need to be obtained from a remote place, the above-mentioned device architecture can not include a network, but only a server or a terminal device.

[0051] Figure 2 A method for gradient screening of retired power batteries is shown, which includes the following steps:

[0052] S1, obtaining charging voltage data of the retired power battery, performing dimension reduction processing on the charging voltage data through a piecewise aggregate approximation algorithm to obtain dimension reduction data.

[0053] In specific embodiments, the retired power battery is subjected to constant current constant voltage charging and constant current discharging in step S1, and when the battery voltage reaches the charge-discharge cutoff voltage, the charging and discharging process is stopped to obtain charge-discharge data, and the charging voltage data is selected from the charge-discharge data.

[0054] In specific embodiments, the charging voltage data is preprocessed to obtain voltage time series data Q=q1, q2,..., qm of length m. m After dimension reduction with a compression ratio of n=m / k, voltage time series data Q' =q'1, q'2,..., q'k of length k after dimension reduction is obtained. k q'k =qk-qk-m. i The dimension reduction data is calculated according to the following formula:

[0055]

[0056] The voltage time series data after dimension reduction corresponds to k time stamps t.

[0057] Specifically, the battery test bench is composed of three parts: BTS-5V50A battery detection equipment for LiFePO4 battery charging and discharging, a hot chamber for controlling the cycle charging and discharging working temperature of the battery, and a PC end for programming control and storing experimental data. 97 retired power batteries are selected. In order to obtain the available capacity and internal resistance of the battery and other parameters, relevant experiments are carried out, including constant current constant voltage charging (CC-CV) and constant current discharging, and when the battery voltage reaches the charge-discharge cutoff voltage (3.6V and 2.0V), the charging and discharging process is stopped, the charge-discharge data is obtained, the charging voltage data is obtained from the charge-discharge data, and the voltage time series data of each battery is obtained after preprocessing the charging voltage data. The voltage time series data of each battery is processed by PAA dimension reduction to obtain dimension reduction data. Generally, the time series is divided into multiple subsequences, each subsequence is represented by the mean value of the original sequence, and the value of the subsequence represents the value of the original sequence at that time. Therefore, the long sequence is mapped to the short sequence, the calculation amount is shortened, and the data trend of the long sequence and the short sequence is consistent. The voltage time series data Q' =q'1, q'2,..., q'k after dimension reduction is a one-dimensional sequence composed of k time stamps t and the corresponding q'k. k i

[0058] S2, performing Gram angle field conversion on the dimension reduction data to obtain polar coordinate data.

[0059] ​​In specific embodiments, in order to make the inner product not biased to the maximum real value, the normalization processing is performed by using the maximum-minimum scaler to scale into the interval [-1, 1]. Step S2 specifically comprises:

[0060] The normalization processing is performed by using the maximum-minimum scaler to scale into the interval [-1, 1], and the formula is as follows:

[0061]

[0062] By taking the scaled value as the angle The timestamp is encoded as the radius r, and the normalized data after dimension reduction is converted into polar coordinate data, and the formula is as follows:

[0063]

[0064] Where t i is the timestamp, and N is a constant factor generated after the regular polar coordinate system, that is, the interval [0, 1] is divided into N equal parts. First, since it is in the interval [-1, 1], has monotonicity, for a given time series, in polar coordinates, a unique result is generated and its inverse function is also unique, so the entire encoding method is a double mapping and does not lose any feature information. Secondly, the dependence on time is maintained through the radius r.

[0065] A Gram matrix (GM) is composed of inner products between different vectors. Since the inner product can determine the angle and direction relationship between vectors, the Gram matrix can reflect the relationship between the vectors in the group. It is defined as a matrix composed of the inner product between any k vectors in n-dimensional Euclidean space, which is called the Gram matrix of the k vectors. The calculation is as follows:

[0066]

[0067] The Gram matrix calculated from the eigenvectors can extract the hidden relationship between image features. The similarity between two images can be determined by the GM of the two images.

[0068] S3, the polar coordinate data is converted into an encoded image by using the Gram angle field.

[0069] Specifically, GAF can be calculated by two encoding methods, i.e., generating GASF / GADF encoding through trigonometric function and difference operation. Since the previous screening sorting battery is based on one-dimensional data of the original battery, such as voltage, IC, etc., a new idea is provided for re-extracting one-dimensional data information by converting time series into polar coordinate system. With the passage of time, the corresponding value will be distorted between different angle points across the circle. GAF encoding obtains Gram angle difference field GADF through trigonometric function difference operation. Gram angle field refers to a method for encoding one-dimensional time series information by using GM. This method converts the signal in Cartesian coordinate system into polar coordinate form. GADF is obtained by converting one-dimensional data into polar coordinate form, calculating the trigonometric difference between each point, and easily utilizing angle perspective to identify time correlation in different time intervals. The final encoding image is GADF image.

