Battery long-term performance screening method and system

Abstract

The invention provides a battery long-term performance screening method and system. The method comprises the steps of obtaining a training set charge-discharge capacity increment curve and a test set charge-discharge capacity increment curve; extracting training sequence fragment data and test sequence fragment data in different first voltage attention windows; performing feature extraction on the training sequence fragment data and the test sequence fragment data to obtain a training sequence feature set and a test sequence feature set; screening a second voltage attention window meeting a preset first condition; obtaining a training sequence feature set in the second voltage attention window, and training a preset classification model to obtain a performance screening model meeting a preset second condition; and obtaining a test sequence feature set in the second voltage attention window, inputting the test sequence feature set into the performance screening model, and performing battery long-term performance screening on the test set battery samples. The method solves the problem that the accuracy of battery long-term performance screening is low due to single input data category and lack of partial key information in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of rapid screening of long-term performance of lithium-ion batteries, and in particular to a method and system for screening long-term performance of batteries. Background Art

[0002] The application fields of lithium-ion batteries are becoming more and more extensive. Batteries with poor long-term performance have a shorter cycle life and will experience accelerated aging during the cycle. Batteries with poor long-term performance will often experience serious loss of active materials and active lithium ions. Accelerated aging will not only affect the long-term performance of the battery, but also bring safety risks during use. Therefore, in the product development stage, it is particularly important to screen excellent batteries to ensure battery performance and safety and improve the overall efficiency and life of the battery pack. From the overall development perspective, battery screening technology is moving towards intelligence and refinement.

[0003] At present, traditional battery screening technology usually relies on actual long-term charge and discharge cycle tests. Although this method can provide relatively accurate data, it is extremely time-consuming and often takes months or even years to complete. Specific battery screening technologies include accelerated aging early diagnosis methods that use combined machine learning, IC curves, DV curves, and discharge curve characteristics, as well as fast-charging battery pack cell screening methods based on three-stage constant-current fast-charging data, and three-level screening methods for cascade batteries based on the open-circuit voltage, voltage change value, and impedance value of the power battery at different charges. The main problems with the existing technology are that the long-term performance test cycle is long and requires a lot of test time and test resources. In addition, because there is no obvious difference in capacity between batteries with long cycle life and short cycle life in the early aging stage, it is impossible to accurately judge the quality of the battery's long-term performance based on the early cycle capacity data. In general, the input data category of the existing technology is single, and some key information is missing, resulting in low accuracy in battery long-term performance screening. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a battery long-term performance screening method and system, which fully considers the sequence fragment data characteristics and voltage attention window interval that have a key impact on the long-term performance of the battery, and improves the accuracy of battery long-term performance screening.

[0005] To achieve the above object, an embodiment of the present invention provides a method for screening long-term performance of a battery, comprising:

[0006] Selecting training set battery samples and test set battery samples to obtain training set charge and discharge capacity increment curves and test set charge and discharge capacity increment curves;

[0007] Extracting training sequence fragment data and test sequence fragment data in different first voltage attention windows based on a preset first voltage attention window, a training set charge-discharge capacity increment curve, and a test set charge-discharge capacity increment curve;

[0008] Performing feature extraction on the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set;

[0009] Training a preset classification model based on a training sequence feature set, and screening a second voltage attention window that meets a preset first condition;

[0010] Acquire a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window, and train a preset classification model to obtain a performance screening model that meets the preset second condition;

[0011] A test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window is obtained and input into the performance screening model to perform long-term battery performance screening on the test set battery samples.

[0012] The embodiment of the present invention proposes a method for screening long-term performance of batteries, which obtains training set charge and discharge capacity increment curves and test set charge and discharge capacity increment curves corresponding to training set battery samples and test set battery samples, and then obtains different first voltage attention windows in combination with the attention mechanism, obtains training sequence fragment data and test sequence fragment data in different first voltage attention windows, and training sequence feature sets and test sequence feature sets, and finally achieves screening of long-term performance of batteries through model training; the sequence features of the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are intercepted through different first voltage attention windows to screen the long-term performance of batteries, which increases the diversity of data input, so that the machine model can automatically focus on those sequence fragment data features and voltage attention window intervals that have a key impact on the long-term performance of the battery, thereby improving the accuracy of long-term performance judgment of the battery; in addition, by screening the second voltage attention window that meets the preset first condition, the machine learning model can perform long-term performance judgment on the test set battery samples for the second voltage attention window, and by judging and processing the key sequence fragment data features in the second voltage window, the utilization rate of data and the accuracy of judgment are further improved.

