A method and device for sorting retired battery modules, an electronic device, and a storage medium

Abstract

This invention discloses a method, apparatus, electronic device, and storage medium for sorting retired power battery modules. The method includes: acquiring the remaining capacity of several retired power battery modules under a first charge-discharge test; acquiring pulse charge-discharge data of the battery modules during the last pulse charge-discharge after three consecutive pulse charge-discharge cycles; calculating the first characteristic parameter, second characteristic parameter, terminal voltage, and internal resistance of the several retired power battery modules based on the pulse charge-discharge data of the last pulse charge-discharge, and performing dimensionality reduction processing to obtain an m-dimensional feature matrix; finally, constructing a neural network model, training the neural network model to obtain a capacity prediction model, predicting the capacity of the retired power battery modules based on the capacity prediction model, and then sorting the battery modules. This invention solves the technical problem that the current inability to accurately predict the capacity of battery modules leads to the ineffective sorting of retired battery modules.

Description

Technical Field

[0001] This invention relates to the field of battery sorting technology, specifically to a method, apparatus, electronic device, and storage medium for sorting retired battery modules. Background Technology

[0002] With the rise of the electric vehicle industry, the sales of electric vehicles are increasing, which in turn drives a sharp increase in the sales of lithium-ion batteries. For electric vehicles, in order to ensure the driving range, when the battery capacity drops to a certain level (e.g., 80%), it needs to be retired from the vehicle. Directly retiring the battery would waste its remaining value. Therefore, it is necessary to reuse retired battery modules to avoid energy waste.

[0003] However, retired battery modules exhibit significant performance differences after undergoing various charge and discharge processes. If they are used directly without being sorted, it will affect the capacity and power performance of the battery for secondary use, as well as the remaining lifespan of the battery pack. Furthermore, retired battery modules are more likely to malfunction.

[0004] Current sorting methods generally involve estimating the battery module capacity and then sorting it based on that capacity. However, current battery module capacity prediction requires a complete charge and discharge process, and the capacity estimation is not accurate enough, thus failing to effectively sort retired battery modules. Summary of the Invention

[0005] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide a method, apparatus, electronic device and storage medium for sorting retired battery modules, thereby solving the technical problem in the prior art that the inability to effectively sort retired battery modules is due to the inability to accurately estimate the capacity of battery modules.

[0006] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for sorting retired battery modules, comprising the following steps:

[0008] Obtain the remaining capacity of several retired power battery modules under their first charge-discharge test;

[0009] The pulse charge-discharge data of several retired power battery modules are obtained at the last pulse charge-discharge after three consecutive pulse charge-discharge cycles. The first characteristic parameter of several retired power battery modules is calculated based on the pulse charge-discharge data at the last pulse charge-discharge.

[0010] The difference between the sum of the first characteristic parameters of each individual cell of a plurality of retired power battery modules at the last pulse charge and discharge and the first characteristic parameters of the plurality of retired power battery modules is obtained, and the difference is used as the second characteristic parameter.

[0011] Obtain the terminal voltage and internal resistance of several retired power battery modules;

[0012] The first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules are reduced in dimension to obtain a feature matrix with m dimensions after dimension reduction.

[0013] A neural network model is constructed, with the m-dimensional feature matrix as input and the discharge capacity of the retired power battery module as output. The neural network model is trained to obtain a capacity prediction model. After predicting the capacity of the retired power battery module based on the capacity prediction model, the retired power battery module is sorted according to the predicted capacity, the terminal voltage and internal resistance of the retired power battery module.

[0014] In some embodiments, obtaining the remaining capacity of several retired power battery modules under the first charge-discharge test includes:

[0015] The remaining capacity of several retired power battery modules was obtained after they were left to stand at the first temperature and then charged and discharged at the first charge and discharge rate.

[0016] In some embodiments, the first characteristic parameter includes at least ohmic internal resistance, polarization internal resistance, and charge / discharge power.

[0017] In some embodiments, obtaining the pulse charge / discharge data of a plurality of retired power battery modules at the last pulse charge / discharge after three consecutive pulse charge / discharge cycles includes:

[0018] The data from the last pulse charge / discharge test was obtained after several retired power battery modules were left to stand at the second temperature and then subjected to three consecutive pulse charge / discharge tests at the second charge / discharge rate.

