Battery reorganization method, device, computer equipment and readable storage medium

Abstract

The embodiment of the present application provides a battery reorganization method, device, computer equipment and readable storage medium, which belongs to the field of battery-related technology. The method includes: obtaining a state index curve diagram of each target battery during the charging process; determining the target time period with the maximum sample significant value according to the distribution of sample significant values ​​in the sample significant map of the sample battery, and extracting the target state index curve corresponding to the target time period from the state index curve diagram; based on the target state index curve, extracting multiple target features corresponding to the target battery and inputting them into the first target model to obtain the battery capacity of each target battery; based on the multiple target features and battery capacity corresponding to the target battery, calculating the characteristic distance between the target battery and other candidate batteries; based on the characteristic distance, selecting a target candidate battery from multiple candidate batteries, and reorganizing the target battery and the target candidate battery. In this way, the efficiency and accuracy of battery reorganization can be improved.

Description

Technical Field

[0001] The present application relates to the technical field related to power batteries, and in particular to a battery reorganization method, device, computer equipment and readable storage medium. Background Art

[0002] In recent years, with the large-scale use of batteries in electronic devices, electric vehicles, and various energy storage scenarios, a large number of batteries will experience capacity decay and internal aging problems every year. At this time, these batteries will be eliminated because they are no longer suitable for the original application scenarios. For example, when the capacity of a battery is less than 80% of the rated capacity, it is considered no longer suitable for use in electric vehicles. However, in application scenarios with lower requirements (such as power grids, energy storage devices, etc.), screening and reorganizing the batteries for reuse without replacing the core materials of the batteries can reduce production costs and improve resource utilization.

[0003] In the related art, when grouping batteries, it often takes a lot of time to detect the capacity of the batteries, and the efficiency of detecting the capacity is low. In addition, the current method of grouping batteries is mainly based on the remaining capacity of the batteries. This grouping method does not take into account the internal state of the batteries, resulting in the reorganized batteries still having shortcomings such as short service life and poor consistency. There may even be safety hazards such as thermal runaway caused by excessive heat due to inaccurate battery reorganization. In summary, the efficiency and accuracy of battery reorganization in the related art are low. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to provide a battery reorganization method, device, computer equipment and readable storage medium, which can improve the efficiency and accuracy of battery reorganization.

[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application provides a battery reassembly method, the method comprising:

[0006] Obtain multiple state indicator curve graphs of each target battery during the charging process;

[0007] According to the distribution of sample significant values ​​of each sample time period in the sample significant graph of the sample battery, determine the target time period with the maximum sample significant value, and extract the target state index curve corresponding to the target time period from the state index curve graph; wherein the sample significant value is determined by the contribution of the sample state index when predicting the battery capacity of the sample battery in the corresponding sample time period;

[0008] Based on the target state index curve, extracting a plurality of target features corresponding to the target battery;

[0009] Inputting a plurality of the target features into a first target model to obtain a battery capacity of each of the target batteries;

[0010] Based on multiple target features and battery capacity corresponding to the target battery, calculating a feature distance between the target battery and other candidate batteries;

[0011] A target candidate battery is selected from the plurality of candidate batteries according to the characteristic distance, and the target battery and the target candidate batteries are recombined.

[0012] Accordingly, a second aspect of an embodiment of the present application provides a battery reorganization device, the device comprising:

[0013] An acquisition module, used to acquire multiple state indicator curve graphs of each target battery during the charging process;

[0014] A determination module, configured to determine a target time period having a maximum sample significant value according to the distribution of sample significant values ​​of each sample time period in the sample significant graph of the sample battery, and extract a target state index curve corresponding to the target time period from the state index curve graph; wherein the sample significant value is determined by the contribution of the sample state index in predicting the battery capacity of the sample battery in the corresponding sample time period;

[0015] An extraction module, used for extracting a plurality of target features corresponding to the target battery based on the target state index curve;

[0016] An input module, used for inputting the plurality of target features into the first target model to obtain the battery capacity of each of the target batteries;

[0017] A calculation module, configured to calculate a characteristic distance between the target battery and other candidate batteries based on a plurality of target characteristics and a battery capacity corresponding to the target battery;

[0018] A reorganization module is used to select a target candidate battery from the plurality of candidate batteries according to the characteristic distance, and to reorganize the target battery with the target candidate batteries.

[0019] In some embodiments, the computing module is further configured to:

[0020] Acquire a plurality of candidate features and a candidate battery capacity of the candidate battery;

[0021] For each of the target features, calculating a first Euclidean distance between the target feature and the corresponding candidate feature of the candidate battery;

[0022] Calculating a second Euclidean distance between the battery capacity and the corresponding candidate battery capacity;

[0023] Based on the first Euclidean distances and the second Euclidean distances, characteristic distances between the target battery and other candidate batteries are calculated.

[0024] In some embodiments, the computing module is further configured to:

[0025] Acquire a first weight set for each of the target features and a second weight set for the battery capacity;

[0026] For each of the first Euclidean distances, adjusting the first Euclidean distance according to the first weight of the corresponding target feature to obtain a first feature distance;

[0027] Adjusting the second Euclidean distance according to the second weight to obtain a second characteristic distance;

[0028] The characteristic distances between the target battery and other candidate batteries are calculated according to the sum of the first characteristic distances and the second characteristic distances.

[0029] In some implementations, the determining module is further configured to:

[0030] Acquire a plurality of sample batteries, and for each of the sample batteries, determine the contribution of each sample state indicator when the second target model predicts the battery capacity of the sample battery in a corresponding sample time period based on the sample state indicator after the sample state indicator is input into the second target model;

[0031] For each of the sample time periods, calculating a plurality of contribution degrees of a plurality of the sample state indicators of a plurality of the sample batteries in the sample time period;

[0032] The average value of the multiple contribution degrees in each sample time period is calculated to obtain the sample significance value of each sample time period.

[0033] In some embodiments, the battery recombination device further includes a first training module for:

[0034] Acquire a plurality of first sample state indicator curve graphs corresponding to a plurality of sample batteries, and add noise data to each of the first sample state indicator curve graphs in any sample time period to obtain a first sample state indicator curve graph processed with the noise data;

[0035] Inputting the sample state indicators in each of the first sample state indicator curve graphs into the second preset model in chronological order, and obtaining the first sample predicted capacity of the sample battery predicted by the second preset model according to the first sample state indicator curve graph;

[0036] Acquire a first real sample capacity of the sample battery, and calculate a first loss of the second preset model based on the first sample predicted capacity and the first real sample capacity;

[0037] Based on the first loss, the model parameters of the second preset model are adjusted until the number of training times for the second preset model reaches the first preset number of times, so as to obtain a second target model.

[0038] In some embodiments, the battery recombination device further includes a second training module for:

[0039] Obtaining a sample battery, and testing the sample battery to obtain a second real sample capacity of the sample battery;

[0040] Charging the sample battery to a preset charging time to obtain a second sample state indicator curve graph of the sample battery;

[0041] Determine the target time period with the maximum sample significance value according to the distribution of sample significance values ​​of each time period in the preset sample significance map;

[0042] Extracting a target sample state index curve corresponding to the target time period from the second sample state index curve graph, and extracting a plurality of sample features corresponding to the sample battery based on the target sample state index curve;

[0043] Inputting a plurality of the sample features into a first preset model to obtain a second sample prediction capacity;

[0044] Calculating a second loss of the first preset model based on the second sample prediction capacity and the second real sample capacity;

[0045] Based on the second loss, the model parameters of the first preset model are adjusted until the number of training times for the first preset model reaches a second preset number of times, so as to obtain a first target model.

[0046] In some implementations, the state indicator curve graph includes a temperature change curve graph and a voltage change curve graph, and the extraction module is further used to:

[0047] Based on the target temperature change curve determined from the temperature change curve graph, determine a minimum temperature value, a first moment corresponding to the minimum temperature value, a first slope between a starting point of the target temperature change curve and a minimum temperature point corresponding to the minimum temperature value, a second slope between the minimum temperature point and an end point of the target temperature change curve, a first difference between a starting temperature value corresponding to the starting point of the target temperature change curve and the minimum temperature value, and a second difference between an end temperature value corresponding to the end point of the target temperature change curve and the minimum temperature value;

[0048] Based on the target voltage change curve determined from the voltage change curve diagram, determine the voltage starting value corresponding to the starting point of the target voltage change curve, the voltage ending value corresponding to the ending point of the target voltage change curve, the third difference between the voltage ending value and the voltage starting value, the third slope corresponding to the starting point of the target voltage change curve and the ending point of the target voltage change curve, and the average voltage value of the target voltage change curve corresponding to the target time period.