[0013] Furthermore, the selecting of training set battery samples and test set battery samples to obtain the training set charge-discharge capacity increment curve and the test set charge-discharge capacity increment curve includes:

[0014] Selecting a number of battery samples, and dividing the battery samples into training set battery samples and test set battery samples;

[0015] By performing cyclic testing experiments on the training set battery samples and the test set battery samples, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are obtained.

[0016] Furthermore, the method of performing a cyclic test experiment on the training set battery samples and the test set battery samples to obtain the training set charge-discharge capacity increment curve and the test set charge-discharge capacity increment curve includes:

[0017] Performing a cyclic testing experiment on the training set battery samples and the test set battery samples;

[0018] Whenever the cycle test experiment meets the preset cycle number interval, the current training set battery sample and the current test set battery sample corresponding to the preset cycle number interval are obtained;

[0019] Performing charge and discharge tests on the current training set battery samples and the current test set battery samples corresponding to the preset cycle number intervals to obtain the current training set charge and discharge curve and the current test set charge and discharge curve;

[0020] When the cycle test experiment is finished, based on several current training set charge and discharge curves and current test set charge and discharge curves, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are calculated.

[0021] Further, extracting the training sequence segment data and the test sequence segment data in different first voltage attention windows based on the preset first voltage attention window, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve includes:

[0022] Based on a preset first voltage attention window, obtaining a first starting voltage and a first voltage interval;

[0023] Based on the first starting voltage and the first voltage interval, sequence extraction is performed on the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve respectively to obtain training sequence fragment data and test sequence fragment data in different first voltage attention windows.

[0024] The embodiment of the present invention proposes a method for screening long-term battery performance. After combining the attention mechanism and the charge and discharge capacity increment curve, machine learning can concentrate on learning knowledge that is strongly related to the long-term performance of the battery from part of the charge and discharge capacity increment curve by changing the width and position of the voltage attention window, thereby improving the training accuracy and judgment accuracy of the machine learning model in judging accelerated battery aging. At the same time, combined with the attention window, multiple voltage ranges that enable the machine learning model to accurately judge accelerated aging can be selected from the charge and discharge capacity increment curve. Compared with the traditional machine learning method of extracting features from the complete capacity increment curve, the data utilization and judgment accuracy are improved.

[0025] Furthermore, the extracting features of the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set includes:

[0026] Respectively extracting the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the training sequence segment data and the test sequence segment data;

[0027] Based on the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the training sequence segment data and the test sequence segment data, a training sequence feature set and a test sequence feature set are obtained.

[0028] Furthermore, the training sequence feature set and the test sequence feature set further include: training set battery sample labels and test set battery sample labels;

[0029] The training set battery sample labels and the test set battery sample labels are obtained by obtaining the attenuation trends of the training set battery samples and the test set battery samples based on the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve, and labeling the training set battery samples and the test set battery samples based on the attenuation trends of the training set battery samples and the test set battery samples.

[0030] Furthermore, the step of training a preset classification model based on a training sequence feature set and screening a second voltage attention window that meets a preset first condition includes:

[0031] Input the training sequence feature set and training set battery sample labels extracted from the training set battery samples into a preset classification model for classification training and scoring, and output the scoring values ​​of the training sequence segment data in different first voltage attention windows;

[0032] The score values ​​of the training sequence segment data in different first voltage attention windows are sorted, and the first voltage attention windows corresponding to the training sequence segment data that meet the preset first condition are selected as the second voltage attention windows.

[0033] Furthermore, the step of acquiring a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window and training a preset classification model to obtain a performance screening model that meets the preset second condition includes:

[0034] Based on the second voltage attention window, acquiring a second starting voltage and a second voltage interval;

[0035] extracting a training sequence feature set corresponding to the training sequence segment data within the second voltage attention window based on the second starting voltage and the second voltage interval;

[0036] Based on the training sequence feature set and the training set battery sample labels corresponding to the training sequence segment data in the second voltage attention window, the preset classification model is trained to obtain a performance screening model that meets the preset second condition.