[0019] In some embodiments, the first temperature is 15°C to 35°C, the second temperature is 15°C to 35°C, the first charge / discharge rate is 0.3C to 1C, and the second charge / discharge rate is 1C to 3C.

[0020] In some embodiments, principal component analysis is used to reduce the dimensionality of the first characteristic parameter, second characteristic parameter, terminal voltage, and internal resistance of several power battery modules.

[0021] In some embodiments, the neural network model is a support vector machine.

[0022] Secondly, the present invention also provides a retired battery module sorting device, comprising:

[0023] The remaining capacity acquisition module is used to acquire the remaining capacity of several retired power battery modules under the first charge and discharge test;

[0024] The first feature parameter acquisition module is used to acquire pulse charge and discharge data of several retired power battery modules during the last pulse charge and discharge after three consecutive pulse charge and discharge cycles, and to calculate the first feature parameter of several retired power battery modules based on the pulse charge and discharge data during the last pulse charge and discharge.

[0025] The second feature parameter acquisition module is used to acquire the difference between the sum of the first feature parameters of each individual cell of a plurality of retired power battery modules during the last pulse charge and discharge and the first feature parameters of the plurality of retired power battery modules, and to use the difference as the second feature parameter.

[0026] The third feature parameter acquisition module is used to acquire the terminal voltage and internal resistance of several retired power battery modules;

[0027] The dimension reduction module is used to reduce the first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules to obtain a feature matrix with m dimensions after dimension reduction.

[0028] The prediction module is used to construct a neural network model. It takes the m-dimensional feature matrix as input and the discharge capacity of the retired power battery module as output, trains the neural network model to obtain a capacity prediction model, and then performs capacity prediction on the retired power battery module based on the capacity prediction model. Finally, it sorts the retired power battery module according to the predicted capacity, the terminal voltage and internal resistance of the retired power battery module.

[0029] Thirdly, the present invention also provides an electronic device, comprising: a processor and a memory;

[0030] The memory stores a computer-readable program that can be executed by the processor;

[0031] When the processor executes the computer-readable program, it implements the steps in the decommissioned battery module sorting method described above.

[0032] Fourthly, the present invention also provides a computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement the steps in the retired battery module sorting method described above.

[0033] Compared with existing technologies, the retired power battery module sorting method, device, electronic equipment, and storage medium provided by this invention perform uniform processing on the initial state of retired battery modules, which can better utilize the parameter differences during the testing process to predict the capacity of retired power battery modules. By using the characteristic parameter values ​​between multiple reaction modules and individual cells obtained by pulse testing, the internal structural characteristics of the module during charging and discharging can be better reflected, and the capacity of the module can be predicted more accurately. By performing dimensionality reduction processing on the characteristic parameters, the amount of information can be reduced, the mutual influence between original data can be eliminated, and the computational load of the model can be reduced. In addition, the more training set samples there are, the more accurate the prediction model will be. Through the technical valve of this invention, the rapid and accurate prediction of retired power battery modules can be achieved without complete charging and discharging. Attached Figure Description

[0034] Figure 1 This is a flowchart of an embodiment of the method for sorting retired power battery modules provided by the present invention;

[0035] Figure 2 This is a schematic diagram of an embodiment of the retired power battery module sorting device provided by the present invention;

[0036] Figure 3 This is a schematic diagram of the operating environment of an embodiment of the retired power battery module sorting program of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0038] Please see Figure 1 The method for sorting retired power battery modules provided in this embodiment of the invention includes the following steps:

[0039] S100: Obtain the remaining capacity of several retired power battery modules under the first charge-discharge test;

[0040] S200. Obtain pulse charge-discharge data of a number of retired power battery modules during the last pulse charge-discharge after three consecutive pulse charge-discharge cycles, and calculate the first characteristic parameter of the number of retired power battery modules based on the pulse charge-discharge data during the last pulse charge-discharge.

[0041] S300: Obtain the difference between the sum of the first characteristic parameters of each individual cell of a plurality of retired power battery modules during the last pulse charge and discharge and the first characteristic parameters of the plurality of retired power battery modules, and use the difference as the second characteristic parameter;

[0042] S400: Obtain the terminal voltage and internal resistance of several retired power battery modules;

[0043] S500 performs dimensionality reduction processing on the first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules to obtain a feature matrix with m dimensions after dimensionality reduction.