[0070] S4, input the encoding image into the trained battery classification model to obtain the battery capacity category, and the battery classification model adopts ConvNeXt model.

[0071] Specifically, after the voltage time series after dimension reduction are sequentially subjected to normalization, polar coordinate conversion processing, and converted into GADF image by using Gram matrix inner product operation, the obtained GADF images of 97 batteries are single-labeled and divided into three categories. That is, the battery capacity category includes a first category, a second category and a third category, wherein the battery capacity of the first category is less than 2Ah, the battery capacity of the second category is between 2-2.4Ah, and the battery capacity of the third category is higher than 2.4Ah.

[0072] Specifically, since the ConvNeXt model has the inductive bias inherent to the convolutional neural network, it has high-accuracy classification performance. The obtained GADF image dataset can be used as the input of the model, the training set and the validation set are divided, the network parameters are constantly adjusted to make the network performance optimal, and finally the trained battery classification model is obtained. The trained battery classification model can obtain accurate screening results. In the embodiments of the present application, the ConvNeXt model adopts ConvNeXt-T network structure.

[0073] The network principle of the ConvNeXt model is as follows:

[0074] The ConvNeXt model is based on the fusion of ResNet-50 and Swin-T network. The ResNet network is modified and optimized to verify the effectiveness of the ViT structure on the CNN network. The Resnet is modified in a transformer-like manner by modifying the convolution kernel size, stage calculation (i.e., the number of times of stacking blocks), stem module, and increasing the feature channel.

[0075] The ConvNeXt model can be divided into five types of T, S, B, L and XL according to the depth of each layer Block and the number of Stage stacking. In the embodiments of the present application, the ConvNeXt-T network structure with the best comprehensive performance is selected to classify and screen the GADF images of the sorted battery voltage curve. After data preprocessing, the two-dimensional feature map after GADF processing is taken as the input of the ConvNeXt model in the embodiments of the present application, which avoids complex feature extraction and data reconstruction processes. The network parameters of the ConvNeXt model in the embodiments of the present application are shown in Table 1.

[0076] Table 1

[0077]

[0078]

[0079] Referring to Table 1 and Figure 3 , the main part of the ConvNeXt-T network structure includes four Stages. Before the GADF image is input into the four Stages, it will pass through a convolution layer with a convolution kernel size of 4*4, a stride of 96 and a padding of 4. In each Stage, a depthwise convolution is introduced, so that each convolution kernel processes a channel separately and performs spatial information mixing and weighting within a single channel. Figure 3 is a Block structure diagram of the ConvNeXt model. As can be seen from the diagram, each Block changes due to the different depths of the feature input of each layer. Through the inverted bottleneck structure (i.e., the dimension changes in the form of small dimension-large dimension-small dimension), information conversion between different dimensional feature spaces is avoided to reduce information loss caused by dimension compression. In order to reduce the amount of calculation, the depthwise convolution layer is moved to the front at the same time. The down-sampling module in front of the Block is composed of a regularization module and a convolution layer with a convolution kernel size and a stride of 2. At the same time, by replacing the activation function GELU and the normalization layer LN (Layer Normalization, LN), better alignment comparison and reduction of negative impact on network performance are achieved. Finally, after global average pooling and full connection layer output, a layer of Softmax detection function is used to convert it into a probability distribution output. The ConvNeXt model has the advantages of accuracy, efficiency and strong scalability due to the inductive bias and more reasonable network parameters.

[0080] To verify the effectiveness of the proposed method, the battery is divided into three types with different capacities. In the embodiment of the application, the battery capacity less than 2 Ah is classified as class A, the battery capacity between 2-2.4 Ah is classified as class B, and the battery capacity higher than 2.4 is classified as class C, and the training data is constructed. The pre-processed GADF image is input into the ConvNeXt model, which first passes through a convolutional layer connected to the regularization module, and then passes through four Block modules with different dimensions in turn, and finally outputs the predicted classification result through global average pooling and Softmax fully connected layer. Random sampling is performed on the sample set, and the training data is divided into training set and validation set according to the ratio of 8:2, which is input into the ConvNeXt model for training. According to the accuracy of the validation set, the hyperparameters of the model are repeatedly adjusted until the model reaches the best ideal state.