[0049] The third aspect of the present application provides a computer device, which includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the battery recombination method described in any one of the first aspect of the present application is implemented.

[0050] The fourth aspect of the present application provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the battery recombination method described in any one of the first aspect of the present application is implemented.

[0051] The present application obtains multiple state index curves of each target battery during the charging process; according to the distribution of sample significant values ​​of each sample time period in the sample significant graph of the sample battery, the target time period with the maximum sample significant value is determined, and the target state index curve corresponding to the target time period is extracted from the state index curve; wherein the sample significant value is determined by the contribution of the sample state index when predicting the battery capacity of the sample battery in the corresponding sample time period; based on the target state index curve, multiple target features corresponding to the target battery are extracted; multiple target features are input into the first target model to obtain the battery capacity of each target battery; based on the multiple target features and battery capacity corresponding to the target battery, the characteristic distance between the target battery and other candidate batteries is calculated; according to the characteristic distance, the target candidate battery is selected from the multiple candidate batteries, and the target battery and the target candidate battery are reorganized. In this way, the target feature that is conducive to accurately predicting the battery capacity of the target battery can be directly selected and input into the first target model, so as to realize the rapid and accurate prediction of the battery capacity. Moreover, in the process of selecting the battery for reorganization for the target battery, the reorganization is not only based on the battery capacity, but also based on the characteristic distance between other multiple target features, so as to fully consider the internal state between the reorganized batteries and improve the accuracy of battery reorganization. In summary, the present application can improve the efficiency and accuracy of battery reorganization. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of the architecture of the battery recombination system provided in an embodiment of the present application;

[0053] Figure 2is a flow chart of a battery reassembly method provided in an embodiment of the present application;

[0054] Figure 3 is an example diagram of a sample saliency map provided in an embodiment of the present application;

[0055] Figure 4 This is an actual test effect diagram provided by the embodiment of the present application;

[0056] Figure 5 This is another actual test effect diagram provided by the embodiment of the present application;

[0057] Figure 6 is a schematic diagram of the functional modules of a battery recombination device provided in an embodiment of the present application;

[0058] Figure 7 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. 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.

[0060] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0062] In recent years, with the large-scale use of batteries in electronic devices, electric vehicles, and various energy storage scenarios, a large number of batteries will experience capacity decay and internal aging problems every year. At this time, these batteries will be eliminated because they are no longer suitable for the original application scenarios. For example, when the capacity of the battery is less than 80% of the rated capacity, it is considered no longer suitable for use in electric vehicles. However, in application scenarios with lower requirements (such as power grids, energy storage devices, etc.), compared to direct disassembly and recycling, screening and reorganizing the batteries for reuse without replacing the core materials of the batteries can reduce production costs and improve resource utilization.

[0063] In the related art, when grouping batteries, it often takes a lot of time to detect the capacity of the batteries, and the efficiency of detecting the capacity is low. In addition, the current method of grouping batteries is mainly based on the remaining capacity of the batteries. This grouping method does not take into account the internal state of the batteries, resulting in the reorganized batteries still having shortcomings such as short service life and poor consistency. There may even be safety hazards such as thermal runaway caused by excessive heat due to inaccurate battery reorganization. In summary, the efficiency and accuracy of battery reorganization in the related art are low.

[0064] Based on this, the embodiments of the present application provide a battery reorganization method, apparatus, computer device and readable storage medium, which can improve the efficiency and accuracy of battery reorganization.

[0065] The battery recombination method, apparatus, computer device and readable storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the battery recombination system in the embodiments of the present application is described.

[0066] Please refer to Figure 1 In some embodiments, the present application provides a battery recombination system, including an electrochemical tester 11 , a sensor 12 , a demodulator 13 , and a capacity prediction and sorting test component 14 .

[0067] Specifically, the electrochemical tester 11 can be used to test and measure the electrochemical performance parameters of the battery, such as voltage, current, and impedance. The sensor 12 can be a temperature sensor and a voltage sensor. The temperature sensor can be used to measure the change of the battery temperature, and the voltage sensor can be used to measure the change of the voltage. The sensor 12 can provide real-time status data and transmit the real-time status data to the demodulator 13 to convert the voltage change signal into a voltage change curve, convert the temperature change signal into a temperature change curve, and transmit the voltage change curve and the temperature change curve to the capacity prediction and sorting test component.

[0068] Furthermore, the capacity prediction and sorting test component 14 can be a server-side component, which can analyze the voltage change curve and temperature change curve obtained from the demodulator 13 by applying a machine learning model, and extract the corresponding multiple target features to accurately predict the battery capacity. In addition, the capacity prediction and sorting test component can also calculate the feature distance based on multiple target features, and select batteries with similar capacity and performance characteristics for reorganization, thereby achieving fast and accurate reorganization of the battery.

[0069] The battery recombination method in the embodiments of the present application can be illustrated by the following embodiments.

[0070] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for enabling the normal operation of the embodiment of the present application will be obtained.

[0071] In the embodiments of the present application, the battery reorganization device will be described from the perspective of the battery reorganization device, which can be integrated into a computer device. Figure 2 , Figure 2 This is a flowchart of the steps of the battery reassembly method provided in the embodiment of the present application. The embodiment of the present application takes the battery reassembly device as an example, which is specifically integrated in a terminal or a server. When the processor on the terminal or the server executes the program instructions corresponding to the battery reassembly method, the specific process is as follows:

[0072] Step 101, obtaining multiple state indicator curve graphs of each target battery during the charging process.

[0073] It is understandable that in order to analyze the target battery, multiple status indicator curves of the target battery during the charging process can be obtained to record the changes in the status indicators of the target battery during the charging process, so as to facilitate the subsequent acquisition of the target characteristics of the target battery and realize accurate identification of the battery capacity.

[0074] Among them, the target battery can be a battery unit that needs to be capacity tested and evaluated, the target battery can be a battery that needs to be sorted during the production process, or it can be a battery that is already in use and needs status monitoring and maintenance.

[0075] Among them, the state index curve graph may include a temperature change curve graph and a voltage change curve graph. For example, the state index curve graph may be obtained by placing a temperature sensor or an infrared temperature measuring device on the target battery, charging the battery with a constant current, and recording the voltage through an electrochemical tester, and transmitting the voltage data in real time to a demodulator for demodulation to obtain a voltage change curve graph. Furthermore, the temperature measured by the temperature sensor may be transmitted to a demodulator for demodulation in real time to obtain a temperature change curve graph.

[0076] Furthermore, in order to speed up the detection efficiency of the target battery, a large current can be used to charge the target battery within the current range that the target battery can withstand, so as to stimulate a larger overpotential and heat generation of the target battery. For example, the current size can be determined according to the actual working conditions of the target battery, such as using a current of 1 / 2 / 3 / 4 / 5C or the like. When the state indicator curve diagram includes a temperature change curve diagram and a voltage change curve diagram, taking 1C as an example, the voltage change curve diagram and the temperature change curve diagram of the target battery within 30 minutes can be collected under the 1C constant current charging state.

[0077] It is understandable that, since the battery has a certain spatial volume, when testing the battery temperature, temperature sensors can be placed at multiple points of the battery, and the temperature range of multiple points can be recorded simultaneously during the test to screen the batteries that need to be reorganized to ensure the safety of battery reorganization. Exemplarily, 9 temperature sensors can be dispersed and controlled at different locations during the battery testing process to obtain 9 temperatures, such as T1, T2, T3, T4, T5, T6, T7, T8, T9, and calculate the standard deviation S of the 9 temperature changes, as well as the difference ΔT corresponding to the highest temperature point and the lowest temperature point. When the standard deviation S is greater than the safety standard deviation threshold S0 and ΔT also satisfies the safety temperature difference threshold T0, it indicates that the temperature change of the battery during the charging process exceeds the preset safety range, and there may be problems inside the battery. Therefore, in order to avoid the safety risks caused by uneven temperature, the battery is not suitable for reorganization prediction.

[0078] Furthermore, during the charging process, a battery whose standard deviation is smaller than a safety standard deviation threshold or whose difference between the highest temperature point and the lowest temperature point is smaller than a safety temperature difference threshold may be determined as a target battery.

[0079] For example, a temperature sensor may be installed on the target battery to monitor the temperature change of the battery during the charging process, and a voltage sensor may be connected to monitor the voltage change of the target battery during the charging process. Afterwards, the battery is charged with a predetermined charging current (such as 1C current), and a data logger is used to synchronously record the voltage and temperature change data of the battery at a set sampling frequency.