[0044] S600. Construct a neural network model, using the m-dimensional feature matrix as input and the discharge capacity of the retired power battery module as output. Train the neural network model to obtain a capacity prediction model. Based on the capacity prediction model, predict the capacity of the retired power battery module.

[0045] Based on the predicted capacity and the terminal voltage and internal resistance of the retired power battery modules, the retired power battery modules were sorted.

[0046] In this embodiment, the initial state of the retired battery module is standardized, which allows for better utilization of parameter differences during testing to predict the capacity of the retired power battery module. Furthermore, by using the characteristic parameter values ​​between multiple reaction modules and individual cells obtained from pulse testing, a better understanding of the reaction module's characteristics can be achieved.

[0047] Understanding the internal structural characteristics of the module during charging and discharging allows for more accurate prediction of its capacity. Dimensionality reduction of the zero-feature parameters reduces the amount of information and eliminates the mutual influence between original data.

[0048] The invention reduces the computational load of the model. Furthermore, the more samples in the training set, the more accurate the prediction model becomes. Through the technical valve stem of this invention, rapid and accurate prediction of retired power battery modules can be achieved without the need for complete charging and discharging.

[0049] In some embodiments, step S100 specifically includes: 5. Obtaining a plurality of retired power battery modules, allowing them to stand at a first temperature, and then charging and discharging them at a first charge / discharge rate.

[0050] The remaining capacity after a charge-discharge test.

[0051] In this embodiment, after several retired power battery modules to be sorted are left to stand at room temperature for a sufficiently long time, a charge-discharge test is performed at the first charge-discharge rate to obtain the retired power battery modules.

[0052] The remaining capacity of the group. The first temperature is 15℃~35℃, and the first charge / discharge rate is 0.3C~1C. Preferably, the first temperature is 25℃, and the first charge / discharge rate is 1C.

[0053] In some embodiments, the first characteristic parameter includes at least ohmic internal resistance, polarization internal resistance, and charge / discharge power.

[0054] In some embodiments, step S200, which involves acquiring pulse charge-discharge data of a plurality of retired power battery modules during the last pulse charge-discharge after three consecutive pulse charge-discharge cycles, includes:

[0055] After several retired power battery modules were left to stand at the second temperature, they underwent three consecutive pulse charge-discharge tests at the second charge-discharge rate of 5. The data from the last pulse charge-discharge test was obtained.

[0056] In this embodiment, several retired power battery modules are adjusted to the same state at 1C (0.3C~1C), and then subjected to three consecutive pulse discharge and pulse charge tests at a second temperature and a second charge / discharge rate. The ohmic internal resistance (R) of the module is calculated using the pulse charge / discharge data from the last test. n1 R n2 ), polarization resistance (R) f1 R f2 The charging and discharging power (P1, P2) is used as the first characteristic parameter. Wherein, the second temperature is 15℃~35℃, and the second charging and discharging rate is 1C~3C. Preferably,

[0057] The second temperature is 25°C, and the second charge / discharge rate is 1C.

[0058] In some embodiments, after three consecutive pulse discharge and pulse charge tests, the ohmic internal resistance, polarization internal resistance, and charge / discharge power of each individual cell during the last pulse charge / discharge are calculated.

[0059] The difference between the sum of the values ​​and the value of the retired power battery module is denoted as ΔR. n1 ΔR n2 ΔR f1 ΔR f2 5ΔP1 and ΔP2 are used as the second characteristic parameters.

[0060] Furthermore, after obtaining the first and second characteristic parameters, the retired power battery module after the pulse is left to stand for a sufficient period of time, and then the terminal voltage and internal resistance of the retired power battery module are tested using a voltage and internal resistance tester to obtain the terminal voltage U and internal resistance R as characteristic parameters.

[0061] 0 In some embodiments, principal component analysis is used to analyze the first characteristics of several power battery modules.

[0062] The first characteristic parameter, the second characteristic parameter, the terminal voltage, and the internal resistance are reduced in dimension.