[0037] Further, the step of obtaining the test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window and inputting the feature set into the performance screening model to perform long-term battery performance screening on the test set battery samples includes:

[0038] extracting a test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window based on the second voltage attention window and the test sequence fragment data in the second voltage attention window;

[0039] Based on the test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window, the long-term performance of the battery samples in the test set is screened and judged through the performance screening model to obtain the long-term performance judgment results of each battery; wherein the long-term performance judgment results of each battery include: normally aged batteries and accelerated aged batteries.

[0040] The embodiment of the present invention also provides a battery long-term performance screening system, including: an incremental curve acquisition module, a sequence fragment extraction module, a sequence feature set extraction module, a second voltage attention window acquisition module, a performance screening model training module and a battery long-term performance screening module;

[0041] The incremental curve acquisition module is used to select training set battery samples and test set battery samples to obtain the training set charge and discharge capacity incremental curve and the test set charge and discharge capacity incremental curve;

[0042] The sequence segment extraction module is used to extract training sequence segment data and test sequence segment data in different first voltage attention windows based on a preset first voltage attention window, a training set charge-discharge capacity increment curve, and a test set charge-discharge capacity increment curve;

[0043] The sequence feature set extraction module is used to extract features from the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set;

[0044] The second voltage attention window acquisition module is used to train a preset classification model based on a training sequence feature set, and screen a second voltage attention window that meets a preset first condition;

[0045] The performance screening model training module is used to obtain a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window, and train a preset classification model to obtain a performance screening model that meets the preset second condition;

[0046] The battery long-term performance screening module is used to obtain the test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window, and input it into the performance screening model to perform battery long-term performance screening on the test set battery samples.

[0047] The embodiment of the present invention proposes a battery long-term performance screening system, which obtains the training set charge and discharge capacity incremental curve and the test set charge and discharge capacity incremental curve corresponding to the training set battery samples and the test set battery samples through the incremental curve acquisition module, and then obtains different first voltage attention windows through the sequence fragment extraction module and the sequence feature set extraction module in combination with the attention mechanism, obtains the training sequence fragment data and the test sequence fragment data in the different first voltage attention windows, as well as the training sequence feature set and the test sequence feature set, and finally the performance screening model training module and the battery long-term performance screening module implement the screening of the battery long-term performance through model training; intercepts the training set charge and discharge capacity incremental curves through different first voltage attention windows The long-term performance of the battery is screened based on the sequence characteristics of the capacity increment curve and the test set charge and discharge capacity increment curve, which increases the diversity of data input and enables the machine model to automatically focus on those sequence fragment data characteristics and voltage attention window intervals that have a key impact on the long-term performance of the battery, thereby improving the accuracy of the long-term performance judgment of the battery; in addition, the second voltage attention window acquisition module is used to screen the second voltage attention window that meets the preset first condition, so that the machine learning model can make long-term performance judgments on the test set battery samples based on the second voltage attention window, and further improve the data utilization and judgment accuracy by judging and processing the key sequence fragment data characteristics in the second voltage window. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic flow chart of the steps of a method for screening long-term performance of a battery provided by an embodiment of the present invention;

[0049] Figure 2 A schematic diagram of two long-term attenuation trends of lithium batteries in a battery long-term performance screening method provided by an embodiment of the present invention;

[0050] Figure 3 A schematic diagram of a normal aging attenuation curve of a lithium battery long-term performance test in a battery long-term performance screening method provided by an embodiment of the present invention;

[0051] Figure 4A schematic diagram of an accelerated aging attenuation curve of a lithium battery long-term performance test in a battery long-term performance screening method provided by an embodiment of the present invention;

[0052] Figure 5 A schematic diagram of a charge-discharge IC curve of a lithium battery before cycling according to a method for screening long-term battery performance provided by a certain embodiment of the present invention;

[0053] Figure 6 A schematic diagram of a 100-cycle charge and discharge IC curve of a lithium battery according to a battery long-term performance screening method provided by an embodiment of the present invention;

[0054] Figure 7 A schematic diagram of a 200-cycle charge-discharge IC curve of a lithium battery according to a battery long-term performance screening method provided by an embodiment of the present invention;