[0080] Afterwards, the collected voltage change data and temperature change data can be converted into voltage change curve graphs and temperature change curve graphs through capacity prediction and sorting test components.

[0081] Through the above method, the preliminary screening of batteries can be achieved, and the state index curve graph of the target battery can be accurately obtained, so as to facilitate the subsequent extraction of the target state index curve from the state index curve graph and complete the accurate prediction of the capacity of the target battery.

[0082] Step 102, based on the distribution of sample significance values ​​in each sample time period in the sample significance graph of the sample battery, determine the target time period with the maximum sample significance value, and extract the target state indicator curve corresponding to the target time period from the state indicator curve graph; wherein the sample significance value is determined by the contribution of the sample state indicator in predicting the battery capacity of the sample battery in the corresponding sample time period.

[0083] It should be noted that since the performance of similar batteries under similar conditions is similar, the time period corresponding to the important characteristics of the sample battery is also the time period corresponding to the important characteristics of the similar batteries. Therefore, the target time period with the maximum sample significance value can be determined by testing the sample battery, and the corresponding target state indicator curve can be extracted from the state indicator curve graph of the target battery based on the target time period to achieve efficient and accurate analysis of the target battery.

[0084] The sample battery can be a specific battery used to collect data and train models in a battery health monitoring system. The data of the sample battery can be used to analyze and predict the behavior and status of similar batteries.

[0085] The sample significant graph may be a graph generated by analyzing the sample status indicators of the sample battery during the charging process, and is used to display significant values ​​of the sample status indicators in different time periods.

[0086] The sample time period may be a time period during which the sample battery is charged and used to collect the status indicator data. The sample time period may be determined arbitrarily or may be set according to a preset time period.

[0087] The sample significance value may refer to the contribution of the sample state indicator to the prediction of the battery capacity within the sample time period. Each sample time period may correspond to at least one sample significance value, and the higher the sample significance value, the higher the contribution of the state indicator within the sample time period to the prediction of the battery capacity.

[0088] The maximum sample saliency value may be a sample saliency value with the largest value among all sample saliency values ​​in the sample saliency map.

[0089] The target time period may be a time period in which the maximum sample significance value is determined based on the sample significance map of sample batteries of the same type as the target battery.

[0090] The target state indicator curve may be a state indicator curve corresponding to a target time period extracted from the state indicator curve diagram, and the state indicator curve is used to reflect changes in state indicators of the target battery within the target time period.

[0091] The sample status indicator may be various physical or chemical parameters used to describe the battery status during the sample battery detection process, such as voltage and temperature.

[0092] The battery capacity may be the amount of electrical energy that the battery can store and release.

[0093] The contribution degree can be the importance of the sample state indicator in predicting the battery capacity of the sample battery, that is, the correlation between the corresponding sample state indicator and the battery capacity prediction operation. The higher the contribution degree, the higher the importance of the sample state indicator in the corresponding time period for correctly predicting the battery capacity of the sample battery, the higher the correlation of the sample state indicator in the corresponding time period for predicting the battery capacity, and the higher the sample significance value.

[0094] For example, by combining Figure 3 The process of determining the target time period based on the sample saliency map is given as an example. Figure 3 In , b represents the sample significance value, ΔT represents the temperature, V represents the voltage, and Q represents the capacity of the sample battery. Specifically, the sample significance map can be divided into three sample time periods, namely 0 to 600 seconds, 600 to 1200 seconds, and 1200 seconds to 1800 seconds. Through the distribution of the significance values ​​of the three sample time periods, it can be determined that the sample time period with the maximum sample significance value is 1200 seconds to 1800 seconds. Therefore, 1200 seconds to 1800 seconds can be used as the target time period. Furthermore, the time distance of the sample time period can be reduced as required and subdivided into multiple sample time periods, such as 6 or 8 sample time periods, etc., and then the target time period can be selected from multiple sample time periods according to the sample significance value. For example, in Figure 3 In the example, grayscale is used to represent the size of the data. The darker the grayscale, the larger the sample significance value. Therefore, we can Figure 3 It is known that the sample significance values ​​corresponding to the sample time periods of 0-600 seconds and 1200-1800 seconds are significantly higher than the sample significance values ​​of other sample time periods. Therefore, the sample time period of 0-600 seconds can be selected as the target time period. In different sample significance maps, the method of selecting sample time periods is the same as above, which will not be repeated here.

[0095] Further, when the target time period is determined, for example, 1200 seconds to 1800 seconds is taken as the target time period, the target state curve corresponding to 1200 seconds to 1800 seconds can be extracted from the state indicator curve diagram to facilitate the subsequent extraction of target features of the target state curve.

[0096] In some embodiments, when each sample time period corresponds to multiple sample significance values, and the maximum sample significance value appears in multiple sample time periods, for example, when the maximum sample significance value of 0.6 appears in both the first sample time period and the second sample time period, all sample significance values ​​of the first sample time period and the second sample time period can be accumulated respectively to obtain a first accumulated significance value corresponding to the first sample time period and a second accumulated significance value corresponding to the second sample time period, and the first accumulated significance value and the second accumulated significance value are compared to select the sample time period with the larger accumulated significance value as the target time period.

[0097] It can be understood that by analyzing the distribution of sample significant values ​​of sample batteries, the target time period that is most important for predicting the battery capacity of the sample battery can be determined, so that the target time period can be directly applied to the prediction of the battery capacity of similar target batteries and the selection of target features, thereby achieving efficient and accurate analysis of the target battery.

[0098] In some embodiments, in order to quickly and accurately select the most important target time period in the charging process, the sample significant value of the sample battery can be calculated, so as to quickly determine the target time period according to the distribution of the sample significant value, and apply the target time period to the significant selection of the target state index curve of the same type of battery, so as to achieve rapid prediction and reorganization of the target battery. Exemplarily, the sample significant value can be obtained by the following steps:

[0099] (A.1) obtaining a plurality of sample batteries, and determining, for each sample battery, a contribution of each sample state indicator after being input into the second target model, when the second target model predicts the battery capacity of the sample battery in a corresponding sample time period based on the sample state indicator;

[0100] (A.2) for each sample time period, calculating multiple contribution degrees of multiple sample state indicators of multiple sample batteries in the sample time period;

[0101] (A.3) Calculate the average of multiple contributions in each sample time period to obtain the sample significance value of each sample time period.

[0102] Among them, the second target model can be a deep learning model, or a machine learning model such as a support vector machine or a random forest.

[0103] It can be understood that only when the second target model can make an accurate prediction of the battery capacity of the sample battery, it indicates that at least part of the sample state indicators of the corresponding sample battery can accurately and stably predict the battery capacity. Therefore, the sample state indicators of the corresponding sample battery should be retained; if all the sample state indicators in the sample battery are input into the second target model, the second target model cannot make an accurate prediction of the battery capacity, indicating that the sample state indicator of the sample battery used to calculate the sample significant value is inaccurate and has no reference significance. The sample state indicator of the corresponding sample battery should be discarded and not considered for calculating the sample significant value. Exemplarily, assuming that there are multiple sample batteries that have obtained real capacity as labels, by charging and discharging the sample batteries, the voltage change curve and temperature change curve corresponding to each sample battery can be detected. Afterwards, the sample state indicators in the voltage change curve and the temperature change curve are input into the second target model, and after obtaining the predicted capacity of the sample battery predicted by the second target model, the capacity difference between the predicted capacity and the actual capacity is calculated. When the capacity difference is less than the preset difference threshold, it indicates that under the sample state indicator corresponding to the sample battery, the second target model can successfully make an accurate prediction of the battery capacity. Therefore, the contribution of the sample battery in each sample time period can be involved in the calculation of the sample significant value of the final sample time period. When the capacity difference is less than the preset difference threshold, it indicates that under the sample state indicator corresponding to the sample battery, the second target model cannot make an accurate prediction. At this time, the corresponding sample battery can be screened out, and the contribution of the sample battery in each sample time period does not participate in the calculation of the sample significant value of the sample time period.

[0104] In some embodiments, after selecting data corresponding to sample batteries with smaller prediction errors from the prediction data of the second target model, the sample significance value can be calculated. Specifically, when training the second target model, the second target model can generate the contribution of the sample battery in different sample time periods based on the importance of the sample state indicator in each sample time period for predicting the battery capacity while predicting the battery capacity of the sample battery.