[0063] In this embodiment, Principal Component Analysis (PCA) is used to reduce the dimensionality of the 14 sets of feature parameters obtained. The original features are projected onto the dimensions with the largest amount of projection information as much as possible. By projecting the original features onto these dimensions, the loss of information after dimensionality reduction is minimized, and the mutual influence between the original data components can be eliminated.

[0064] Specifically, suppose there are X sample sizes {x1, x2, ... x}. n Each sample has k-dimensional features x1 = {x} 11 x 12 , ...x 1k To reduce the dimensionality to m dimensions, the dataset is first de-averaged by subtracting the average value from each feature. Then, the covariance matrix of the features is calculated to obtain the eigenvalues ​​and eigenvectors of the covariance matrix. The eigenvectors are arranged in rows from top to bottom according to the corresponding eigenvalues ​​to form a matrix. The first m rows are taken to form matrix P, which is the feature matrix after dimensionality reduction to m dimensions.

[0065] In some embodiments, the neural network model is a support vector machine.

[0066] In this embodiment, Support Vector Machine (SVM) is a supervised learning model with relevant learning algorithms. SVM is used to reflect the mapping relationship between the m-dimensional feature matrix and the module discharge capacity, establishing a battery module capacity prediction model. The feature matrix serves as the input value of the capacity prediction model, and the module discharge capacity is the output value. Specifically, training and prediction sets are set up. The training set contains at least 20 modules, and the prediction set contains at least 10 modules. KS-test is used to verify the identical distribution of the training and test sets, thus validating the rationality of the training and test set partitioning. Finally, the battery module capacity prediction model is trained.

[0067] Finally, the battery modules are sorted based on the predicted capacity, module terminal voltage, and internal resistance, thereby achieving rapid sorting.

[0068] The technical solution provided by this invention standardizes the initial state of retired battery modules, enabling better capacity prediction of retired power battery modules by utilizing parameter differences during testing. By using the characteristic parameter values ​​between multiple reaction modules and individual cells obtained from pulse testing, the internal structural characteristics of the module during charging and discharging can be better reflected, resulting in more accurate capacity prediction. By reducing the dimensionality of the characteristic parameters, the amount of information can be reduced, eliminating the mutual influence between original data and reducing the computational load of the model. Furthermore, the larger the number of training set samples, the more accurate the prediction model will be. Through the technical valve of this invention, rapid and accurate prediction of retired power battery modules can be achieved without complete charging and discharging.

[0069] Based on the above-described method for monitoring bird damage on power transmission lines, this invention also provides a corresponding bird damage monitoring device 700 for power transmission lines. Please refer to [link to relevant documentation]. Figure 2 The retired battery module sorting device 700 includes a remaining capacity acquisition module 710, a first feature parameter acquisition module 720, a second feature parameter acquisition module 730, a third feature parameter acquisition module 740, a dimensionality reduction processing module 750, and a prediction module 760.

[0070] The remaining capacity acquisition module 710 is used to acquire the remaining capacity of several retired power battery modules under the first charge and discharge test.

[0071] The first feature parameter acquisition module 720 is used to acquire pulse charge and discharge data of several retired power battery modules during the last pulse charge and discharge after three consecutive pulse charge and discharge cycles, and to calculate the first feature parameters of several retired power battery modules based on the pulse charge and discharge data during the last pulse charge and discharge.

[0072] The second feature parameter acquisition module 730 is used to acquire the difference between the sum of the first feature parameters of each individual cell of a plurality of retired power battery modules during the last pulse charge and discharge and the first feature parameters of the plurality of retired power battery modules, and to use the difference as the second feature parameter.

[0073] The third feature parameter acquisition module 740 is used to acquire the terminal voltage and internal resistance of several retired power battery modules.

[0074] The dimension reduction processing module 750 is used to perform dimension reduction processing on the first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules to obtain a dimension-reduced m-dimensional feature matrix.

[0075] The prediction module 760 is used to construct a neural network model. The model is trained with the m-dimensional feature matrix as input and the discharge capacity of the retired power battery module as output to obtain a capacity prediction model. After predicting the capacity of the retired power battery module based on the capacity prediction model, the retired power battery module is sorted according to the predicted capacity, the terminal voltage and internal resistance of the retired power battery module.

