Battery sorting method and battery automatic sorting equipment

Abstract

The present application relates to the field of battery management technology, and to a battery sorting method and automatic battery sorting equipment. The battery sorting method includes: performing an electrochemical impedance spectrum test on the battery to be sorted to obtain an electrochemical impedance spectrum corresponding to the battery to be sorted; performing equivalent circuit fitting on the electrochemical impedance spectrum to obtain a first evaluation parameter; obtaining a second evaluation parameter corresponding to the battery to be sorted, the second evaluation parameter is directly obtained based on the battery to be sorted, and the second evaluation parameter is used to indicate the physical and chemical properties of the battery to be sorted; the first evaluation parameter and the second evaluation parameter are input into a pre-trained aging prediction model to obtain the degree of aging corresponding to the battery to be sorted; and the sorting result corresponding to the battery to be sorted is determined based on the degree of aging. By adopting the above technical solution, the degree of aging of the battery can be judged based on the pre-trained aging prediction model by comprehensively considering factors of different dimensions to achieve battery sorting, which can help improve the accuracy of the battery sorting results.

Description

Technical Field

[0001] The present application relates to the field of battery management technology, and in particular to a battery sorting method and automatic battery sorting equipment. Background Art

[0002] In new energy vehicles, batteries are not used individually, but dozens or hundreds of batteries are connected in series and parallel to form a battery pack. Therefore, the status of each battery in the pack needs to be kept consistent as much as possible, otherwise the failure of one battery may cause the failure of the entire pack. Therefore, before assembling into a pack, a battery sorting machine is needed to classify the batteries, and only batteries with good consistency are used for grouping and pack assembly to ensure the service life of the pack.

[0003] The current battery sorting machines mainly test the open circuit voltage (OCV) and 1000Hz internal resistance of the battery. The sorting results are relatively rough and do not effectively sort out inconsistent batteries, resulting in the consistency of the pack still being a problem. Summary of the invention

[0004] In order to help improve the battery sorting effect, the present application provides a battery sorting method and an automatic battery sorting device.

[0005] In a first aspect, the present application provides a battery sorting method, which adopts the following technical solution:

[0006] A battery sorting method, the method comprising:

[0007] Performing an electrochemical impedance spectroscopy test on the battery to be sorted to obtain an electrochemical impedance spectroscopy corresponding to the battery to be sorted;

[0008] Performing equivalent circuit fitting on the electrochemical impedance spectrum to obtain a first evaluation parameter;

[0009] Acquire a second evaluation parameter corresponding to the battery to be sorted, where the second evaluation parameter is directly obtained based on the battery to be sorted, and the second evaluation parameter is used to indicate the physical and chemical properties of the battery to be sorted;

[0010] Inputting the first evaluation parameter and the second evaluation parameter into a pre-trained aging prediction model to obtain an aging degree corresponding to the battery to be sorted;

[0011] A sorting result corresponding to the battery to be sorted is determined based on the aging degree.

[0012] By adopting the above technical solution, the first evaluation parameter can be determined based on the electrochemical impedance spectrum corresponding to the battery to be sorted, and the first evaluation parameter and the second evaluation parameter directly obtained based on the battery to be sorted can be comprehensively judged through the pre-trained aging model, and finally the aging degree corresponding to the battery to be sorted can be obtained. In this way, the aging degree of the battery can be judged by comprehensively considering factors of different dimensions, which can help improve the accuracy of the battery sorting results.

[0013] Optionally, the step of inputting the first evaluation parameter and the second evaluation parameter into a pre-trained aging prediction model to obtain an aging degree corresponding to the battery to be sorted includes:

[0014] Generate and determine two third evaluation parameters based on the first evaluation parameter and the second evaluation parameter;

[0015] The third evaluation parameter is input into the pre-trained aging model to obtain the aging degree corresponding to the battery to be sorted.

[0016] By adopting the above technical solution, the third evaluation parameter can be comprehensively determined based on the first evaluation parameter and the second evaluation parameter, and the aging degree can be predicted by the third evaluation parameter, which can enhance the model's understanding and prediction capabilities of battery aging.