[0105] Further, for example, sample state indicator 1 corresponds to sample time period 1, sample state indicator 2 corresponds to sample time period 2, and sample state indicator 3 corresponds to sample time period 3. The contribution of sample state indicator 1 to the second target model's ability to accurately predict the battery capacity is 60%, so the contribution of the second target model outputting sample time period 1 is 0.6; the contribution of sample state indicator 2 to the second target model's ability to accurately predict the battery capacity is 30%, so the contribution of sample time period 2 is 0.3; the contribution of sample state indicator 3 to the second target model's ability to accurately predict the battery capacity is 10%, so the contribution of sample time period 3 is 0.1.

[0106] It is understandable that while the second target model predicts the battery capacity of the sample battery, a sample significance map can be generated based on the predicted sample capacity and contribution. It is also possible that after predicting the sample battery, the sample status indicators of the sample batteries with good prediction effects can be selected and re-input into the second target model to generate a significance map. The present application does not impose any specific restrictions on this.

[0107] It can be understood that the contribution degree can also be called a sub-sample significant value. Each sample battery can correspond to a sub-sample significant value in a time period. In order to distinguish it from the final sample significant value, the contribution degree is used to represent it. The sum of multiple contribution degrees in the sample time period is averaged, and the average value obtained can be used as the sample significant value of the sample time period, so as to help more accurately evaluate and quantify the contribution of the sample state index to the prediction of battery capacity, so that the sample significant value finally calculated is more representative and referenced among similar batteries. For example, for sample battery 1, its contribution degree in sample time period 1 is 0.1, its contribution degree in sample time period 2 is 0.3, and its contribution degree in sample time period 3 is 0.6; for sample battery 2, its contribution degree in sample time period 1 is 0.2, its contribution degree in sample time period 2 is 0.3, and its contribution degree in sample time period 3 is 0.5; for sample battery 3, its contribution degree in sample time period 1 is 0.1, its contribution degree in sample time period 2 is 0.2, and its contribution degree in sample time period 3 is 0.7. Then, the sample significance value corresponding to sample time period 1 is (0.1+0.2+0.1) / 3=0.13, the sample significance value corresponding to sample time period 2 is (0.3+0.3+0.2) / 3=0.27, and the sample significance value corresponding to sample time period 3 is (0.6+0.5+0.7) / 3=0.6.

[0108] Through the above method, the average value of multiple contributions in each sample time period can be calculated to obtain the sample significance value of each sample time period, so as to evaluate and quantify the contribution of the sample status index to the predicted battery capacity, which is conducive to selecting the target time period and applying it to the prediction of similar batteries, helping to improve the accuracy of predicted battery capacity and making the calculated sample significance value more representative among similar batteries.

[0109] It can be understood that in order to make the second target model predict the sample battery more accurately and calculate the contribution corresponding to different sample time periods as much as possible, the second target model can be trained. Exemplarily, the second target model can be trained by the following steps:

[0110] (B.1) obtaining a plurality of first sample state index curve graphs corresponding to a plurality of sample batteries, and adding noise data to each first sample state index curve graph in any sample time period to obtain a first sample state index curve graph after noise data processing;

[0111] (B.2) Inputting the sample state indicators in each first sample state indicator curve graph into the second preset model in chronological order, obtaining the first sample predicted capacity predicted by the second preset model for the sample battery according to the first sample state indicator curve graph

[0112] (B.3) obtaining a first real sample capacity of the sample battery, and calculating a first loss of the second preset model based on the first sample predicted capacity and the first real sample capacity;

[0113] (B.4) Based on the first loss, the model parameters of the second preset model are adjusted until the number of training times for the second preset model reaches the first preset number of times, thereby obtaining a second target model.

[0114] The first sample state index curve diagram may be a voltage change curve and a temperature change curve obtained by monitoring the charge and discharge process of the sample battery, and the voltage change curve and the temperature change curve reflect the performance index of the sample battery in different states.

[0115] The noise data may be the noise data added when training the second preset model in order to facilitate the calculation of the contribution of each sample state indicator. After the noise data is added to the sample state indicator corresponding to a sample time period in the first sample state indicator curve, the second preset model can still make an accurate prediction of the battery capacity of the sample battery based on other sample state indicators, indicating that the sample state indicator with the added noise data is not an important indicator and has a low contribution to the successful prediction of the sample battery. If, after adding the noise data, the first sample predicted capacity predicted by the second preset model is very different from the first true sample capacity, it indicates that the sample state indicator with the added noise data is an important indicator and has a high corresponding contribution.

[0116] Among them, the second preset model can be an initial model for predicting battery capacity and generating sample significance values ​​for the corresponding sample time period. The second preset model can be a deep learning model, a support vector machine, a random forest or other machine learning model. After the second preset model is trained, the second target model can be obtained.

[0117] The first sample predicted capacity may be a battery capacity predicted by the second preset model according to an input sample state indicator.

[0118] The first real sample capacity may be a battery capacity actually measured on a sample battery, which is used for comparison with the first sample predicted capacity to evaluate the prediction accuracy of the second preset model.

[0119] The first loss may be the difference between the first sample predicted capacity predicted by the second preset model for the sample battery and the first real sample capacity. The first loss may be measured by a mean square error or other loss functions.

[0120] The first preset number of times may be the number of times the model parameters are repeatedly adjusted to optimize the model performance when training the second preset model, and the first preset number of times may be obtained by setting, for example, 500 times, 1000 times, etc. Alternatively, the first preset number of times may not be set, and the training of the second preset model may be stopped when the second preset model converges to obtain the second target model.

[0121] It can be understood that in the process of training the second preset model, the second preset model can obtain the voltage change, temperature change, capacity change, etc. of the sample battery during charging, and generate a corresponding sample significance map to represent the importance of the sample status indicator in each sample time period. The sample significance value is generally in the (0, 1) range.

[0122] Furthermore, the noise data sample x after being processed by the noise data can be generated according to the following formula: noise :

[0123]

[0124] Among them, x represents the first sample state indicator curve graph, M represents the significance graph, represents noise data, which is the data after processing the first sample state indicator curve graph.

[0125] It is understandable that when training the second preset model, it is necessary to make effective capacity predictions based on both the processed noise data samples and the original samples. At the same time, the total saliency value in the saliency map is required to be reduced as much as possible so that the second preset model is forced to focus on those highly significant areas that are retained. Therefore, after continuous training iterations, the sample saliency map only retains higher sample saliency values ​​for important information of the second preset model.

[0126] Exemplarily, by charging the same type of sample batteries for a uniform time, for example, charging for 30 minutes, a first sample state indicator curve graph showing the voltage and temperature changing with time during the charging process can be obtained. Furthermore, noise data can be added to the first sample state indicator curve corresponding to at least one sample time period in the first sample state indicator curve graph, and the first sample state indicators with added noise data can be sequentially input into the second preset model in chronological order to obtain the first sample predicted capacity of the sample battery predicted by the second preset model based on the first sample state indicators. It can be understood that the first sample state indicator can be temperature data and voltage data. In some embodiments, the first sample state indicator curve graph with added noise can also be directly input into the second preset model.

[0127] Furthermore, noise data may be added to all time periods in the first sample state indicator curve graph in sequence. For example, the first sample state indicator curve graph has a first time period, a second time period and a third time period, and the first time period, the second time period and the third time period are all sample time periods covered in the first sample state indicator curve graph. Afterwards, noise data can be added to the first time period in the first sample state indicator curve graph, and the first sample state indicator graph with the added noise data is input into the second preset model in chronological order to obtain the first sample predicted capacity of the sample battery predicted by the second preset model according to the first sample state indicator; then, noise data is added to the second time period in the first sample state indicator curve graph, and the first sample state indicator graph with the added noise data is input into the second preset model in chronological order to obtain the first sample predicted capacity of the sample battery predicted by the second preset model according to the first sample state indicator graph; further, noise data is added to the third time period in the first sample state indicator curve graph, and the first sample state indicator graph with the added noise data is input into the second preset model in chronological order to obtain the first sample predicted capacity of the sample battery predicted by the second preset model according to the first sample state indicator graph, so that the second preset model can obtain the corresponding multiple first sample predicted capacities according to the first sample state indicator curve graph after adding noise data in different sample time periods.