[0076] In this embodiment, the initial state of the retired battery module is standardized, which can better utilize the parameter differences during the testing process to predict the capacity of the retired power battery module. By using the characteristic parameter values ​​between multiple reaction modules and individual cells obtained by pulse testing, the internal structural characteristics of the module during charging and discharging can be better reflected, and the capacity of the module can be predicted more accurately. By reducing the dimensionality of the characteristic parameters, the amount of information can be reduced, the mutual influence between the original data can be eliminated, and the computational load of the model can be reduced. In addition, the more training set samples there are, the more accurate the prediction model will be. Through the technical valve of this invention, the rapid and accurate prediction of retired power battery modules can be achieved without complete charging and discharging.

[0077] In some embodiments, the remaining capacity acquisition module 710 is specifically used for:

[0078] The remaining capacity of several retired power battery modules was obtained after they were left to stand at the first temperature and then charged and discharged at the first charge and discharge rate.

[0079] In some embodiments, the first characteristic parameter includes at least ohmic internal resistance, polarization internal resistance, and charge / discharge power.

[0080] In some embodiments, obtaining the pulse charge / discharge data of a plurality of retired power battery modules at the last pulse charge / discharge after three consecutive pulse charge / discharge cycles includes:

[0081] The data from the last pulse charge / discharge test was obtained after several retired power battery modules were left to stand at the second temperature and then subjected to three consecutive pulse charge / discharge tests at the second charge / discharge rate.

[0082] In some embodiments, the first temperature is 15°C to 35°C, the second temperature is 15°C to 35°C, the first charge / discharge rate is 0.3C to 1C, and the second charge / discharge rate is 1C to 3C.

[0083] In some embodiments, principal component analysis is used to reduce the dimensionality of the first characteristic parameter, second characteristic parameter, terminal voltage, and internal resistance of several power battery modules.

[0084] In some embodiments, the neural network model is a support vector machine.

[0085] like Figure 3 As shown, based on the above-described method for sorting retired battery modules, the present invention also provides an electronic device, which can be a mobile terminal, desktop computer, laptop, handheld computer, server, or other computing device. The electronic device includes a processor 10, a memory 20, and a display 30. Figure 3Only some components of the electronic device are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0086] In some embodiments, memory 20 may be an internal storage unit of the electronic device, such as a hard disk or memory. In other embodiments, memory 20 may be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. Furthermore, memory 20 may include both internal and external storage units. Memory 20 is used to store application software and various types of data installed on the electronic device, such as program code installed on the electronic device. Memory 20 may also be used to temporarily store data that has been output or will be output. In one embodiment, memory 20 stores a retired battery module sorting program 40, which can be executed by processor 10 to implement the retired battery module sorting methods of the embodiments of this application.

[0087] In some embodiments, processor 10 may be a central processing unit (CPU), microprocessor or other data processing chip, used to run program code stored in memory 20 or process data, such as executing a method for sorting retired battery modules.

[0088] In some embodiments, display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 30 is used to display information from the decommissioned battery module sorting device and to display a user interface for visualization. Components 10-30 of the electronic device communicate with each other via a system bus.

[0089] In one embodiment, when the processor 10 executes the retired battery module sorting program 40 in the memory 20, it implements the steps in the retired battery module sorting method as described in the above embodiments. Since the retired battery module sorting method has been described in detail above, it will not be repeated here.

[0090] In summary, the retired battery module sorting method, apparatus, electronic device, and storage medium provided by this invention unify the initial state of retired battery modules, enabling better capacity prediction of retired power battery modules by utilizing parameter differences during testing. By utilizing the characteristic parameter values ​​between multiple reaction modules and individual cells obtained from pulse testing, the internal structural characteristics of the module during charging and discharging can be better reflected, leading to more accurate capacity prediction. Dimensionality reduction of the characteristic parameters reduces the amount of information, eliminates mutual influence between original data, and reduces model computation. Furthermore, the larger the training set sample size, the more accurate the established prediction model. Through the technical lever of this invention, rapid and accurate prediction of retired power battery modules can be achieved without requiring a complete charge and discharge cycle.

[0091] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The storage medium can be a memory, magnetic disk, optical disk, etc.

