A method, system, device and medium for secondary sorting of retired batteries based on random charging fragments

Abstract

The present invention belongs to the technical field of retired battery sorting, and specifically relates to a method, system, device and medium for secondary sorting of retired batteries based on random charging fragments. The method includes generating a single battery identification code, building a battery code database, controlling the charge and discharge of the single battery based on a preset charge and discharge strategy, obtaining a time series sample data set of temperature, voltage and capacity changes, converting it into two-dimensional image data and storing it, and training a prediction model through a two-dimensional convolutional neural network to estimate the maximum available capacity of the battery to achieve secondary sorting of the battery. The system includes code scanning, charge and discharge control, data acquisition, data processing and sorting modules. The present invention can reduce test time and energy consumption, improve battery sorting efficiency and accuracy, reduce memory usage and back-end model running time, and is suitable for efficient and accurate sorting of retired batteries.

Description

Technical Field

[0001] The present invention belongs to the technical field of retired battery sorting, and specifically relates to a retired battery secondary sorting method, system, equipment and medium based on random charging fragments. Background Art

[0002] After long-term vehicle-mounted applications, when these battery packs reach the end of their service life and need to be retired, the single cells will show significant differences in maximum available capacity, ohmic internal resistance, and polarization internal resistance. This inconsistency not only weakens the overall performance of the battery pack, but also poses a challenge to the reuse of retired batteries. Therefore, before reusing retired batteries, the health status of the single cells or modules must be evaluated, and they must be classified and screened based on the evaluation results.

[0003] Existing retired battery screening technologies mainly include the following methods: First, technicians can use professional tools, such as DC internal resistance testers and multimeters, to measure various parameters of the same group of batteries and grade the batteries based on these data. However, the effectiveness of this method is highly dependent on all batteries having similar remaining state of charge (SOC), otherwise the measurement results may not be accurate enough. Secondly, the capacity of the battery is fully tested through charge and discharge cycles, and then classified based on the capacity displayed by the battery during full discharge or charge and discharge. Although this method can provide a relatively accurate assessment, it is time-consuming and energy-consuming. Finally, specially designed charge and discharge conditions can be used to evaluate battery performance, and the internal parameters of the battery can be obtained through simulation calculations, thereby achieving efficient classification. Although this method is efficient, its implementation requires high expertise in the fields of electrochemistry and control, and the technical threshold is relatively high, requiring high personnel costs.

[0004] In summary, the current screening technology for retired batteries still has the following difficulties and shortcomings:

[0005] 1. It requires a lot of tedious and meticulous feature engineering work, requires high expertise in the field of electrochemistry, and has high technical barriers. At the same time, different types of batteries often require specially designed different features, which leads to the problem that the scope of application of the method is not wide enough.

[0006] 2. Methods based on charge and discharge cycles often bring great application costs, resulting in low efficiency in actual industrial scenarios. Although the method of sampling a specific charging interval using feature engineering does not require a complete charge and discharge test, it requires a cumbersome charging test process. The time required for different batteries varies greatly, making it difficult to automate.

[0007] 3. Traditional algorithms based on XGBoost, clustering, etc. have limited effects and insufficient reference value, and are difficult to meet the accuracy required for actual production. Summary of the invention

[0008] The purpose of the present invention is to provide a method, system, equipment and medium for secondary sorting of retired batteries based on random charging fragments, so as to solve the problems of retired battery screening technology in the prior art being cumbersome, time-consuming and energy-consuming, high technical threshold, limited application scope, low efficiency and insufficient precision.

[0009] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0010] In a first aspect, the present invention proposes a secondary sorting method for retired batteries based on random charging fragments, the method comprising:

[0011] Generate identification codes for single cells and build a battery coding database;

[0012] Controlling the charging and discharging of single cells based on preset charging and discharging strategies and preset operating conditions;

[0013] Acquire a time series sample data set of a single battery; wherein the time series sample data is determined according to a temperature change time series, a voltage change time series, and a capacity change time series under the charge and discharge control scenario;

[0014] Converting the time series sample data set into two-dimensional image data and storing the data in the battery coding database, and establishing a corresponding relationship between the identification code, the single battery and the two-dimensional image data in the battery coding database;

[0015] In combination with the battery coding database, a prediction model based on a two-dimensional convolutional neural network is trained, and a stochastic gradient descent algorithm is combined to minimize the difference between the actual maximum available capacity of the battery and the estimated value, so as to optimize the prediction model parameters;

[0016] The two-dimensional image data of the target single battery to be sorted is input into the trained prediction model, and the estimated value of the maximum available capacity of the battery is output, and the battery is secondary sorted based on the estimated value of the maximum available capacity.