[0128] Exemplarily, all sample time periods corresponding to the first sample state index curve graph include the first time period, the second time period and the third time period. Specifically, in the first sample state index curve graph, the predicted capacity of the first sample corresponding to the addition of noise data to the first time period is 38Ah (ampere-hour), the predicted capacity of the first sample corresponding to the addition of noise data to the second time period in the first sample state index curve graph is 50Ah, and the predicted capacity of the first sample corresponding to the addition of noise data to the third time period in the first sample state index curve graph is 44Ah. The battery capacity of the correct sample battery is 50Ah. It can be seen that after adding noise data to the data of the first time period, the battery capacity predicted by the second preset model is more inaccurate. Therefore, the first sample state index corresponding to the first time period is more critical for accurately predicting the battery capacity, and the sample significance value of the first time period is higher; and after adding noise data to the data of the second time period, the battery capacity can still be accurately predicted based on the remaining normal data. It can be seen that the data corresponding to the second time period is not the key data that can accurately predict the battery capacity of the sample battery, and the sample significance value corresponding to the second time period is lower. It can be seen that the sample significance value of the first time period output by the second preset model> the sample significance value of the third time period> the sample significance value of the second time period.

[0129] Furthermore, after the first sample predicted capacity is obtained through the second preset model prediction, the first loss of the second preset model can be calculated based on the first real sample capacity corresponding to the sample battery, which can be calculated specifically through the mean square error function. Afterwards, the model parameters of the second preset model can be adjusted based on the first loss, for example, the parameters of the second preset model can be adjusted using gradient descent to reduce the first loss, until the number of training times for the second preset model reaches the preset 500 times, the training of the second preset model is stopped, and the second target model is obtained.

[0130] Through the above method, while continuously improving the accuracy of the battery capacity predicted by the second preset model, the second preset model is forced to retain areas with higher significance values, which is beneficial for the second preset model to capture truly useful features and also beneficial for the second preset model to obtain accurate sample significance values.

[0131] Step 103: extract multiple target features corresponding to the target battery based on the target state index curve.

[0132] In some embodiments, in order to predict the capacity of the target battery and facilitate the subsequent calculation of the characteristic distance between the target battery and other candidate batteries, the target battery of the target battery can be extracted through the target state indicator curve to improve the accuracy of the target battery capacity prediction and the accuracy and safety of the subsequent reorganization of the target battery.

[0133] The target state index curve may be a curve corresponding to a target time period selected from the state index curve diagram, and may also be a curve having the highest significance value calculated by sample batteries of the same type.

[0134] Among them, the target feature can be the state index of the target state index curve in different target time periods or the change value of the state index. For example, when the state index curve diagram is a temperature change curve diagram, the target feature can be the minimum temperature value, the first moment corresponding to the minimum temperature value, the first slope between the starting point of the target temperature change curve and the minimum temperature point corresponding to the minimum temperature value, the second slope between the minimum temperature point and the end point of the target temperature change curve, the first difference between the starting temperature value and the minimum temperature value corresponding to the starting point of the target temperature change curve, and the second difference between the ending temperature value and the minimum temperature value corresponding to the end point of the target temperature change curve. When the state index curve diagram is a voltage change curve diagram, the target feature can be the starting voltage value corresponding to the starting point of the target voltage change curve, the ending voltage value corresponding to the end point of the target voltage change curve, the third difference between the ending voltage value and the starting voltage value, the third slope corresponding to the starting point of the target voltage change curve and the end point of the target voltage change curve, and the average voltage value of the target voltage change curve corresponding to the target time period.

[0135] It is understandable that by analyzing the characteristic values ​​of key indicators such as temperature and voltage, we can deeply understand the working status and change trend of the target battery, thereby predicting the capacity change of the target battery and optimizing the reorganization plan, promoting the continuous development and improvement of battery management and application.

[0136] It should be noted that the state indicator curve graph may include a temperature change curve graph and a voltage change curve graph. By analyzing the characteristic values ​​of the temperature change curve graph, the performance of the target battery under different temperature conditions can be understood, and the capacity change of the battery can be inferred; and by analyzing the characteristic values ​​of the voltage change curve graph, the charging and discharging conditions of the battery can be revealed, and the battery changes can be inferred. Therefore, by analyzing the target features in the temperature change curve graph and the voltage change curve graph, an important basis can be provided for the prediction and reorganization of the battery capacity. Exemplarily, step 103 may include:

[0137] Based on the target temperature change curve determined from the temperature change curve graph, determine the minimum temperature value, the first moment corresponding to the minimum temperature value, the first slope between the starting point of the target temperature change curve and the minimum temperature point corresponding to the minimum temperature value, the second slope between the minimum temperature point and the end point of the target temperature change curve, the first difference between the starting temperature value corresponding to the starting point of the target temperature change curve and the minimum temperature value, and the second difference between the ending temperature value corresponding to the end point of the target temperature change curve and the minimum temperature value;

[0138] Based on the target voltage change curve determined from the voltage change curve diagram, determine the voltage starting value corresponding to the starting point of the target voltage change curve, the voltage ending value corresponding to the ending point of the target voltage change curve, the third difference between the voltage ending value and the voltage starting value, the third slope corresponding to the starting point of the target voltage change curve and the ending point of the target voltage change curve, and the average voltage value of the target voltage change curve corresponding to the target time period.

[0139] It should be noted that since the state indicator curve diagram may include a temperature change curve diagram and a voltage change curve diagram, the target temperature change curve corresponding to the target time period may be determined from the temperature change curve diagram, and the target voltage change curve corresponding to the target time period may be determined from the voltage change curve diagram.

[0140] The minimum temperature value may be the lowest temperature value reached in the target temperature change curve, and the minimum temperature value may reflect the heat absorption effect caused by entropy change of the target battery within the target time period.

[0141] Among them, the first moment corresponding to the minimum temperature value may be a specific time point corresponding to the minimum temperature value. The first moment may be associated with the aging degree of the target battery, and plays an important role in analyzing the current state of the target battery.

[0142] Among them, the first slope can be the slope from the starting point of the target temperature change curve to the temperature minimum point corresponding to the temperature minimum value. The first slope can reflect the temperature drop rate of the target battery at the beginning of discharge, thereby providing information about the initial state of the target battery.

[0143] The second slope may be the slope from the minimum temperature point to the end point of the target temperature change curve. The second slope may reflect the rate of change of the temperature of the target battery during the charging process and is used to evaluate the thermal behavior of the battery under charging conditions.

[0144] The first difference may be the difference between the temperature value corresponding to the starting point of the target temperature change curve and the minimum temperature value, and the first difference may be used to measure the distance between the temperature of the target battery at the beginning of discharge and the minimum temperature.

[0145] The second difference may be the difference between the temperature value corresponding to the end point of the target temperature change curve and the minimum temperature value, and the second difference may be used to characterize the temperature rise of the target battery at the end of discharge.

[0146] The voltage start value may be a voltage reading at the start point of the target voltage variation curve, and the voltage start value may be used as a reference for the voltage state when the target battery starts to discharge.

[0147] The voltage end value may be a voltage reading at the end point of the target voltage variation curve, and the voltage end value may be used to characterize the voltage state of the target battery at the end of discharge.

[0148] The third difference may be the difference between the voltage end value and the voltage start value, and the third difference may be used to characterize the voltage drop of the target battery during the discharge process.

[0149] The third slope may be the slope between the starting point and the ending point of the target voltage change curve, and the third slope may be used to characterize the voltage change speed of the target battery during the discharge process.

[0150] The average voltage value may be an average voltage value of a target voltage variation curve within a significant time period, and the average voltage value may be used as a reference to characterize an average voltage level of a target battery within a target time period.

[0151] Through the above method, it is convenient to predict the battery capacity of the target battery according to each target feature, and it is also convenient to calculate the feature distance between the target battery and other candidate batteries.

[0152] Step 104 , inputting a plurality of target features into the first target model to obtain the battery capacity of each target battery.

[0153] The first target model may be a trained prediction model used to estimate the battery capacity of the target battery according to the target feature. The first target model may be a Gaussian model, a regression model, a decision tree or other types of prediction models.

[0154] The battery capacity may be the amount of electrical energy that the target battery can store and provide for use.

[0155] It is understandable that if the battery remaining energy detection standard is followed, the time required to complete the remaining energy detection in sequence will be as high as more than 10 hours, and the equipment occupancy rate required for the test will also increase significantly, limiting the battery detection efficiency. By collecting the status indicator curve within a certain time period (such as within 30 minutes) and selecting the target feature corresponding to the target time period and inputting it into the first target model, the first target model can quickly predict the battery capacity according to the target feature to obtain the battery capacity of the target battery. Moreover, since the significance value corresponding to the target feature is also the highest (see above for the principle), the first target model can accurately predict the battery capacity based on the target feature.