[0017] Optionally, the first evaluation parameter includes a charge transfer resistance R CT , the second evaluation parameter includes the health state SOH, and the generating at least two third evaluation parameters based on the first evaluation parameter and the second evaluation parameter includes:

[0018] Based on the charge transfer resistance R CT The product of the health state SOH determines the third evaluation parameter.

[0019] By adopting the above technical solutions, the model can capture the intrinsic connection between the charge transfer process and the overall health status of the battery, thereby improving the accuracy of aging prediction results.

[0020] Optionally, the second evaluation parameter also includes the internal temperature, the charge transfer resistance R CT The third evaluation parameter is determined by multiplying the product of the health state SOH, including:

[0021] Determining a temperature coefficient corresponding to the internal temperature based on a preset temperature function;

[0022] Based on the temperature coefficient, the charge transfer resistor R CT The third evaluation parameter is obtained by correcting the product of the health state SOH.

[0023] By adopting the above technical solution, the charge transfer resistance R can be adjusted based on the temperature coefficient corresponding to the internal temperature. CT The third evaluation parameter is obtained by correcting the product of the health state SOH, which can help reduce the influence of temperature on resistance detection, and further help improve the accuracy of the third evaluation parameter.

[0024] Optionally, the loss function used in the aging prediction model training process includes a prediction error term and a feature interaction term, the prediction error term is used to measure the difference between the aging degree predicted by the model and the actual aging degree, and the feature interaction term is used to measure the difference between the aging degrees predicted by different third evaluation parameters.

[0025] By adopting the above technical solution, the learning effect of the model on different third evaluation parameters can be balanced through the interactive error term in the loss function, which can help improve the accuracy of the model prediction results.

[0026] Optionally, the loss function is expressed by the following formula:

[0027]

[0028] Wherein, L is the loss value; N is the total number of samples in the data set; α is the weight coefficient corresponding to the prediction error term; is the predicted value of the i-th group of samples; Y i is the true value of the i-th group of samples; p is a hyperparameter used to adjust the loss degree of the prediction error term; is the weight coefficient corresponding to the feature interaction term; j and k are feature indexes used to traverse different combinations of the third evaluation parameters, j is greater than or equal to 1, and k is less than or equal to the number of the third evaluation parameters; φ jk is the interaction weight between the j-th third evaluation parameter and the k-th third evaluation parameter; is the predicted aging degree corresponding to the jth third evaluation parameter in the i-th group of samples; is the predicted aging degree corresponding to the kth third evaluation parameter in the i-th group of samples; q is a hyperparameter used to adjust the loss degree of the feature interaction term.

[0029] By adopting the above technical solution, the degree of learning of the model for different features can be balanced by reasonably setting the interaction weights between different third evaluation parameters, so that the model can better learn the impact of different features on the degree of aging, which can help improve the accuracy of aging judgment results.

[0030] Optionally, the interaction weights φ between different third evaluation parameters jk Initialized before training and updated during training.

[0031] By adopting the above technical solution, the interaction weights between the third evaluation parameters can be dynamically adjusted during the model training process, thereby dynamically adjusting the impact of the differences in predicted aging degrees corresponding to the third evaluation parameters on the model training results, improving the training effect of the model, and further helping to improve the accuracy of aging prediction.

[0032] Optionally, the first evaluation parameter includes an ohmic resistance R S , SEI film resistance R SEI And the charge transfer resistance R CT At least one of;

[0033] The second evaluation parameter includes at least one of the state of charge SOC, the state of health SOH, the internal temperature and the internal short circuit of the battery cell.

[0034] In the second aspect, the present application provides an automatic battery sorting device, which adopts the following technical solution:

[0035] A battery automatic sorting device, the device comprising at least one test station, the test station being used to accommodate batteries to be tested, each of the test stations comprising a control component and a first measurement component and a second measurement component connected to the control component by signal;

[0036] The first measuring component is used to perform an electrochemical impedance spectroscopy test on the battery to be tested at the testing station to obtain an electrochemical impedance spectrum, the second measuring component is used to collect a second evaluation parameter corresponding to the battery to be tested at the testing station, and the control component is used to execute any one of the battery sorting methods provided in the first aspect.