[0017] Furthermore, the charging and discharging control of the single battery based on the preset charging and discharging strategy and the preset working condition includes:

[0018] Set a time period with the same time interval and time starting point as the charging and discharging time interval;

[0019] The single battery is controlled to be charged and discharged at a constant current based on the time interval.

[0020] Furthermore, the step of obtaining a time series sample data set of a single battery includes:

[0021] Get temperature change time series , record the temperature change during the set charging phase, ,in, Indicates time, from arrive ;

[0022] Get voltage change sequence , record the voltage change during the set charging stage, ,in, Indicates time, from arrive ;

[0023] Get capacity change time series , record the cumulative charging capacity change time series during the set charging stage , ,in, Indicates time, from arrive ;

[0024] Temperature change time series , voltage change time series and capacity change time series Normalize to get the normalized sequence , and , get the time series sample data set S of the set charge and discharge stage, .

[0025] Furthermore, converting the time series sample data set into two-dimensional image data includes:

[0026] Adaptive average pooling operation Convert a time series sample data set of any length into a fixed length Time series samples of

[0027] By reshaping the operation The fixed length is The time series samples are converted into an image of a set size to form two-dimensional image data, where Indicates the image height, Indicates the image width.

[0028] Furthermore, the prediction model constructed based on the two-dimensional convolutional neural network is trained in combination with the battery coding database, and the difference between the actual maximum available capacity of the battery and the estimated value is minimized in combination with the stochastic gradient descent algorithm to optimize the prediction model parameters, including:

[0029] Build a prediction model based on a two-dimensional convolutional neural network, including building convolutional layers, pooling layers, and fully connected layers;

[0030] The converted two-dimensional image data is input into the prediction model for training, the features in the two-dimensional image data are gradually extracted through the convolution layer and the pooling layer, and the estimated value of the maximum available capacity of the battery is output through the fully connected layer ;

[0031] Use stochastic gradient descent algorithm to minimize the actual maximum available capacity With estimated value The difference between them is used to optimize the prediction model parameters and obtain the prediction model that has completed training.

[0032] In the second aspect, the present invention proposes a secondary sorting system for retired batteries based on random charging segments, the system comprising a code scanning module, a charge and discharge control module, a data acquisition module, a data processing module and a sorting module, wherein:

[0033] The code scanning module is used to generate an identification code for each single battery and transmit it to a battery code database pre-built in the data processing module to store the battery identity code;

[0034] The charge and discharge control module is used to control the charge and discharge of the single battery based on a preset charge and discharge strategy and preset working conditions;

[0035] The data acquisition module is used to obtain a time series sample data set of a single battery; wherein the time series sample data is determined according to a temperature change time series, a voltage change time series and a capacity change time series under the charge and discharge control scenario;

[0036] The data processing module includes a processor and a memory;

[0037] The processor is used to convert the time series sample data set into two-dimensional image data, establish a corresponding relationship between the identification code, the single battery and the two-dimensional image data in the battery coding database; and train the prediction model built based on the two-dimensional convolutional neural network in combination with the battery coding database, and minimize the difference between the actual maximum available capacity of the battery and the estimated value in combination with the stochastic gradient descent algorithm to optimize the prediction model parameters;

[0038] The memory is used to store the battery coding database, and is used to store the two-dimensional image data in the battery coding database;

[0039] The sorting module is used to input the two-dimensional image data of the target single battery to be sorted into the trained prediction model, output the estimated value of the maximum available capacity of the battery, and perform secondary sorting of the battery based on the estimated value of the maximum available capacity.

[0040] Furthermore, the charge and discharge control module includes:

[0041] A parameter setting unit, used to set a time period with the same time interval and time starting point as a charging and discharging time interval;

[0042] The control unit is used to perform constant current charging and discharging control on the single battery based on the time interval.