[0081] By selecting different dimension reduction scales, different deep learning image classification models, and different battery sample quantities, comparative experiments are performed. The retired power battery gradient screening method proposed in the embodiment of the application can realize accurate classification, and has high precision and applicability.

[0082] The retired power battery gradient screening method proposed in the embodiment of the application can retain feature information while reducing the amount of calculation by using PAA dimension reduction on the original charging voltage data. Then the voltage time series after dimension reduction is converted into GADF image by Gram angle field, and further classified and screened by ConvNeXt model. Compared with other typical classification models, the screening accuracy of GADF-ConvNeXt model used in the embodiment of the application is the highest, which can reach 95.26%, which fully verifies the accuracy and applicability of the GADF-ConvNeXt method.

[0083] Table 2

[0084]

[0085] Further reference Figure 4 , as an implementation of the method shown in the above figures, the application provides an embodiment of a retired power battery gradient screening device. The device embodiment corresponds to the method embodiment shown in Figure 2 , and the device can be applied in various electronic devices.

[0086] The embodiment of the application provides a retired power battery gradient screening device, which comprises:

[0087] The dimension reduction module 1 is configured to obtain the charging voltage data of the retired power battery, perform dimension reduction processing on the charging voltage data by using the piecewise aggregate approximation algorithm, and obtain dimension reduction data;

[0088] The conversion module 2 is configured to perform Gram angle field conversion on the dimension-reduced data to obtain polar coordinate data.

[0089] The encoding module 3 is configured to convert the polar coordinate data into an encoded image by using Gram angle field conversion.

[0090] The classification module 4 is configured to input the encoded image into a trained battery classification model to obtain a battery capacity category, the battery classification model adopting a ConvNeXt model.

[0091] Reference will be made to the following description Figure 5 which shows a structural schematic diagram of a computer device 500 of an electronic device (for example Figure 1 a server or a terminal device) suitable for being used to implement the embodiments of the present application. Figure 5 The electronic device shown is merely an example and should not bring any limitation to the functions and use range of the embodiments of the present application.

[0092] As shown in Figure 5 , the computer device 500 includes a central processing unit (CPU) 501 and a graphics processor (GPU) 502, which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 503 or programs loaded from a storage portion 509 into a random access memory (RAM) 504. In the RAM 504, various programs and data required for the operation of the device 500 are also stored. The CPU 501, the GPU 502, the ROM 503, and the RAM 504 are connected to each other through a bus 505. An input / output (I / O) interface 506 is also connected to the bus 505.

[0093] The following components are connected to the I / O interface 506: an input portion 507 including a keyboard, a mouse, and the like; an output portion 508 including a display such as a cathode ray tube (CRT) display, a liquid crystal display (LCD), and the like, and a speaker, and the like; a storage portion 509 including a hard disk, and the like; and a communication portion 510 including a network interface card such as a LAN card, a modem, and the like. The communication portion 510 performs communication processing via a network such as the Internet. A drive 511 can also be connected to the I / O interface 506 as needed. A removable medium 512 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 511 as needed, so that a computer program read therefrom is installed in the storage portion 509 as needed.

[0094] In particular, the processes described above with reference to the flowcharts can be implemented as a computer software program in accordance with embodiments of the present disclosure. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer readable medium, the computer program comprising program code for executing the methods illustrated by the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by the communication section 510, and / or installed from the removable medium 512. When the computer program is executed by the central processing unit (CPU) 501 and the graphics processor (GPU) 502, the above-described functions defined in the methods of the present application are executed.

[0095] It should be noted that the computer readable medium described in the present application can be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor device, device or means, or any combination of the above. More specific examples of computer readable medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer readable medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution device, device or means. In the present application, the computer readable signal medium can include a data signal carried in a baseband or as a part of a carrier wave, which carries computer readable program code. Such a propagated data signal can take many forms, including but not limited to an electromagnetic signal, an optical signal or any suitable combination of the above. The computer readable signal medium can also be any computer readable medium other than the computer readable medium that can send, propagate or transmit the program for use by or in conjunction with an instruction execution device, device or means. The program code contained on the computer readable medium can be transmitted by any suitable medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0096] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0097] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0098] The modules described in the embodiments of the present application can be implemented by software, or by hardware. The modules described can also be arranged in a processor.