[0037] In a third aspect, the present application provides an electronic device, which adopts the following technical solution:

[0038] An electronic device, comprising:

[0039] at least one processor;

[0040] Memory;

[0041] At least one application, wherein the at least one application is stored in a memory and configured to be executed by at least one processor, and the at least one application is configured to: execute any one of the battery sorting methods provided in the first aspect.

[0042] In summary, the present application includes at least one of the following beneficial technical effects:

[0043] 1. The first evaluation parameter can be determined based on the electrochemical impedance spectrum corresponding to the battery to be sorted, and the first evaluation parameter and the second evaluation parameter directly obtained based on the battery to be sorted can be comprehensively judged through the pre-trained aging model, and finally the aging degree corresponding to the battery to be sorted is obtained, thereby realizing the sorting of the battery. In this way, the aging degree of the battery can be judged by comprehensively considering factors of different dimensions, which can help improve the accuracy of the battery sorting results;

[0044] 2. An interactive error term is added to the loss function to measure the difference between the aging degrees predicted by different third evaluation parameters. In this way, the learning effect of the model on different third evaluation parameters can be balanced through the interactive error term, which can help improve the accuracy of the model prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of a process of a battery sorting method provided in an embodiment of the present application;

[0046] Figure 2 is a schematic diagram of an equivalent circuit model provided in an embodiment of the present application;

[0047] Figure 3 It is a flow chart of an aging degree prediction method provided in an embodiment of the present application;

[0048] Figure 4 It is a structural schematic diagram of an automatic battery sorting device provided in an embodiment of the present application;

[0049] Figure 5 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1-5 It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0051] The present application embodiment discloses a battery sorting method. Figure 1 , the battery sorting method comprises the following steps:

[0052] Step 101 , performing an electrochemical impedance spectroscopy test on the battery to be sorted to obtain an electrochemical impedance spectrum corresponding to the battery to be sorted.

[0053] Among them, the electrochemical impedance spectroscopy test, also known as the EIS (Electrochemical Impedance Spectroscopy) test, is an important technology for analyzing electrochemical systems, especially battery electrode processes and interface characteristics. Through the electrochemical impedance spectroscopy, we can analyze the kinetic characteristics of the battery level, evaluate the battery health state (SOH), and monitor the defects inside the battery, such as internal short circuits, which can help sort the batteries.

[0054] There are many ways to obtain electrochemical impedance spectra, such as constructing an electrochemical impedance spectrum based on the time domain information or frequency domain information of the battery in a machine learning manner, which will not be described in detail in this embodiment.

[0055] Step 102: Perform equivalent circuit fitting on the electrochemical impedance spectrum to obtain a first evaluation parameter.

[0056] Among them, equivalent circuit fitting refers to matching the impedance data of a battery or other electrochemical system with an equivalent circuit model composed of resistance (R), capacitance (C) and constant phase element (CPE), so as to extract information from the electrochemical impedance spectrum that can be used to evaluate battery performance and use it as the first evaluation parameter.

[0057] Optionally, the first evaluation parameter includes an ohmic resistance R S , SEI film resistance R SEI And the charge transfer resistance R CT At least one of .

[0058] Specifically, the ohmic resistance R S It refers to the resistance of the electrolyte and electrode materials inside the battery to the flow of current. During the battery cycle, due to the degradation of the electrolyte and the structural changes of the electrode materials, the ohmic resistance R S may gradually increase, so the ohmic resistance R S It can reflect the performance of the battery to a certain extent.

[0059] SEI film resistance R SEI Refers to the resistance of the solid electrolyte interface (SEI) film to the flow of current. The increase in SEI film resistance will lead to a decrease in the charge and discharge efficiency of the battery because more energy is used to overcome the resistance of the SEI film. Therefore, the SEI film resistance R SEI It can reflect the performance of the battery to a certain extent.

[0060] Charge transfer resistance R CT It refers to the resistance when charge transfer (such as electron transfer or ion transfer) occurs at the interface between the electrode and the electrolyte. Charge transfer resistance R CT The increase of charge transfer resistance R CTIt can reflect the performance of the battery to a certain extent.

[0061] In one example, the equivalent circuit model reference Figure 2 .

[0062] Step 103, obtaining a second evaluation parameter corresponding to the battery to be sorted.