[0156] It is understandable that in order to improve the accuracy of the first target model in predicting the battery capacity of the target battery, the first preset model may be trained. Exemplarily, the first target model may be trained through the following steps:

[0157] (C.1) obtaining a sample battery and testing the sample battery to obtain a second true sample capacity of the sample battery;

[0158] (C.2) charging the sample battery to a preset charging time to obtain a second sample state indicator curve graph of the sample battery;

[0159] (C.3) determining the target time period with the maximum sample significance value according to the distribution of sample significance values ​​of each time period in the preset sample significance map;

[0160] (C.4) extracting a target sample state index curve corresponding to the target time period from the second sample state index curve graph, and extracting a plurality of sample features corresponding to the sample battery based on the target sample state index curve;

[0161] (C.5) inputting the plurality of sample features into the first preset model to obtain a second sample prediction capacity;

[0162] (C.6) calculating a second loss of the first preset model based on the second sample prediction capacity and the second actual sample capacity;

[0163] (C.7) Based on the second loss, the model parameters of the first preset model are adjusted until the number of training times for the first preset model reaches a second preset number of times, thereby obtaining a first target model.

[0164] The second real sample capacity may be a battery capacity actually measured on a sample battery, which is used for comparison with the second sample predicted capacity to evaluate the prediction accuracy of the first preset model.

[0165] Among them, the preset charging time can be the time for fully charging the sample battery using a preset current. It is understandable that in order to improve the detection efficiency of the sample battery, the sample battery can be charged with a large current within the current range that the sample battery can withstand, so as to stimulate a larger electric potential and heat generation of the sample battery. For example, the magnitude of the current can be determined according to the actual working conditions of the sample battery, such as using a current of 1 / 2 / 3 / 4 / 5C or the like. Taking 1C as an example, a sample voltage change curve graph and a sample temperature change curve graph of the sample battery within 30 minutes can be collected under a 1C constant current charging state.

[0166] The second sample state index curve diagram may be a voltage change curve and a temperature change curve obtained by monitoring the charge and discharge process of the sample battery, and the voltage change curve and the temperature change curve reflect the performance index of the sample battery in different states.

[0167] The target time period may be a time period in which the maximum sample significance value is determined based on the sample significance map of sample batteries of the same type as the target battery.

[0168] The target sample state index curve may be a sample state index curve corresponding to a target time period extracted from the sample state index curve diagram, and the target sample state index curve may be used to reflect changes in sample state indicators of the sample battery within the target time period.

[0169] Among them, the sample feature can be the sample state index of the target sample state index curve under different target time periods or the change value of the sample state index. For example, when the sample state index curve diagram is a sample temperature change curve diagram, the sample feature can be the minimum temperature value, the time corresponding to the minimum temperature value, the slope between the starting point of the sample temperature change curve and the minimum temperature point corresponding to the minimum temperature value, the slope between the minimum temperature point and the end point of the sample temperature change curve, the difference between the starting temperature value and the minimum temperature value corresponding to the starting point of the sample temperature change curve, and the difference between the end temperature value and the minimum temperature value corresponding to the end point of the sample temperature change curve. When the sample state index curve diagram is a voltage change curve diagram, the sample feature can be the starting voltage value corresponding to the starting point of the sample voltage change curve, the ending voltage value corresponding to the end point of the sample voltage change curve, the difference between the ending voltage value and the starting voltage value, the slope corresponding to the starting point of the sample voltage change curve and the end point of the sample voltage change curve, and the average voltage value of the sample voltage change curve corresponding to the significant time period.

[0170] The first preset model may be an initial model for predicting the battery capacity according to the sample characteristics. The first preset model may be a Gaussian model, or a regression model, a decision tree, or other types of prediction models.

[0171] The second sample predicted capacity may be a battery capacity predicted by the first preset model according to input sample features.

[0172] The second loss may be the difference between the second sample predicted capacity predicted by the first preset model for the sample battery and the second real sample capacity. The second loss may be measured by a mean square error or other loss functions.

[0173] The second preset number of times may be the number of times the model parameters are repeatedly adjusted to optimize the model performance when training the first preset model, and the second preset number of times may be obtained by setting, for example, 500 times, 1000 times, etc. Alternatively, the second preset number of times may not be set, and the training of the first preset model may be stopped when the first preset model converges to obtain the first target model.

[0174] Exemplarily, when obtaining the second real sample capacity of the sample battery, the sample battery may be charged with a constant current of 0.2C to obtain the second real sample capacity.

[0175] For example, for sample battery A, sample battery A can be charged at a constant current of 0.2C to obtain the actual capacity of sample battery A. It is detected that the actual capacity of sample battery A is 5000mAh. Then, sample battery A is discharged at a constant current of 0.2C to a cut-off voltage.

[0176] Furthermore, a sample voltage change curve graph and a sample temperature change curve graph of a sample battery collected within 30 minutes under a 1C constant current charging state can be used to obtain a sample significance map obtained in advance through the first preset model training process. The category of the sample battery in the sample significance map should be the same as the category of the sample battery trained this time.

[0177] Furthermore, the target time period with the maximum sample significance value can be determined based on the distribution of sample significance values ​​in each time period in the preset sample significance graph. For example, from the 23rd to the 26th minute, the sample significance value is the largest, and the 23rd to the 26th minute can be determined as the target time period. Afterwards, the target sample voltage change curve and the target sample temperature change curve from the 23rd to the 26th minute are selected from the sample voltage change curve graph and the sample temperature change curve graph.

[0178] Specifically, multiple sample features of the sample battery can be extracted based on the target sample voltage change curve and the target sample temperature change curve, and the multiple sample features can be input into the first preset model to obtain the second sample predicted capacity output by the first preset model. Furthermore, the second loss of the first preset model can be calculated based on the second sample predicted capacity and the second true sample capacity, and then the model parameters of the first preset model can be adjusted according to the second loss until the preset number of training times is reached, such as when the number of training times reaches 500 times, the training of the first preset model is stopped to obtain the first target model. Alternatively, the training is stopped when the first preset model converges to obtain the first target model.

[0179] Through the above method, the first preset model can be continuously trained, so that the trained first target model can quickly and accurately predict the battery capacity of the target battery.

[0180] Step 105 , based on multiple target features and battery capacity corresponding to the target battery, calculate the feature distance between the target battery and other candidate batteries.

[0181] It is understandable that in order to facilitate subsequent battery selection and reorganization to improve the overall performance, safety and life of the battery pack while reducing maintenance costs, the characteristic distance between the target battery and other candidate batteries can be calculated to ensure that the battery pack can operate stably under different working conditions and meet the needs of energy management and system optimization.

[0182] The candidate battery may be a battery used in combination with the target battery. Specifically, the candidate battery should have performance compatibility, physical compatibility and good health with the target battery. The candidate battery may also be understood as other target batteries.

[0183] The feature distance may be a metric used to quantify the performance difference between the target battery and the candidate battery. For example, the feature distance may be Euclidean distance, Manhattan distance, or other distance metrics.

[0184] For example, when the target battery is battery A and one of the candidate batteries is battery B, taking the target feature as the minimum temperature value, when the minimum temperature value of battery A is 25°C and the minimum temperature value of battery B is 24°C, the Euclidean distance between battery A and battery B is calculated to be 1, and the characteristic distance between battery A and battery B for the target feature of the minimum temperature value is 1. Similarly, the characteristic distances of other target features and battery capacity can also be calculated in the same way as above, which will not be elaborated here.

[0185] It should be noted that if batteries are reorganized based only on battery capacity, the differences in the battery's internal chemical state, failure mode, and consistency in voltage rise during charging cannot be considered during the reorganization process. The reorganized batteries have disadvantages such as short service life and poor consistency. The target features carry key information about the degree of battery aging and have a strong correspondence with the battery's capacity and thermal behavior. Reorganizing batteries based on the feature distance calculated based on multiple target features and battery capacity can make the reorganized battery pack have better charge and discharge consistency and cycle life.

[0186] In some embodiments, in order to calculate the characteristic distance between the target battery and the candidate battery, the Euclidean distance may be calculated one by one between each target feature of the target battery and the candidate feature of the corresponding candidate battery, thereby obtaining the characteristic distance between the target battery and other candidate batteries. Exemplarily, step 105 may include:

[0187] (105.1) Obtain multiple candidate features and candidate battery capacities of candidate batteries;

[0188] (105.2) For each target feature, calculating the first Euclidean distance between the target feature and the candidate feature of the corresponding candidate battery;

[0189] (105.3) Calculate the second Euclidean distance between the battery capacity and the corresponding candidate battery capacity;

[0190] (105.4) Based on the plurality of first Euclidean distances and second Euclidean distances, a characteristic distance between the target battery and other candidate batteries is calculated.