[0099] As another aspect, the application also provides a computer readable medium, which can be included in the electronic device described in the above embodiments, or can exist independently without being assembled into the electronic device. The computer readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire charging voltage data of the retired power battery, perform dimension reduction processing on the charging voltage data by using a piecewise aggregate approximation algorithm to obtain dimension reduction data; perform Gram angle field conversion on the dimension reduction data to obtain polar coordinate data; convert the polar coordinate data into an encoded image by using Gram angle field conversion; input the encoded image into a trained battery classification model to obtain a battery capacity category, and the battery classification model uses a ConvNeXt model.

[0100] The above description is merely the preferred embodiments of the present application and the explanation of the technical principles used. It should be understood by those skilled in the art that the scope of the application involved in the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combinations of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features can be replaced with the technical features disclosed in the present application (but not limited to) having similar functions to form technical solutions.

Claims

1. A method for gradient screening of decommissioned power batteries, characterized in that, The method comprises the following steps: S1, obtaining charging voltage data of the retired power battery, performing dimension reduction processing on the charging voltage data through a piecewise aggregation approximation algorithm to obtain dimension reduction data, and preprocessing the charging voltage data to obtain voltage time series data Q=q1, q2, …, qm of length m m , after a dimension reduction process with a compression ratio of n=m / k, obtaining voltage time series data Q'=q'1, q'2, …, q'k of length k after dimension reduction k , wherein q' i is calculated according to the following formula: ； The reduced dimension voltage time series data correspond to k time stamps t; S2, the gram angle field conversion is carried out on the reduced dimension data to obtain polar coordinate data, and specifically comprises: The maximum-minimum scaler is used for normalization processing and scaling to the interval [-1, 1], and the formula is as follows: ； By taking the scaled value as an angle φ, the timestamp is encoded as a radius The normalized data of the reduced dimension data is converted into polar coordinate data, and the formula is as follows: ； wherein, is the time stamp, N is a constant factor resulting from the normalization of the polar coordinate system; S3, the polar coordinate data is converted into an encoded image by gram angle field conversion; S4, the encoded image is input into a trained battery classification model to obtain a battery capacity category, and the battery classification model adopts a ConvNeXt model.

2. The retired power battery gradient screening method according to claim 1, characterized in that, In the step S1, the retired power battery is charged and discharged by constant current and constant voltage, and when the battery voltage reaches the charge-discharge cutoff voltage, the charging and discharging process is stopped, and charge-discharge data is obtained, and the charge voltage data is selected from the charge-discharge data.

3. The retired power battery gradient screening method according to claim 1, characterized in that, The encoded image is a GADF image.

4. The retired power battery gradient screening method according to claim 1, characterized in that, The battery capacity category includes a first category, a second category and a third category, wherein the battery capacity of the first category is less than 2 Ah, the battery capacity of the second category is between 2 and 2.4 Ah, and the battery capacity of the third category is higher than 2.4 Ah.

5. The retired power cell gradient screening method of claim 1, wherein, The ConvNeXt model adopts a ConvNeXt-T network structure.

6. A device for gradient screening of decommissioned power batteries, characterized in that, The method comprises: The dimension reduction module is configured to obtain charging voltage data of the retired power battery, perform dimension reduction processing on the charging voltage data through a piecewise aggregation approximation algorithm to obtain dimension reduction data, and obtain voltage time sequence data Q=q1, q2, …, qm of length m by preprocessing the charging voltage data. m After a dimension reduction process with a compression ratio of n=m / k, voltage time sequence data Q'=q'1, q'2, …, q'k of length k after dimension reduction is obtained. k q'k=qk-k+1+qk-k+2+…+qk-1+qk i is calculated according to the following formula: ； The reduced dimension voltage time series data correspond to k time stamps t; The conversion module is configured to convert the reduced dimension data by gram angle field conversion to obtain polar coordinate data, and specifically comprises: The maximum-minimum scaler is used for normalization processing and scaling to the interval [-1, 1], and the formula is as follows: ； By taking the scaled value as an angle φ, the timestamp is encoded as a radius The normalized data of the reduced dimension data is converted into polar coordinate data, and the formula is as follows: ； wherein, is a time stamp, N is a constant factor resulting from the normalization of the polar coordinate system; The encoding module is configured to convert the polar coordinate data into an encoded image by gram angle field conversion; The classification module is configured to input the encoded image into a trained battery classification model to obtain a battery capacity category, and the battery classification model adopts a ConvNeXt model.

7. An electronic device comprising: One or more processors; Storage means for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1-5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method of any one of claims 1-5.

Images