[0063] The second evaluation parameter is directly obtained based on the battery to be sorted, and the second evaluation parameter is used to indicate the physical and chemical properties of the battery to be sorted. The type of the second evaluation parameter is different from the type of the first evaluation parameter. In one example, the second evaluation parameter is directly measured from the battery to be sorted.

[0064] Optionally, the second evaluation parameter includes at least one of the state of charge SOC, state of health SOH, internal temperature and internal short circuit of the battery cell.

[0065] Among them, the state of charge SOC refers to the ratio of the current remaining capacity of the battery to its rated capacity. The state of charge SOC is a key parameter for determining the available capacity of the battery. In actual implementation, the state of charge SOC of the battery to be sorted can be measured by the open circuit voltage (OCV) method and other methods.

[0066] The health state SOH refers to the ratio of the current maximum available capacity of the battery to its initial rated capacity. The health state SOH is an indicator to measure the state of the battery relative to a new battery. In actual implementation, the health state SOH can be determined by measuring the charge and discharge capacity of the battery to be sorted.

[0067] Internal temperature refers to the temperature inside the battery cell. In actual implementation, the internal temperature can be measured by a temperature sensor built into the battery to be sorted, or by an infrared temperature sensor.

[0068] Internal short circuit refers to electrical contact between the positive and negative electrodes inside the battery, which may cause the battery to heat up, catch fire or even explode. In actual implementation, internal short circuit can be judged by monitoring the battery voltage, voltage or temperature.

[0069] It should be noted that, during the use of the battery, the second evaluation parameter is generally monitored by the battery management system (BMS). Therefore, in actual implementation, the second evaluation parameter corresponding to the battery can also be directly derived through the BMS system corresponding to the battery to be sorted.

[0070] Step 104: input the first evaluation parameter and the second evaluation parameter into the pre-trained aging prediction model to obtain the aging degree corresponding to the battery to be sorted.

[0071] The aging model is obtained by training the initial neural network model using training data, and the training data includes at least one group of training samples, and each group of training samples is set based on the first evaluation parameter of the sample battery, the second evaluation parameter and the aging degree of the sample battery. In actual implementation, the initial neural network model can select a convolutional neural network model (CNN), a recurrent neural network model (RNN) and other models, and this embodiment does not limit the type of the initial network neural network model.

[0072] In one example, the aging degree is represented by a numerical value, such as a percentage, and the aging degree of the battery may increase as the numerical value increases. At this time, the aging degree of a new battery is at a lower level, and the aging degree of a battery that cannot be used or has a high risk of use is at a higher level.

[0073] Step 105, determining the sorting result corresponding to the battery to be sorted based on the aging degree.

[0074] In one example, the aging degree is represented by a numerical value, and the sorting result is represented by a battery grade. At this time, the sorting result corresponding to the battery to be sorted is determined based on the aging degree, including: determining the numerical range corresponding to the aging degree; determining the battery grade corresponding to the numerical range as the battery grade corresponding to the battery to be sorted.

[0075] The numerical interval is obtained by dividing the value range of the aging degree, and the corresponding relationship between the numerical interval and the battery level is preset.

[0076] In another example, the degree of aging is represented by a numerical value, the batteries to be sorted include multiple batteries, and the sorting results are represented by a battery group. At this time, the sorting results corresponding to the batteries to be sorted are determined based on the degree of aging, including: sorting each battery to be sorted according to the degree of aging; dividing a preset number of batteries to be sorted into a group in order from low to high according to the degree of aging, and the battery group that has been divided or has the highest degree of aging is divided to obtain at least two battery groups.

[0077] The preset number is less than the number of batteries to be sorted. For example, if there are 98 batteries to be sorted and the preset number is 20, the batteries to be sorted will be divided into battery groups, the number of batteries in the first four battery groups is 20, and the number of batteries in the last group to be sorted is 18.

[0078] Furthermore, after obtaining at least two battery groups, the method further includes: for each battery group, determining whether the maximum difference in aging degree corresponding to each battery to be sorted in the battery group is greater than a preset difference threshold; if so, determining that the battery group is non-compliant. This can help ensure the reliability of battery group classification.