[0043] Furthermore, the data acquisition module includes:

[0044] The first acquisition unit is used to acquire the temperature change time series , record the temperature change during the set charging phase, ,in, Indicates time, from arrive ;

[0045] The second acquisition unit is used to acquire the voltage change sequence , record the voltage change during the set charging stage, ,in, Indicates time, from arrive ;

[0046] The third acquisition unit is used to obtain the capacity change time series , record the cumulative charging capacity change time series during the set charging stage , ,in, Indicates time, from arrive ;

[0047] Data integration unit, used to convert temperature change time series , voltage change time series and capacity change time series Normalize to get the normalized sequence , and , get the time series sample data set S of the set charge and discharge stage, .

[0048] In a third aspect, the present invention provides an electronic device, comprising:

[0049] a processor unit; a memory unit for storing instructions executable by the processor unit;

[0050] Wherein, the processor unit is configured to execute the instructions to implement the retired battery secondary sorting method as described in any one of the above items.

[0051] In a fourth aspect, the present invention proposes a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by a processor unit of an electronic device, the electronic device is enabled to execute the retired battery secondary sorting method as described in any one of the above items.

[0052] The beneficial effects of the present invention are:

[0053] 1. The sorting method proposed in the present invention uses the random charging fragments generated by retired batteries during use as input data, combined with a two-dimensional convolutional neural network (2D-CNN) for analysis, to achieve an effective assessment of the battery health status. Compared with traditional retired battery screening technology, the present invention does not require a complete charge and discharge cycle test, and only requires random charging fragments to evaluate the battery status, thereby significantly reducing test time and energy consumption. This method not only reduces experimental costs, but also improves test efficiency, making the secondary sorting of retired batteries more efficient and environmentally friendly.

[0054] 2. The present invention can fully extract the features in the battery charging data through the designed two-dimensional convolutional neural network model, and establish a nonlinear relationship between the random charging segment and the health status of the retired battery. In practical applications, this method can achieve a low mean absolute percentage error (MAPE) on a battery with a rated capacity of 25Ah, showing a high-precision battery status assessment capability. In addition, compared with the general neural network model, the 2D-CNN model proposed in the present invention can achieve better estimation effects while significantly reducing the memory (video memory) occupancy and back-end model running time, thereby improving the practicality and efficiency of the model. This makes the present invention have broad application prospects and important economic value in the field of secondary sorting of retired batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of a process of a method for secondary sorting of retired batteries based on random charging segments provided in an embodiment of the present application;

[0056] Figure 2 A schematic diagram of the structure of a retired battery secondary sorting system based on random charging segments provided in an embodiment of the present application;

[0057] Figure 3 A schematic diagram of another structure of a retired battery secondary sorting system based on random charging segments provided in an embodiment of the present application;

[0058] Figure 4 A schematic diagram of a sorting process of a retired battery secondary sorting method based on random charging segments provided in an embodiment of the present application;

[0059] Figure 5A schematic diagram of a sorting algorithm in a method for secondary sorting of retired batteries based on random charging fragments provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0061] Example 1

[0062] like Figure 1 As shown, this embodiment proposes a secondary sorting method for retired batteries based on random charging fragments, and the method includes the following steps:

[0063] S1. Generate an identification code of a single battery and build a battery code database, wherein the identification code is specifically an identity code of the single battery, such as a QR code, a barcode, an electronic tag, etc.;

[0064] S2, controlling the charging and discharging of the single battery based on a preset charging and discharging strategy and preset operating conditions;

[0065] More specifically, the charging and discharging of the single cell is controlled based on a preset charging and discharging strategy and preset working conditions, including: setting a time period with the same time interval and time starting point as the charging and discharging time interval; and, performing constant current charging and discharging control on the single cell based on the time interval.

[0066] For example, the time interval can be selected as 10 seconds, 30 seconds, etc., and the time starting point can be selected as any time. The specific working condition can be set to constant current charging for 10 minutes, constant current charging for 20 minutes, etc. The data under these specific working conditions can reflect the performance changes of the battery under different conditions.