[0055] Figure 8 A schematic diagram of the module structure of a battery long-term performance screening system provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] With the rapid development of lithium batteries, screening excellent batteries can ensure battery performance and safety. At present, the more common battery screening strategy is to perform long-term performance screening on batteries. The traditional battery long-term performance evaluation and screening method usually relies on actual long-term charge and discharge cycle tests. Although this method can provide more accurate data results, the overall solution is time-consuming and not conducive to production and research and development; because batteries with poor long-term performance have a short cycle life and are prone to serious loss of active materials and active lithium ions. In addition, accelerated aging will occur during the cycle process, further affecting the safety performance of the battery pack.

[0058] Based on the problems existing in the prior art and the current battery screening methods mainly consider multiple parameters and characteristics of the battery, such as voltage, capacity, charge and discharge curves and impedance, the embodiment of the present invention proposes a battery long-term performance screening method to solve the problem that the long-term performance test cycle of the prior art is long and consumes a lot of test time and test resources, and solves the problem that the quality of the long-term performance of the battery cannot be accurately judged based on the early cycle capacity data, and there is no obvious difference in capacity between batteries with long cycle life and short cycle life in the early aging stage. For details, see Figure 2 , Figure 2 A schematic diagram of two long-term attenuation trends of lithium batteries in a battery long-term performance screening method provided by an embodiment of the present invention; Figure 2 As shown, the capacity characteristics of the accelerated aged and normally aged batteries in the early stages of cycling are very similar.

[0059] In the embodiment of the present invention, the charge and discharge IC curve and the charge and discharge capacity increment curve have the same meaning, which will not be described in detail below.

[0060] Example 1

[0061] See also Figure 1 , Figure 1 A schematic flow chart of the steps of a battery long-term performance screening method provided by an embodiment of the present invention. Figure 1 As shown, the embodiment of the present invention provides a battery long-term performance screening method, including steps 101 to 106, each step is specifically as follows:

[0062] Step 101, selecting training set battery samples and test set battery samples, and obtaining training set charge and discharge capacity increment curve and test set charge and discharge capacity increment curve;

[0063] As an example of this embodiment, a number of battery samples are selected, and the battery samples are divided into training set battery samples and test set battery samples; by performing a cyclic test experiment on the training set battery samples and the test set battery samples, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are obtained. Specifically, a cyclic test experiment is performed on the training set battery samples and the test set battery samples; whenever the cyclic test experiment meets the preset cycle number interval, the current training set battery samples and the current test set battery samples corresponding to the preset cycle number interval are obtained; the current training set battery samples and the current test set battery samples corresponding to the preset cycle number interval are subjected to charge and discharge tests to obtain the current training set charge and discharge curve and the current test set charge and discharge curve; when the cyclic test experiment is completed, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are calculated based on the current training set charge and discharge curves and the current test set charge and discharge curves.

[0064] A specific implementation method is to select battery samples, divide the battery samples into training set battery samples and test set battery samples, and carry out battery cycle decay experiments under different cycle conditions, wherein the different cycle conditions may be: before the cycle and during the cycle test, the battery is subjected to a 0.05C small current charge and discharge test every 100 cycles, the ambient temperature of the small current charge and discharge test is 25°C, and the upper and lower limit voltages of the battery charge and discharge are 3V and 4.35V respectively; the decay trend of the test battery includes normal aging and accelerated aging, and the decay trend of the normally aged battery is as follows: Figure 3 As shown, Figure 3A schematic diagram of a normal aging attenuation curve of a lithium battery long-term performance test in a battery long-term performance screening method provided by an embodiment of the present invention; the attenuation trend of an accelerated aging battery is as follows Figure 4 As shown, Figure 4 A schematic diagram of an accelerated aging decay curve of a lithium battery long-term performance test of a battery long-term performance screening method provided by an embodiment of the present invention; by conducting a battery cycle decay experiment, the battery cycle capacity decay data and the 0.05C small current charge and discharge curve and the charge and discharge capacity increment curve at different cycle periods are obtained; specifically, the battery small current charge and discharge curves before the cycle, 100 cycles and 200 cycles are extracted and the charge and discharge capacity increment curve is calculated, and the charge and discharge capacity increment curve before the cycle is as follows Figure 5 As shown, Figure 5 A schematic diagram of a lithium battery charge and discharge IC curve before cycling according to a battery long-term performance screening method provided by an embodiment of the present invention; a charge and discharge capacity increment curve after 100 cycles as shown in FIG. Figure 6 As shown, Figure 6 A schematic diagram of a lithium battery IC curve for 100 cycles of charge and discharge according to a method for long-term battery performance screening provided by an embodiment of the present invention; a curve of capacity increment for 200 cycles of charge and discharge as shown in FIG. Figure 7 As shown, Figure 7 A schematic diagram of a lithium battery IC curve for 200 cycles of charge and discharge according to a battery long-term performance screening method provided by an embodiment of the present invention; wherein the capacity increment calculation formula at different voltage positions is as follows:

[0065]

[0066] Where dQ / dV is the capacity increment, Q V1 and Q V2 are the capacities at voltages V1 and V2 respectively.

[0067] Step 102, extracting training sequence segment data and test sequence segment data in different first voltage attention windows based on a preset first voltage attention window, a training set charge-discharge capacity increment curve, and a test set charge-discharge capacity increment curve;

[0068] As an example of this embodiment, based on a preset first voltage attention window, a first starting voltage and a first voltage interval are obtained; based on the first starting voltage and the first voltage interval, sequence extraction is performed on the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve, respectively, to obtain training sequence fragment data and test sequence fragment data in different first voltage attention windows.

[0069] A specific implementation method is to set the initial voltage attention window (V0, V0+V interval), starting voltage V0∈(3.05V-4.05V), voltage interval V interval Take 0.1V, extract the sequence fragment data of the charge and discharge capacity increment curve of the training set and test set battery samples within the attention window, change the starting voltage V0 and the voltage attention window width V interval , extract the sequence fragment data of the charge and discharge capacity increment curve in different voltage attention windows {(V0, dQ / dV0), (V1, dQ / dV1), …, (V n , dQ / dV n )}.

[0070] Step 103, extracting features from the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set;

[0071] As an example of this embodiment, the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the training sequence fragment data and the test sequence fragment data are extracted respectively; based on the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the training sequence fragment data and the test sequence fragment data, the training sequence feature set and the test sequence feature set are obtained. The training sequence feature set and the test sequence feature set also include: training set battery sample labels and test set battery sample labels; the training set battery sample labels and the test set battery sample labels are obtained based on the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve, and the training set battery samples and the test set battery samples are labeled based on the decay trend of the training set battery samples and the test set battery samples.

[0072] A specific implementation method is to extract the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the charge and discharge capacity increment curve sequence fragment data in different voltage attention windows as feature parameters, construct a training sequence feature set and a test sequence feature set, and then mark the normal aging and accelerated aging batteries in the training set and the test set respectively. The normal aging battery is a battery that does not experience a capacity dive during the cycle decay process, and the accelerated aging battery is a battery that experiences a capacity dive during the cycle decay process.

[0073] Step 104, training a preset classification model based on the training sequence feature set, and screening a second voltage attention window that meets a preset first condition;

[0074] As an example of this embodiment, the training sequence feature set and the training set battery sample labels extracted from the training set battery samples are input into a preset classification model for classification training and scoring, and the scoring values ​​of the training sequence fragment data in different first voltage attention windows are output; the scoring values ​​of the training sequence fragment data in different first voltage attention windows are sorted, and the first voltage attention window corresponding to the training sequence fragment data that meets the preset first condition is selected as the second voltage attention window.

[0075] A specific implementation method is to use the sequence segment feature set of the charge and discharge capacity increment curve extracted from the training set and the battery sample labels of the training set to train a decision tree classification model or other machine learning classification model, wherein the other machine learning classification model can be: a support vector machine classification model, a neural network classification model, and a logistic regression classification model, etc.; by outputting the classification accuracy of each training set in the decision tree classification model or other machine learning model as the scoring value of the training sequence segment data in different first voltage attention windows, the calculation formula of the classification accuracy is as follows:

[0076]

[0077] Where N true is the number of correct battery samples, N all is the number of all battery samples;