[0079] The implementation principle of a battery sorting method in an embodiment of the present application is as follows: perform an electrochemical impedance spectrum test on the battery to be sorted to obtain the electrochemical impedance spectrum corresponding to the battery to be sorted; perform equivalent circuit fitting on the electrochemical impedance spectrum to obtain a first evaluation parameter; obtain a second evaluation parameter corresponding to the battery to be sorted, the second evaluation parameter is directly obtained based on the battery to be sorted, and the second evaluation parameter is used to indicate the physical and chemical properties of the battery to be sorted; input the first evaluation parameter and the second evaluation parameter into a pre-trained aging prediction model to obtain the degree of aging corresponding to the battery to be sorted; determine the sorting result corresponding to the battery to be sorted based on the degree of aging. By adopting the above technical solution, the first evaluation parameter can be determined based on the electrochemical impedance spectrum corresponding to the battery to be sorted, and the first evaluation parameter and the second evaluation parameter directly obtained based on the battery to be sorted are comprehensively judged by the pre-trained aging model, and finally the degree of aging corresponding to the battery to be sorted is obtained, thereby realizing the sorting of the battery. In this way, the degree of aging of the battery can be judged by comprehensively considering factors of different dimensions, which can help improve the accuracy of the battery sorting results.

[0080] In some embodiments, reference Figure 3 Step 104, inputting the first evaluation parameter and the second evaluation parameter into the pre-trained aging prediction model to obtain the aging degree corresponding to the battery to be sorted, includes the following steps:

[0081] Step 201: determine at least two third evaluation parameters based on the first evaluation parameter and the second evaluation parameter.

[0082] Optionally, the third evaluation parameter may be obtained by mathematical calculation based on the first evaluation parameter and the second evaluation parameter. Specifically, the data calculation method may be to calculate the product of one evaluation parameter and another evaluation parameter, or to divide one evaluation parameter by another evaluation parameter. The calculation methods of different evaluation parameters may differ and need to be determined according to actual conditions.

[0083] In one example, the first evaluation parameter includes the charge transfer resistance R CT , the second evaluation parameter includes the health state SOH, at this time, at least two third evaluation parameters are generated based on the first evaluation parameter and the second evaluation parameter, including: based on the charge transfer resistance R CT The product of the charge transfer resistance R and the health state SOH determines the third evaluation parameter. This can help the model capture the intrinsic relationship between the charge transfer process and the overall health of the battery, thereby improving the accuracy of the aging prediction results. In addition, since in the early stage of battery aging, the charge transfer resistance R CT The changes may not be obvious, but when combined with the health status SOH, these subtle changes can be discovered more accurately, which can help improve the accuracy of aging prediction results.

[0084] Furthermore, the second evaluation parameter also includes the internal temperature, based on the charge transfer resistance R CT The product of the internal temperature and the health state SOH determines the third evaluation parameter, including: determining the temperature coefficient corresponding to the internal temperature based on a preset temperature function; and calculating the charge transfer resistance R based on the temperature coefficient. ct The third evaluation parameter is obtained by correcting the product of the health state SOH.

[0085] The temperature function records the corresponding relationship between temperature and temperature coefficient. In one example, the temperature function records temperature intervals corresponding to different temperature coefficients. In this case, the temperature coefficient corresponding to the temperature interval to which the internal temperature recorded in the temperature function belongs can be determined as the temperature coefficient corresponding to the internal temperature.

[0086] Since the temperature of the battery cell will affect the resistance, this may affect the accuracy of the first evaluation parameter, and thus affect the accuracy of the judgment result. Therefore, in the above technical solution, the charge transfer resistance R is calculated based on the temperature coefficient corresponding to the internal temperature. CT The third evaluation parameter is obtained by correcting the product of the health state SOH, which can help reduce the influence of temperature on resistance detection, and further help improve the accuracy of the third evaluation parameter.

[0087] In another example, the first evaluation parameter includes an ohmic resistance R S and the charge transfer resistance R CT At this time, at least two third evaluation parameters are generated based on the first evaluation parameter and the second evaluation parameter, including: S and the charge transfer resistor R CT The ratio of is determined as the third evaluation parameter. In this way, the battery degradation and electrolyte performance decline can be effectively represented by the ratio as a quantitative indicator, reducing the model's dependence on specific resistance values, making the model applicable to aging prediction of batteries of different types and capacities, which can help broaden the scope of use of the model.