[0067] S3, obtaining a time series sample data set of a single battery; wherein the time series sample data is determined according to a temperature change time series, a voltage change time series, and a capacity change time series under a charge and discharge control scenario;

[0068] S4, converting the time series sample data set into two-dimensional image data and storing it in a battery coding database, and establishing a corresponding relationship between the identification code, the single battery and the two-dimensional image data in the battery coding database;

[0069] S5. Combined with the battery coding database, the prediction model based on the two-dimensional convolutional neural network is trained, and combined with the stochastic gradient descent algorithm, the difference between the actual maximum available capacity of the battery and the estimated value is minimized to optimize the prediction model parameters;

[0070] S6. Input the two-dimensional image data of the target single battery to be sorted into the trained prediction model, output the estimated value of the maximum available capacity of the battery, and perform secondary sorting of the battery based on the estimated value of the maximum available capacity.

[0071] It can be understood that the core of the sorting method proposed in this embodiment is to use the random charging fragments generated by retired batteries during use as input data, and analyze these data through machine learning technology, especially a specially designed two-dimensional convolutional neural network (2D-CNN), so as to achieve effective evaluation of the battery health status. 2D-CNN can automatically extract useful features from the input data through a series of convolutional layers and pooling layers. The convolution operation allows the network to slide a small filter or kernel on the input data to identify specific patterns or features; while the pooling layer reduces the dimensionality of the data by downsampling while retaining the most important features. This mechanism enables 2D-CNN to capture local correlations and global structures in the input data.

[0072] Accordingly, in this algorithm, in order for 2D-CNN to effectively process battery charging data, it is first necessary to convert the one-dimensional time series data (time series sample data set) into two-dimensional image data. This is because the original one-dimensional time series data not only contains short-term fluctuation information that changes over time, but also implies long-term trend patterns. By converting these data into a two-dimensional format, the relationship between adjacent time points can be observed in the horizontal direction (i.e., the column direction), which reflects the short-term fluctuation characteristics of the data; while in the vertical direction (i.e., the row direction), the correlation between different time intervals can be captured, revealing the long-term trend of the data. This conversion method provides an ideal input form for 2D-CNN, enabling it to simultaneously consider the temporal locality and long-term dependence of the data. In this way, 2D-CNN can effectively process the converted battery charging data without making any major modifications to the existing deep learning framework.

[0073] Further preferably, in step S3, obtaining a time series sample data set of a single battery includes:

[0074] Get temperature change time series , record the temperature change during the set charging phase, ,in, Indicates time, from arrive ;

[0075] Get voltage change sequence , record the voltage change during the set charging stage, ,in, Indicates time, from arrive ;

[0076] Get capacity change time series , record the cumulative charging capacity change time series during the set charging stage , ,in, Indicates time, from arrive ;

[0077] Temperature change time series , voltage change time series and capacity change time series Normalize to get the normalized sequence , and , get the time series sample data set S of the set charge and discharge stage,

[0078] .

[0079] Further preferably, in step S4, before converting the time series sample data set into two-dimensional image data, the step further includes dividing the time series sample data set into a training set and a test set.

[0080] Further preferably, converting the time series sample data set into two-dimensional image data includes:

[0081] Adaptive average pooling operation Convert a time series sample data set of any length into a fixed length Time series samples of

[0082] By reshaping the operation The fixed length is The time series samples are converted into an image of a set size to form two-dimensional image data, where Indicates the image height, Indicates the image width.

[0083] Further preferably, in step S5, the prediction model based on the two-dimensional convolutional neural network (2D-CNN) is trained in combination with the battery coding database, and the actual maximum available capacity of the battery is minimized in combination with the stochastic gradient descent algorithm. and estimated values The method comprises: constructing a prediction model based on a two-dimensional convolutional neural network, including constructing a convolution layer, a pooling layer, and a fully connected layer; and inputting the converted two-dimensional image data into the prediction model for training, gradually extracting features in the two-dimensional image data through the convolution layer and the pooling layer, and outputting an estimated value of the maximum available capacity of the battery through the fully connected layer. ; Use stochastic gradient descent algorithm to minimize the actual maximum available capacity With estimated value The difference between them is used to optimize the prediction model parameters and obtain the prediction model that has completed training.