[0078] Then, by sorting the score values ​​of the feature set data of each sequence segment, the first voltage attention window corresponding to the training sequence segment data that meets the preset first condition is selected as the second voltage attention window, which is equivalent to the optimal voltage attention window (V 0,best , V 0,best +V interval ), wherein the first preset condition is: the maximum value of the score value of each sequence segment feature set data; for more specific explanation, see Table 1:

[0079] Table 1 Battery long-term performance screening results Table 1

[0080] Cycle times Voltage lower limit Voltage upper limit Training set accuracy Before loop 3.35 3.45 100％ Before loop 3.85 3.95 100％ 100 3.25 3.35 100％ 100 3.35 3.45 100％ 100 3.65 3.75 100％ 100 3.95 4.05 100％ 200 3.15 3.25 100％

[0081] The voltage interval with 100% judgment accuracy of the training set decision tree classification model is taken as the optimal voltage attention interval (second voltage attention window) of the charge and discharge IC curve. It can be seen from Table 1 that the optimal voltage attention interval (second voltage attention window) before the cycle is (3.35V, 3.45V) and (3.85V, 3.95V), and the optimal voltage attention windows before the cycle, 100 cycles and 200 cycles are different.

[0082] Step 105, obtaining a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window, and training a preset classification model to obtain a performance screening model that meets the preset second condition;

[0083] As an example of this embodiment, based on the second voltage attention window, the second starting voltage and the second voltage interval are obtained; based on the second starting voltage and the second voltage interval, the training sequence feature set corresponding to the training sequence fragment data in the second voltage attention window is extracted; based on the training sequence feature set and the training set battery sample labels corresponding to the training sequence fragment data in the second voltage attention window, the preset classification model is trained to obtain a performance screening model that meets the preset second condition.

[0084] A specific implementation method is based on the second voltage attention window (V 0,best , V 0,best +V interval ), which is the second starting voltage V of the optimal voltage attention window 0,best and a second voltage interval, wherein the second voltage interval and the first voltage interval take the same value, both of which are 0.1V; extracting the training sequence feature set and the training set battery sample label corresponding to the training sequence fragment data in the optimal voltage attention window, training the decision tree classification model, and obtaining multiple trained decision tree classification models, and obtaining a performance screening model based on a preset second condition, wherein the preset second condition is: the optimal decision tree classification model trained in the optimal voltage attention window.

[0085] Step 106, obtaining a test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window, and inputting the feature set into the performance screening model to perform long-term battery performance screening on the test set battery samples.

[0086] As an example of this embodiment, based on the second voltage attention window and the test sequence fragment data within the second voltage attention window, a test sequence feature set corresponding to the test sequence fragment data within the second voltage attention window is extracted; based on the test sequence feature set corresponding to the test sequence fragment data within the second voltage attention window, the long-term performance of the battery samples in the test set is screened and judged through a performance screening model to obtain a long-term performance judgment result of each battery; wherein the long-term performance judgment result of each battery includes: normally aged batteries and accelerated aged batteries.

[0087] A specific implementation method is to obtain the second voltage attention window (V 0,best , V 0,best +V interval), and construct the corresponding test sequence feature set based on the test sequence fragment data, input the test sequence feature set into the performance screening model, and output the battery long-term performance judgment result through the performance screening model to judge whether the test set battery sample is a normal aging battery or an accelerated aging battery; based on the charge and discharge IC curve sequence fragments extracted by the optimal voltage attention window before the cycle, 100 cycles and 200 cycles, the accuracy of the decision tree classification model in both the training set and the test set can reach 100%, as shown in Table 2:

[0088] Table 2 Battery long-term performance screening results Table 2

[0089] Cycle times Voltage lower limit Voltage upper limit Training set accuracy Test set accuracy Before loop 3.35 3.45 100％ 100％ Before loop 3.85 3.95 100％ 100％ 100 3.25 3.35 100％ 100％ 100 3.35 3.45 100％ 100％ 100 3.65 3.75 100％ 100％ 100 3.95 4.05 100％ 100％ 200 3.15 3.25 100％ 100％