[0088] Furthermore, the second evaluation parameter includes the state of charge SOC. At this time, the ohmic resistance R S and the charge transfer resistor R CT The ratio of the charge state SOC to the ohmic resistance R is determined as the third evaluation parameter, including: S and the charge transfer resistor R CT The ratio of is adjusted to obtain the third evaluation parameter. For example: the ohmic resistance R S and the charge transfer resistor R CTThe third evaluation parameter is obtained by multiplying the ratio of the charge state SOC. Since the aging characteristics of the battery may be affected by the charge and discharge state SOC, the ohmic resistance RS and the charge transfer resistance R are combined with the charge and discharge state SOC. CT The ratio of is adjusted to obtain the third evaluation parameter, which can facilitate the model to learn the aging characteristics under different charge and discharge states SOC, and thus help to improve the accuracy of aging prediction.

[0089] In other examples, some parameters of the first evaluation parameter and the second evaluation parameter may be directly determined as the third evaluation parameter, for example, the internal short circuit in the second evaluation parameter may be directly determined as the third evaluation parameter.

[0090] Step 202: input the third evaluation parameter into the pre-trained aging model to obtain the aging degree corresponding to the battery to be sorted.

[0091] Among them, in the training data corresponding to the aging model, each group of training samples includes a third evaluation parameter, and the third evaluation parameter is determined based on the first evaluation parameter and the evaluation parameter of the sample battery.

[0092] In the above implementation, the third evaluation parameter can be comprehensively determined based on the first evaluation parameter and the second evaluation parameter, and the aging degree can be predicted by the third evaluation parameter, which can enhance the model's understanding and prediction capabilities of battery aging. In particular, battery aging is a complex and nonlinear relationship affected by multi-dimensional factors. Therefore, by fusing the features to a certain extent, the model can better learn the relationship between different evaluation parameters, thereby improving the accuracy of the prediction results.

[0093] Furthermore, since the third evaluation parameter is calculated based on the first evaluation parameter and the second evaluation parameter, in actual implementation, in order to enrich the dimension of the data, one first evaluation parameter may participate in the calculation of more than two third evaluation parameters, that is, the first evaluation parameter is combined with other different evaluation parameters to obtain different third evaluation parameters, and the same is true for the second evaluation parameter. This may lead to feature overlap between different third evaluation parameters, and then lead to differences in the degree of learning of different features by the model, which ultimately affects the accuracy of the aging judgment results.

[0094] Based on the above problems, further, the loss function used in the training process of the aging prediction model includes a prediction error term and a feature error term.

[0095] Among them, the prediction error term is used to measure the difference between the aging degree predicted by the model and the actual aging degree. The feature interaction term is used to measure the difference between the aging degrees predicted by different third evaluation parameters.

[0096] In the above technical solution, since an interactive error term is added to the loss function to measure the difference between the aging degrees predicted by different third evaluation parameters, the interactive error term can be used to balance the learning effect of the model on different third evaluation parameters, which can help improve the accuracy of the model prediction results.

[0097] This can help balance the model's learning effects on different third evaluation parameters.

[0098] In one example, the loss function is expressed as follows:

[0099]

[0100] Among them, L is the loss value; N is the total number of samples in the data set; α is the weight coefficient corresponding to the prediction error term, and α is the preset value; is the predicted value of the i-th group of samples; Y i is the true value of the i-th group of samples; p is a hyperparameter used to adjust the loss degree of the prediction error term; β is the weight coefficient corresponding to the feature interaction term; j and k are feature indexes used to traverse different combinations of third evaluation parameters, j is greater than or equal to 1, and k is less than or equal to the number of third evaluation parameters; φ jk is the interaction weight between the jth third evaluation parameter and the kth third evaluation parameter; is the predicted aging degree corresponding to the jth third evaluation parameter in the i-th group of samples; is the predicted aging degree corresponding to the kth third evaluation parameter in the i-th group of samples; q is a hyperparameter used to adjust the loss degree of the feature interaction term.

[0101] In the above loss function, since the interaction weight can control the influence of the difference between the predicted aging degrees corresponding to different third evaluation parameters on the loss value, the model's learning degree of different features can be balanced by reasonably setting the interaction weight, so that the model can better learn the influence of different features on the aging degree, which can help improve the accuracy of the aging judgment results.