[0092] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for sorting retired battery modules, characterized in that, Includes the following steps: Obtain the remaining capacity of several retired power battery modules under their first charge-discharge test; The pulse charge-discharge data of several retired power battery modules are obtained at the last pulse charge-discharge after three consecutive pulse charge-discharge cycles. The first characteristic parameter of several retired power battery modules is calculated based on the pulse charge-discharge data at the last pulse charge-discharge. The difference between the sum of the first characteristic parameters of each individual cell of a plurality of retired power battery modules at the last pulse charge and discharge and the first characteristic parameters of the plurality of retired power battery modules is obtained, and the difference is used as the second characteristic parameter. Obtain the terminal voltage and internal resistance of several retired power battery modules; The first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules are reduced in dimension to obtain a feature matrix with m dimensions after dimension reduction. A neural network model is constructed, with the m-dimensional feature matrix as input and the discharge capacity of the retired power battery module as output. The neural network model is trained to obtain a capacity prediction model. After predicting the capacity of the retired power battery module based on the capacity prediction model, the retired power battery module is sorted according to the predicted capacity, the terminal voltage and internal resistance of the retired power battery module. The first characteristic parameter includes at least ohmic internal resistance, polarization internal resistance, and charge / discharge power.

2. The method for sorting retired battery modules according to claim 1, characterized in that, The process of obtaining the remaining capacity of several retired power battery modules under the first charge-discharge test includes: The remaining capacity of several retired power battery modules was obtained after they were left to stand at the first temperature and then charged and discharged at the first charge and discharge rate.

3. The method for sorting retired battery modules according to claim 2, characterized in that, The acquisition of pulse charge / discharge data of several retired power battery modules during the last pulse charge / discharge after three consecutive pulse charge / discharge cycles includes: The data from the last pulse charge / discharge test was obtained after several retired power battery modules were left to stand at the second temperature and then subjected to three consecutive pulse charge / discharge tests at the second charge / discharge rate.

4. The method for sorting retired battery modules according to claim 3, characterized in that, The first temperature is 15℃~35℃, the second temperature is 15℃~35℃, the first charge / discharge rate is 0.3C~1C, and the second charge / discharge rate is 1C~3C.

5. The method for sorting retired battery modules according to claim 4, characterized in that, Principal component analysis was used to reduce the dimensionality of the first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules.

6. The method for sorting retired battery modules according to claim 5, characterized in that, The neural network model is a support vector machine.

7. A sorting device for retired battery modules, characterized in that, include: The remaining capacity acquisition module is used to acquire the remaining capacity of several retired power battery modules under the first charge and discharge test; The first feature parameter acquisition module is used to acquire pulse charge and discharge data of several retired power battery modules during the last pulse charge and discharge after three consecutive pulse charge and discharge cycles, and to calculate the first feature parameter of several retired power battery modules based on the pulse charge and discharge data during the last pulse charge and discharge. The second feature parameter acquisition module is used to acquire the difference between the sum of the first feature parameters of each individual cell of a plurality of retired power battery modules during the last pulse charge and discharge and the first feature parameters of the plurality of retired power battery modules, and to use the difference as the second feature parameter. The third feature parameter acquisition module is used to acquire the terminal voltage and internal resistance of several retired power battery modules; The dimension reduction module is used to reduce the first characteristic parameter, second characteristic parameter, terminal voltage and internal resistance of several power battery modules to obtain a feature matrix with m dimensions after dimension reduction. The prediction module is used to construct a neural network model. The model is trained with the m-dimensional feature matrix as input and the discharge capacity of the retired power battery module as output to obtain a capacity prediction model. After predicting the capacity of the retired power battery module based on the capacity prediction model, the retired power battery module is sorted according to the predicted capacity, the terminal voltage and internal resistance of the retired power battery module. The first characteristic parameter includes at least ohmic internal resistance, polarization internal resistance, and charge / discharge power.

8. An electronic device, characterized in that, include: Processor and memory; The memory stores a computer-readable program that can be executed by the processor; When the processor executes the computer-readable program, it implements the steps in the method for sorting retired battery modules as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which can be executed by one or more processors to implement the steps in the method for sorting retired battery modules as described in any one of claims 1-6.

Images