[0084] It is understandable that in practical applications, after the prediction model is trained and its effectiveness is verified by the test set, it can be directly applied to the health status assessment of retired batteries. At this time, it is only necessary to perform the test process, that is, input new battery data into the trained model, to quickly and accurately obtain the maximum available capacity estimate of the battery, thereby realizing efficient secondary sorting and utilization of retired batteries.

[0085] like Figure 2-5 As shown, based on the same inventive concept, in another embodiment of the present invention, a secondary sorting system for retired batteries based on random charging segments is proposed, the system includes a code scanning module, a charge and discharge control module, a data acquisition module, a data processing module and a sorting module, wherein:

[0086] The code scanning module is used to generate an identification code for each single battery and transmit it to a battery code database pre-built in the data processing module to store the battery identity code;

[0087] The charge and discharge control module is used to control the charge and discharge of the single battery based on the preset charge and discharge strategy and preset working conditions;

[0088] The data acquisition module is used to obtain a time series sample data set of a single battery; wherein the time series sample data is determined according to a temperature change time series, a voltage change time series and a capacity change time series under a charge and discharge control scenario;

[0089] The data processing module includes a processor and a memory;

[0090] The processor is used to convert the time series sample data set into two-dimensional image data, establish a corresponding relationship between the identification code, the single battery and the two-dimensional image data in the battery coding database; and is used to train the prediction model built based on the two-dimensional convolutional neural network in combination with the battery coding database, and minimize the difference between the actual maximum available capacity of the battery and the estimated value in combination with the stochastic gradient descent algorithm to optimize the prediction model parameters;

[0091] The memory is used to store the battery coding database, and is used to store the two-dimensional image data in the battery coding database;

[0092] The sorting module is used to input the two-dimensional image data of the target single battery to be sorted into the trained prediction model, output the estimated value of the maximum available capacity of the battery, and perform secondary sorting of the battery based on the estimated value of the maximum available capacity.

[0093] It should be noted here that each module in the above-mentioned secondary sorting system corresponds to steps S1 to S6 in implementing the above-mentioned secondary sorting method, and the instances and application scenarios implemented by multiple modules and corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned embodiment 1.

[0094] Further preferably, the charge and discharge control module includes:

[0095] A parameter setting unit, used to set a time period with the same time interval and time starting point as a charging and discharging time interval;

[0096] The control unit is used to perform constant current charging and discharging control on the single battery based on the time interval.

[0097] Further preferably, the data acquisition module includes:

[0098] The first acquisition unit is used to acquire the temperature change time series , record the temperature change during the set charging phase, ,in, Indicates time, from arrive ;

[0099] The second acquisition unit is used to acquire the voltage change sequence , record the voltage change during the set charging stage, ,in, Indicates time, from arrive ;

[0100] The third acquisition unit is used to obtain the capacity change time series , record the cumulative charging capacity change time series during the set charging stage , ,in, Indicates time, from arrive ;

[0101] Data integration unit, used to convert temperature change time series , voltage change time series and capacity change time series Normalize to get the normalized sequence , and , get the time series sample data set S of the set charge and discharge stage, .

[0102] Exemplarily, in some specific implementations, based on the above-mentioned sorting system, the required equipment includes a barcode scanner, a charging and discharging instrument, a data acquisition instrument, a storage and computing device, which respectively correspond to the barcode scanning module, charging and discharging control module, data acquisition module, data processing module and sorting module in the above-mentioned system.

[0103] The specific implementation includes the following division of labor:

[0104] The barcode scanner generates a recognizable identity code for each single battery, allowing the entire automated test screening and classification process to be quickly mapped to the battery cell;

[0105] Charge and discharge instruments are the most commonly used test equipment for batteries. Most charge and discharge instruments can perform charge and discharge tests on batteries according to the set charge and discharge strategies.

[0106] The data acquisition instrument is used to collect battery charging and discharging information, such as current, voltage, time, etc., and external physical parameters, such as temperature changes, size changes, etc.;

[0107] The storage device stores the collected battery information, and the computing device executes algorithms to analyze the data, and they are generally integrated into the same test host computer.