[0102] Optionally, different third evaluation parameter interaction weight φ jk Specifically, for any two third evaluation parameters, the interaction weight φ of the two evaluation parameters can be determined based on the overlap between the first evaluation parameter and the second evaluation parameter used to determine the two third evaluation parameters. jk .

[0103] Specifically, the overlap degree and the interaction weight φ jkPositive correlation, the higher the overlap, the more common features there are between the two third evaluation parameters. At this time, the difference between the predicted aging degrees corresponding to the two third evaluation parameters is more important in the training of the model, so it is necessary to increase the interaction weight φ between the two third evaluation parameters. jk , so that the model can pay more attention to the learning of features that have differences in the two third evaluation parameters.

[0104] Furthermore, the interaction weight φ between the third evaluation parameters jk can be updated during the training process. In one example, the interaction weights φ are updated using gradient descent or other optimization algorithms. jk Adjustments are made to minimize the loss value. For example, during training, the interaction weight φ is updated according to the gradient of the loss function. jk .

[0105] In one example, the interaction weight φ is adjusted using gradient descent jk Update. The specific implementation is as follows:

[0106]

[0107] in, is the updated interaction weight; is the interaction weight before updating; is the learning rate, which controls the size of the update step; is the loss function Relative to the interaction weight gradient.

[0108] During each parameter update, not only the model parameters of the model need to be updated, but also the interaction weights need to be updated. In this way, the interaction weights between the third evaluation parameters can be dynamically adjusted during the model training process, thereby dynamically adjusting the impact of the differences in predicted aging degrees corresponding to the third evaluation parameters on the model training results, improving the model training effect, and thus helping to improve the accuracy of aging prediction.

[0109] The present application also discloses an automatic battery sorting device, referring to Figure 4 The battery sorting equipment includes at least one test station, which is used to accommodate the battery to be tested. Each test station includes a control component 310 and a first measurement component 320 and a second measurement component 330 connected to the control component 310 by signal.

[0110] The first measuring component 320 is used to perform an electrochemical impedance spectroscopy test on the battery to be tested at the testing station to obtain an electrochemical impedance spectrum.

[0111] The second measuring component 330 is used to collect the second evaluation parameter corresponding to the battery to be tested at the testing station.

[0112] The control component 310 is used to execute the battery sorting method provided by the above method embodiment.

[0113] Furthermore, the test station also includes a conveying mechanism, and the control component is also connected to the conveying mechanism signal; the control component is also used to, when the sorting result corresponding to the battery to be tested is determined, determine that the test of the battery to be tested is completed, and control the rotation of the conveying mechanism to move the tested battery out of the test station and move the next battery to be tested into the test station. Specifically, the conveying mechanism can be implemented as a conveyor belt. In this way, the automatic classification of the batteries whose sorting results have been determined can be achieved, which can help improve the degree of automation of the battery sorting process.

[0114] The present application also provides an electronic device. Figure 5 As shown, Figure 5 The electronic device 400 shown includes: a processor 401 and a memory 403. The processor 401 and the memory 403 are connected, such as through a bus 402. Optionally, the electronic device 400 may also include a transceiver 404. It should be noted that in actual applications, the transceiver 404 is not limited to one, and the structure of the electronic device 400 does not constitute a limitation on the embodiments of the present application.

[0115] Processor 401 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 401 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0116] The bus 402 may include a path to transmit information between the above components. The bus 402 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 402 may be divided into an address bus, a data bus, etc. For ease of representation, Figure 5Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0117] The memory 403 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0118] The memory 403 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 401. The processor 401 is used to execute the application code stored in the memory 403 to implement the contents shown in the above method embodiment.

[0119] The electronic devices include, but are not limited to, mobile terminals such as mobile phones, laptop computers, PDAs (personal digital assistants), PADs (tablet computers), etc., and fixed terminals such as digital TVs, desktop computers, etc. It can also be a server terminal, etc. Figure 5 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0120] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the instructions of the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise clearly stated in this document, the execution of these steps is not strictly limited in order and can be performed in other orders.