[0191] Among them, the candidate features can be multiple features of the candidate batteries. Since each candidate battery can be understood as a target battery, the candidate features can also correspond to target features, that is, multiple target features extracted based on the target temperature change curve and the target voltage transformation curve.

[0192] The capacity of the candidate battery may be the amount of electrical energy that the candidate battery can store and release.

[0193] There may be multiple first Euclidean distances, and the Euclidean distance between each target feature and the corresponding candidate feature serves as a first Euclidean distance.

[0194] The second Euclidean distance may be the degree of difference between the target battery capacity and the candidate battery capacity.

[0195] For example, taking the target characteristic of the target battery as the average voltage value of the target voltage change curve corresponding to the target time period as an example, when the average voltage value of the target characteristic is 3.9V and the average voltage value of the candidate characteristic is 3.8V, then Similarly, the calculation methods of the first Euclidean distances between other target features and candidate features, and the second Euclidean distances between the battery capacity and the corresponding candidate battery capacity are the same as above and will not be repeated here.

[0196] Specifically, after all the first Euclidean distances and the second Euclidean distances are calculated, the weights of the first Euclidean distances and the second Euclidean distances can be adjusted according to the importance of each feature to obtain an adjusted feature distance. Furthermore, the adjusted weights can be flexibly set.

[0197] It should be noted that different target features and battery capacities may have different importance in matching. In order to more accurately calculate their influence in the matching process, different weights may be set for different target features and battery capacities to improve the accuracy and stability of battery recombination. For example, (105.4) may include:

[0198] (105.4.1) obtain the values ​​set for each target characteristic and the values ​​set for the battery capacity;

[0199] (105.4.2) For each first Euclidean distance, the first Euclidean distance is adjusted according to the first weight of the corresponding target feature to obtain;

[0200] (105.4.3) adjusting the second Euclidean distance according to the second weight to obtain a second characteristic distance;

[0201] (105.4.4) Calculate the characteristic distance between the target battery and other candidate batteries based on the sum of the multiple first characteristic distances and the second characteristic distances.

[0202] The first weight may be a weight set for the first Euclidean distance between each target feature and the corresponding candidate feature, and the first weight is used to adjust the first Euclidean distance. The first weights of different first Euclidean distances may be the same or different.

[0203] The second weight may be a weight set for the battery capacity and used to adjust the second Euclidean distance.

[0204] The first characteristic distance can be obtained by adjusting the first Euclidean distance by the first weight.

[0205] The second characteristic distance can be obtained by adjusting the second Euclidean distance through the second weight.

[0206] In some embodiments, the importance of different target features and battery capacities can be analyzed to determine the relative importance of each target feature and battery capacity to the matching degree, thereby determining the value of each first weight and second weight. Alternatively, the first weight and second weight can also be set based on experience.

[0207] Exemplarily, for each target feature and battery capacity, the following weights are set: the first Euclidean distance 1 is 1, and the corresponding first weight is 0.6; the first Euclidean distance 2 is 2.2, and the corresponding first weight is 0.4; the second Euclidean distance is 2.6, and the corresponding second weight is 0.3.

[0208] Thus, it can be obtained that the sum of the first characteristic distances and the second characteristic distances is: 1*0.6+2.2*0.4+2.6*0.3=2.26.

[0209] In the above manner, the corresponding Euclidean distance can be adjusted using different weights to more accurately evaluate the feature similarity between the target battery and the candidate batteries, so as to facilitate more accurate battery recombination in the subsequent process.

[0210] Step 106 , selecting a target candidate battery from a plurality of candidate batteries according to the characteristic distance, and recombining the target battery and the target candidate battery.

[0211] It can be understood that when the feature distance between the target battery and the candidate battery is close, it indicates that the feature similarity between the two is high and recombination is possible; when the feature distance between the target battery and the candidate battery is far, it indicates that the feature difference between the two is large and it is not suitable for recombination.

[0212] The target candidate battery may be a candidate battery whose characteristic distance with the target battery is less than a characteristic distance threshold, and the target candidate battery may be recombined with the target battery.

[0213] In some embodiments, a characteristic distance threshold may be set, and the characteristic distances of the target battery and each candidate battery are compared in turn to obtain a comparison result. When the comparison result indicates that the characteristic distance between the target battery and the candidate battery is less than the characteristic distance threshold, the corresponding candidate battery is determined as the target candidate battery. Furthermore, the characteristic distance threshold may be set according to actual conditions, for example, it may be set to 3, 6, etc.

[0214] Furthermore, multiple target candidate batteries can be sorted according to the characteristic distance to obtain a sorting result, and the number of battery reorganizations required for the target battery is obtained. Based on the sorting result, target candidate batteries with a number equal to the number of battery reorganizations and the smallest characteristic distance to the target battery are selected from the target candidate batteries.

[0215] Furthermore, the characteristic distances between multiple batteries can be calculated simultaneously, and a clustering algorithm can be used to cluster and reorganize the batteries based on the characteristic distances. Exemplarily, the clustering algorithm can be k-means clustering, or a clustering analysis model can be used, such as dynamic time warping, curve shape-based clustering, kernel method clustering, deep learning-based clustering, etc., to achieve rapid reorganization of multiple batteries.

[0216] In the above manner, by setting a feature distance threshold, candidate batteries with similar features to the target battery can be screened out, and target candidate batteries suitable for recombination can be quickly found, reducing the computational complexity and time consumption in the matching process and improving the matching efficiency.

[0217] Please refer to Figure 4 and Figure 5 In some embodiments, the first target model can be a support vector machine, a random forest, a decision tree, a Gaussian model, and an adaptive boosting model. This application uses the constructed target features to make predictions on the lithium iron phosphate / graphite dataset. Figure 4 It can be seen that the results of the prediction by the method of the present application show that the mean absolute percentage error (MAPE) of the prediction is very low, the error between the predicted value and the actual value of the first target model is small, the prediction contribution is high, and the determination coefficient (R-squared, R 2 ) is close to 1, and the fit of the first target model is also good. Figure 5 In this application, the constructed target features are used to make predictions on lithium manganese oxide / graphite batteries, and good results are also achieved.

[0218] The present application obtains multiple state index curves of each target battery during the charging process; according to the distribution of sample significant values ​​of each sample time period in the sample significant graph of the sample battery, the target time period with the maximum sample significant value is determined, and the target state index curve corresponding to the target time period is extracted from the state index curve; wherein the sample significant value is determined by the contribution of the sample state index when predicting the battery capacity of the sample battery in the corresponding sample time period; based on the target state index curve, multiple target features corresponding to the target battery are extracted; multiple target features are input into the first target model to obtain the battery capacity of each target battery; based on the multiple target features and battery capacity corresponding to the target battery, the characteristic distance between the target battery and other candidate batteries is calculated; according to the characteristic distance, the target candidate battery is selected from the multiple candidate batteries, and the target battery and the target candidate battery are reorganized. In this way, the target feature that is conducive to accurately predicting the battery capacity of the target battery can be directly selected and input into the first target model, so as to realize the rapid and accurate prediction of the battery capacity. Moreover, in the process of selecting the battery for reorganization for the target battery, the reorganization is not only based on the battery capacity, but also based on the characteristic distance between other multiple target features, so as to fully consider the internal state between the reorganized batteries and improve the accuracy of battery reorganization. In summary, the present application can improve the efficiency and accuracy of battery reorganization.

[0219] See also Figure 6 The present application also provides a battery reassembly device, which can implement the above-mentioned battery reassembly method. The battery reassembly device includes:

[0220] An acquisition module 61 is used to acquire multiple state indicator curve graphs of each target battery during the charging process;

[0221] A determination module 62 is used to determine the target time period with the maximum sample significant value according to the distribution of the sample significant values ​​in each sample time period in the sample significant graph of the sample battery, and extract the target state index curve corresponding to the target time period from the state index curve graph; wherein the sample significant value is determined by the contribution of the sample state index in predicting the battery capacity of the sample battery in the corresponding sample time period;

[0222] An extraction module 63, used to extract multiple target features corresponding to the target battery based on the target state index curve;

[0223] An input module 64, used to input a plurality of target features into the first target model to obtain a battery capacity of each target battery;

[0224] A calculation module 65, configured to calculate a characteristic distance between the target battery and other candidate batteries based on a plurality of target characteristics and a battery capacity corresponding to the target battery;

[0225] The reorganization module 66 is used to select a target candidate battery from a plurality of candidate batteries according to the characteristic distance, and to reorganize the target battery and the target candidate battery.