[0108] Exemplarily, based on the above-mentioned sorting system, the sorting method includes the following steps:

[0109] (1) Generate an identification code (such as a QR code, barcode, electronic tag, etc.) for each retired single battery through a barcode scanner, and establish a corresponding database in the data storage device corresponding to the identification code;

[0110] (2) The host computer controls the charging and discharging instrument to execute the charging and discharging strategy of the agreed working conditions. The specific strategy is as follows;

[0111] (4) The data acquisition instrument collects the charge and discharge data of the retired single battery and the required external physical parameter data of the battery, and stores the collected data in the database established for the corresponding retired single battery after dimension reduction through the autoencoder;

[0112] (5) The host computing device obtains the retired battery data from the database, processes and analyzes it according to the designed battery health status assessment algorithm, and finally obtains the status assessment result.

[0113] According to the above embodiments of the present invention, the artificial neural network supervised learning method of the present invention can achieve high-precision evaluation of the health status of retired batteries, thereby realizing secondary sorting of batteries;

[0114] And using randomly segmented charging data as input reduces experimental costs. Since the initial capacity of each battery cell is different at the beginning of the test, the method based on specific charge-discharge curve intervals will introduce additional experimental overhead. The method using random segmentation meets the actual needs of real-world scenarios.

[0115] In some specific implementations, the sorting method proposed by the present invention has the following advantages:

[0116] (1) It is possible to estimate the maximum available capacity of a battery using only random battery charging test data, which can reduce the test time and energy consumption compared to the method of determining the maximum available capacity of a battery using a complete charge and discharge cycle;

[0117] (2) 2D-CNN can fully extract features from the data, thereby establishing a nonlinear relationship between random charging segments and the health status of retired batteries. Experimental results show that a mean absolute percentage error (MAPE) of 1.47% can be achieved on a battery with a rated capacity of 25Ah, and a mean absolute percentage error (MAPE) of 0.72% can be achieved on a battery with a rated capacity of 15.5Ah.

[0118] (3) Since the proposed 2D-CNN is designed specifically for battery state estimation, compared with general-purpose ResNet, ShuffleNet, EfficientNet, etc., it can achieve better estimation results while reducing memory (video memory) usage by an average of 59% and reducing backend model running time by 77%.

[0119] In another embodiment provided by the present invention, an electronic device is also provided, including: a processor unit; a memory unit for storing instructions executable by the processor unit; wherein the processor unit is configured to execute instructions to implement the retired battery secondary sorting method as described above.

[0120] In another embodiment provided by the present invention, a computer-readable storage medium is also provided. When the instructions in the computer-readable storage medium are executed by a processor unit of an electronic device, the electronic device can perform the retired battery secondary sorting method as described above.

[0121] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0122] In addition, each functional module in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0123] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A secondary sorting method for retired batteries based on random charging fragments, characterized in that: The method comprises: Generate identification codes for single cells and build a battery coding database; Based on a preset charging and discharging strategy and a preset working condition, the charging and discharging of the single battery is controlled; specifically, a time period with the same time interval and time starting point is set as the charging and discharging time interval; based on the time interval, the single battery is controlled to be charged and discharged at a constant current; Acquire a time series sample data set of a single battery; wherein the time series sample data is determined according to a temperature change time series, a voltage change time series, and a capacity change time series under the charge and discharge control scenario; The time series sample data set is converted into two-dimensional image data and stored in the battery coding database, and a corresponding relationship between the identification code, the single battery and the two-dimensional image data is established in the battery coding database; wherein, the time series sample data set is converted into two-dimensional image data including using an adaptive average pooling operation Convert the time series sample data set of any length into a fixed length A time series sample of The fixed length is The time series samples are converted into an image of a set size to form two-dimensional image data, where Indicates the image height, Indicates the image width; In combination with the battery coding database, a prediction model based on a two-dimensional convolutional neural network is trained, and a stochastic gradient descent algorithm is combined to minimize the difference between the actual maximum available capacity of the battery and the estimated value, so as to optimize the prediction model parameters; The two-dimensional image data of the target single battery to be sorted is input into the trained prediction model, and the estimated value of the maximum available capacity of the battery is output, and the battery is secondary sorted based on the estimated value of the maximum available capacity.