[0121] The above are only some implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A battery sorting method, characterized in that: The method comprises: Performing an electrochemical impedance spectroscopy test on the battery to be sorted to obtain an electrochemical impedance spectroscopy corresponding to the battery to be sorted; Performing equivalent circuit fitting on the electrochemical impedance spectrum to obtain a first evaluation parameter; Acquire a second evaluation parameter corresponding to the battery to be sorted, where the second evaluation parameter is directly obtained based on the battery to be sorted, and the second evaluation parameter is used to indicate the physical and chemical properties of the battery to be sorted; Inputting the first evaluation parameter and the second evaluation parameter into a pre-trained aging prediction model to obtain an aging degree corresponding to the battery to be sorted; Determine a sorting result corresponding to the battery to be sorted based on the aging degree; The step of inputting the first evaluation parameter and the second evaluation parameter into a pre-trained aging prediction model to obtain the aging degree corresponding to the battery to be sorted includes: determining at least two third evaluation parameters based on the first evaluation parameter and the second evaluation parameter; Inputting the third evaluation parameter into the pre-trained aging model to obtain the aging degree corresponding to the battery to be sorted; The loss function used in the aging prediction model training process includes a prediction error term and a feature interaction term, wherein the prediction error term is used to measure the difference between the aging degree predicted by the model and the actual aging degree, and the feature interaction term is used to measure the difference between the aging degrees predicted by different third evaluation parameters; The loss function is expressed by the following formula: , Wherein, L is the loss value; N is the total number of samples in the data set; α is the weight coefficient corresponding to the prediction error term; is the predicted value of the i-th group of samples; Y i is the true value of the i-th group of samples; p is a hyperparameter used to adjust the loss degree of the prediction error term; is the weight coefficient corresponding to the feature interaction term; j and k are feature indexes used to traverse different combinations of the third evaluation parameters, j is greater than or equal to 1, and k is less than or equal to the number of the third evaluation parameters; φ jk is the interaction weight between the j-th third evaluation parameter and the k-th third evaluation parameter; is the predicted aging degree corresponding to the jth third evaluation parameter in the i-th group of samples; is the predicted aging degree corresponding to the kth third evaluation parameter in the i-th group of samples; q is a hyperparameter used to adjust the loss degree of the feature interaction term.

2. The method according to claim 1, characterized in that The first evaluation parameter includes the charge transfer resistance R CT , the second evaluation parameter includes the health state SOH, and the generating at least two third evaluation parameters based on the first evaluation parameter and the second evaluation parameter includes: Based on the charge transfer resistance R CT The product of the health state SOH determines the third evaluation parameter.

3. The method according to claim 2, characterized in that The second evaluation parameter also includes the internal temperature, the charge transfer resistance R CT The third evaluation parameter is determined by multiplying the product of the health state SOH, including: Determining a temperature coefficient corresponding to the internal temperature based on a preset temperature function; Based on the temperature coefficient, the charge transfer resistor R CT The third evaluation parameter is obtained by correcting the product of the health state SOH.

4. The method according to claim 1, characterized in that: The interaction weights φ between different third evaluation parameters jk Initialized before training and updated during training.

5. The method according to claim 4, characterized in that Interaction weight φ jk Update via: ， in, is the updated interaction weight; is the interaction weight before updating; is the learning rate, which controls the size of the update step; is the loss function Relative to the interaction weight gradient.

6. The method according to claim 1, characterized in that The first evaluation parameter includes the ohmic resistance R S , SEI film resistance R SEI And the charge transfer resistance R CT At least one of; The second evaluation parameter includes at least one of the state of charge SOC, the state of health SOH, the internal temperature and the internal short circuit of the battery cell.

7. An automatic battery sorting device, characterized in that: The device comprises at least one test station, the test station is used to accommodate a battery to be tested, and each of the test stations comprises a control component and a first measurement component and a second measurement component connected to the control component by signal; The first measuring component is used to perform an electrochemical impedance spectroscopy test on the battery to be tested at the test station to obtain an electrochemical impedance spectrum, the second measuring component is used to collect a second evaluation parameter corresponding to the battery to be tested at the test station, and the control component is used to execute the battery sorting method described in any one of claims 1 to 6.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; Memory; At least one application, wherein the at least one application is stored in a memory and configured to be executed by at least one processor, and the at least one application is configured to: execute the battery sorting method according to any one of claims 1 to 6.

Images