[0226] The specific implementation of the battery reassembly device is basically the same as the specific implementation of the battery reassembly method described above, and will not be repeated here. On the premise of meeting the requirements of the embodiment of the present application, the battery reassembly device can also be provided with other functional modules to implement the battery reassembly method in the above embodiment.

[0227] The embodiment of the present application also provides a computer device, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned battery recombination method when executing the computer program. The computer device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0228] See also Figure 7 , Figure 7 The hardware structure of a computer device according to another embodiment is shown, and the computer device includes:

[0229] The processor 71 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0230] The memory 72 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 72 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 72, and the processor 71 is called to execute the battery reassembly method of the embodiment of this application;

[0231] Input / output interface 73, used to implement information input and output;

[0232] Communication interface 74, used to realize communication interaction between the device and other devices, which can be realized by wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0233] A bus 75 that transmits information between the various components of the device (e.g., the processor 71, the memory 72, the input / output interface 73, and the communication interface 74);

[0234] The processor 71 , the memory 72 , the input / output interface 73 and the communication interface 74 are connected to each other in communication within the device via a bus 75 .

[0235] The embodiment of the present application further provides a computer-readable storage medium, which stores a computer program, and the computer program implements the above-mentioned battery recombination method when executed by a processor.

[0236] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0237] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0238] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0239] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0240] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0241] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0242] It should be understood that in the present application, "at least one (item)" and "several" refer to one or more, and "plurality" refers to two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and subsequent associated objects are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0243] In the several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0244] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0245] In addition, each functional unit 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. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0246] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0247] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A battery reassembly method, characterized in that: The method comprises: Obtain multiple state indicator curve graphs of each target battery during the charging process; According to the distribution of sample significant values ​​of each sample time period in the sample significant graph of the sample battery, determine the target time period with the maximum sample significant value, and extract the target state index curve corresponding to the target time period from the state index curve graph; wherein the sample significant value is determined by the contribution of the sample state index when predicting the battery capacity of the sample battery in the corresponding sample time period; Based on the target state index curve, extracting a plurality of target features corresponding to the target battery; Input multiple target features into the first target model to obtain the battery capacity of each target battery; wherein the first target model is trained by the following steps: obtain a sample battery, and test the sample battery to obtain a second real sample capacity of the sample battery; charge the sample battery to a preset charging time to obtain a second sample state index curve diagram of the sample battery; determine the target time period with the maximum sample significance value according to the distribution of sample significance values ​​in each time period in the preset sample significance diagram; extract the target sample state index curve corresponding to the target time period from the second sample state index curve diagram, and extract multiple sample features corresponding to the sample battery based on the target sample state index curve; input multiple sample features into the first preset model to obtain a second sample predicted capacity; calculate the second loss of the first preset model based on the second sample predicted capacity and the second real sample capacity; adjust the model parameters of the first preset model based on the second loss until the training times of the first preset model reach the second preset times, and obtain the first target model; Based on multiple target features and battery capacity corresponding to the target battery, calculating a feature distance between the target battery and other candidate batteries; A target candidate battery is selected from the plurality of candidate batteries according to the characteristic distance, and the target battery and the target candidate batteries are recombined.

2. The battery reassembly method according to claim 1, characterized in that: The calculating, based on the multiple target features and battery capacity corresponding to the target battery, the feature distance between the target battery and other candidate batteries includes: Acquire a plurality of candidate features and a candidate battery capacity of the candidate battery; For each of the target features, calculating a first Euclidean distance between the target feature and the corresponding candidate feature of the candidate battery; Calculating a second Euclidean distance between the battery capacity and the corresponding candidate battery capacity; Based on the first Euclidean distances and the second Euclidean distances, characteristic distances between the target battery and other candidate batteries are calculated.

3. The battery reassembly method according to claim 2, characterized in that: The calculating, based on the plurality of the first Euclidean distances and the second Euclidean distances, the characteristic distances between the target battery and other candidate batteries includes: Acquire a first weight set for each of the target features and a second weight set for the battery capacity; For each of the first Euclidean distances, adjusting the first Euclidean distance according to the first weight of the corresponding target feature to obtain a first feature distance; Adjusting the second Euclidean distance according to the second weight to obtain a second characteristic distance; The characteristic distances between the target battery and other candidate batteries are calculated according to the sum of the first characteristic distances and the second characteristic distances.

4. The battery reassembly method according to claim 1, characterized in that: The sample significance value is obtained by the following steps: Acquire a plurality of sample batteries, and for each of the sample batteries, determine the contribution of each sample state indicator when the second target model predicts the battery capacity of the sample battery in a corresponding sample time period based on the sample state indicator after the sample state indicator is input into the second target model; For each of the sample time periods, calculating a plurality of contribution degrees of a plurality of the sample state indicators of a plurality of the sample batteries in the sample time period; The average value of the multiple contribution degrees in each sample time period is calculated to obtain the sample significance value of each sample time period.

5. The battery reassembly method according to claim 4, characterized in that: The second target model is trained by the following steps: Acquire a plurality of first sample state indicator curve graphs corresponding to a plurality of sample batteries, and add noise data to each of the first sample state indicator curve graphs in any sample time period to obtain a first sample state indicator curve graph processed with the noise data; Inputting the sample state indicators in each of the first sample state indicator curve graphs into the second preset model in chronological order, and obtaining the first sample predicted capacity of the sample battery predicted by the second preset model according to the first sample state indicator curve graph; Acquire a first real sample capacity of the sample battery, and calculate a first loss of the second preset model based on the first sample predicted capacity and the first real sample capacity; Based on the first loss, the model parameters of the second preset model are adjusted until the number of training times for the second preset model reaches the first preset number of times, so as to obtain a second target model.

6. The battery reassembly method according to claim 1, characterized in that: The state index curve graph includes a temperature change curve graph and a voltage change curve graph. Based on the target state index curve, extracting multiple target features corresponding to the target battery includes: Based on the target temperature change curve determined from the temperature change curve graph, determine a minimum temperature value, a first moment corresponding to the minimum temperature value, a first slope between a starting point of the target temperature change curve and a minimum temperature point corresponding to the minimum temperature value, a second slope between the minimum temperature point and an end point of the target temperature change curve, a first difference between a starting temperature value corresponding to the starting point of the target temperature change curve and the minimum temperature value, and a second difference between an end temperature value corresponding to the end point of the target temperature change curve and the minimum temperature value; Based on the target voltage change curve determined from the voltage change curve diagram, determine the voltage starting value corresponding to the starting point of the target voltage change curve, the voltage ending value corresponding to the ending point of the target voltage change curve, the third difference between the voltage ending value and the voltage starting value, the third slope corresponding to the starting point of the target voltage change curve and the ending point of the target voltage change curve, and the average voltage value of the target voltage change curve corresponding to the target time period.

7. A battery reorganization device, characterized in that: The device comprises: An acquisition module, used to acquire multiple state indicator curve graphs of each target battery during the charging process; A determination module, configured to determine a target time period having a maximum sample significant value according to the distribution of sample significant values ​​of each sample time period in the sample significant graph of the sample battery, and extract a target state index curve corresponding to the target time period from the state index curve graph; wherein the sample significant value is determined by the contribution of the sample state index in predicting the battery capacity of the sample battery in the corresponding sample time period; An extraction module, used for extracting a plurality of target features corresponding to the target battery based on the target state index curve; An input module, used for inputting a plurality of the target features into a first target model to obtain a battery capacity of each of the target batteries; wherein the first target model is trained by the following steps: obtaining a sample battery, and testing the sample battery to obtain a second real sample capacity of the sample battery; charging the sample battery to a preset charging time to obtain a second sample state index curve diagram of the sample battery; determining a target time period having a maximum sample significance value according to the distribution of sample significance values ​​in each time period in a preset sample significance diagram; extracting a target sample state index curve corresponding to the target time period from the second sample state index curve diagram, and extracting a plurality of sample features corresponding to the sample battery based on the target sample state index curve; inputting a plurality of the sample features into a first preset model to obtain a second sample predicted capacity; calculating a second loss of the first preset model based on the second sample predicted capacity and the second real sample capacity; adjusting a model parameter of the first preset model based on the second loss until the number of training times for the first preset model reaches a second preset number of times to obtain a first target model; A calculation module, configured to calculate a characteristic distance between the target battery and other candidate batteries based on a plurality of target characteristics and a battery capacity corresponding to the target battery; A reorganization module is used to select a target candidate battery from the plurality of candidate batteries according to the characteristic distance, and to reorganize the target battery with the target candidate batteries.

8. A computer device, characterized in that: The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the battery recombination method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the battery reassembly method according to any one of claims 1 to 6 is implemented.

Images