2. According to claim 1, a method for secondary sorting of retired batteries based on random charging fragments is characterized in that: The step of obtaining a time series sample data set of a single battery includes: Get temperature change time series , record the temperature change during the set charging phase, ,in, Indicates time, from arrive ; Get voltage change sequence : Record the voltage change during the set charging stage, ,in, Indicates time, from arrive ; Get capacity change time series : Record the cumulative charging capacity change time series within the set charging stage ,in, Indicates time, from arrive ; Temperature change time series , voltage change time series and capacity change time series Normalize to get the normalized sequence and , get the time series sample data set of the set charge and discharge stage .

3. According to claim 1, a method for secondary sorting of retired batteries based on random charging fragments is characterized in that: The prediction model constructed based on the two-dimensional convolutional neural network is trained in combination with the battery coding database, and the difference between the actual maximum available capacity of the battery and the estimated value is minimized in combination with the stochastic gradient descent algorithm to optimize the prediction model parameters, including: Build a prediction model based on a two-dimensional convolutional neural network, including building convolutional layers, pooling layers, and fully connected layers; The converted two-dimensional image data is input into the prediction model for training, the features in the two-dimensional image data are gradually extracted through the convolution layer and the pooling layer, and the estimated value of the maximum available capacity of the battery is output through the fully connected layer ; Use stochastic gradient descent algorithm to minimize the actual maximum available capacity With estimated value The difference between them is used to optimize the prediction model parameters and obtain the prediction model that has completed training.

4. A secondary sorting system for retired batteries based on random charging fragments, applied to perform the sorting method described in any one of claims 1 to 3, characterized in that: The system includes a code scanning module, a charge and discharge control module, a data acquisition module, a data processing module and a sorting module, wherein: The code scanning module is used to generate an identification code for each single battery and transmit it to a battery code database pre-built in the data processing module to store the battery identity code; The charge and discharge control module is used to control the charge and discharge of the single battery based on a preset charge and discharge strategy and preset working conditions; The data acquisition module is used to obtain a time series sample data set of a single battery; wherein the time series sample data is determined according to a temperature change time series, a voltage change time series and a capacity change time series under the charge and discharge control scenario; The data processing module includes a processor and a memory; The processor is used to convert the time series sample data set into two-dimensional image data, establish a corresponding relationship between the identification code, the single battery and the two-dimensional image data in the battery coding database; and train the prediction model built based on the two-dimensional convolutional neural network in combination with the battery coding database, and minimize the difference between the actual maximum available capacity of the battery and the estimated value in combination with the stochastic gradient descent algorithm to optimize the prediction model parameters; The memory is used to store the battery coding database, and is used to store the two-dimensional image data in the battery coding database; The sorting module is used to input the two-dimensional image data of the target single battery to be sorted into the trained prediction model, output the estimated value of the maximum available capacity of the battery, and perform secondary sorting of the battery based on the estimated value of the maximum available capacity.

5. A secondary sorting system for retired batteries based on random charging segments according to claim 4, characterized in that: The charge and discharge control module comprises: A parameter setting unit, used to set a time period with the same time interval and time starting point as a charging and discharging time interval; The control unit is used to perform constant current charging and discharging control on the single battery based on the time interval.

6. A secondary sorting system for retired batteries based on random charging segments according to claim 4, characterized in that: The data acquisition module comprises: The first acquisition unit is used to acquire the temperature change time series , record the temperature change during the set charging phase, ,in, Indicates time, from arrive ; The second acquisition unit is used to acquire the voltage change sequence : Record the voltage change during the set charging stage, ,in, Indicates time, from arrive ; The third acquisition unit is used to obtain the capacity change time series : Record the cumulative charging capacity change time series within the set charging stage ,in, Indicates time, from arrive ; Data integration unit, used to convert temperature change time series , voltage change time series and capacity change time series Normalize to get the normalized sequence and , get the time series sample data set of the set charge and discharge stage .

7. An electronic device, characterized in that: include: Processor unit; a memory unit for storing instructions executable by the processor unit; The processor unit is configured to execute the instructions to implement the retired battery secondary sorting method according to any one of claims 1 to 3.

8. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor unit of an electronic device, the electronic device is enabled to perform the retired battery secondary sorting method as claimed in any one of claims 1 to 3.

Images