[0090] The embodiment of the present invention proposes a method for screening long-term performance of batteries, which obtains training set charge and discharge capacity increment curves and test set charge and discharge capacity increment curves corresponding to training set battery samples and test set battery samples, and then obtains different first voltage attention windows in combination with the attention mechanism, obtains training sequence fragment data and test sequence fragment data in different first voltage attention windows, and training sequence feature sets and test sequence feature sets, and finally achieves screening of long-term performance of batteries through model training; the sequence features of the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are intercepted through different first voltage attention windows to screen the long-term performance of batteries, which increases the diversity of data input, so that the machine model can automatically focus on those sequence fragment data features and voltage attention window intervals that have a key impact on the long-term performance of the battery, thereby improving the accuracy of long-term performance judgment of the battery; in addition, by screening the second voltage attention window that meets the preset first condition, the machine learning model can perform long-term performance judgment on the test set battery samples for the second voltage attention window, and by judging and processing the key sequence fragment data features in the second voltage window, the utilization rate of data and the accuracy of judgment are further improved.

[0091] Example 2

[0092] See also Figure 8 , Figure 8 A schematic diagram of the module structure of a battery long-term performance screening system provided by an embodiment of the present invention. Figure 8 As shown, an embodiment of the present invention proposes a battery long-term performance screening system, including: an incremental curve acquisition module, a sequence fragment extraction module, a sequence feature set extraction module, a second voltage attention window acquisition module, a performance screening model training module and a battery long-term performance screening module;

[0093] The incremental curve acquisition module is used to select training set battery samples and test set battery samples to obtain the training set charge and discharge capacity incremental curve and the test set charge and discharge capacity incremental curve;

[0094] The sequence segment extraction module is used to extract training sequence segment data and test sequence segment data in different first voltage attention windows based on a preset first voltage attention window, a training set charge-discharge capacity increment curve, and a test set charge-discharge capacity increment curve;

[0095] The sequence feature set extraction module is used to extract features from the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set;

[0096] The second voltage attention window acquisition module is used to train a preset classification model based on a training sequence feature set, and screen a second voltage attention window that meets a preset first condition;

[0097] The performance screening model training module is used to obtain a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window, and train a preset classification model to obtain a performance screening model that meets the preset second condition;

[0098] The battery long-term performance screening module is used to obtain the test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window, and input it into the performance screening model to perform battery long-term performance screening on the test set battery samples.

[0099] The embodiment of the present invention proposes a battery long-term performance screening system, which obtains the training set charge and discharge capacity incremental curve and the test set charge and discharge capacity incremental curve corresponding to the training set battery samples and the test set battery samples through the incremental curve acquisition module, and then obtains different first voltage attention windows through the sequence fragment extraction module and the sequence feature set extraction module in combination with the attention mechanism, obtains the training sequence fragment data and the test sequence fragment data in the different first voltage attention windows, as well as the training sequence feature set and the test sequence feature set, and finally the performance screening model training module and the battery long-term performance screening module implement the screening of the battery long-term performance through model training; intercepts the training set charge and discharge capacity incremental curves through different first voltage attention windows The long-term performance of the battery is screened based on the sequence characteristics of the capacity increment curve and the test set charge and discharge capacity increment curve, which increases the diversity of data input and enables the machine model to automatically focus on those sequence fragment data characteristics and voltage attention window intervals that have a key impact on the long-term performance of the battery, thereby improving the accuracy of the long-term performance judgment of the battery; in addition, the second voltage attention window acquisition module is used to screen the second voltage attention window that meets the preset first condition, so that the machine learning model can make long-term performance judgments on the test set battery samples based on the second voltage attention window, and further improve the data utilization and judgment accuracy by judging and processing the key sequence fragment data characteristics in the second voltage window.

[0100] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

[0101] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0102] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

Claims

1. A method for screening long-term battery performance, characterized in that: include: Selecting training set battery samples and test set battery samples to obtain training set charge and discharge capacity increment curves and test set charge and discharge capacity increment curves; Extracting training sequence fragment data and test sequence fragment data in different first voltage attention windows based on a preset first voltage attention window, a training set charge-discharge capacity increment curve, and a test set charge-discharge capacity increment curve; Performing feature extraction on the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set; Training a preset classification model based on a training sequence feature set, and screening a second voltage attention window that meets a preset first condition; Acquire a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window, and train a preset classification model to obtain a performance screening model that meets the preset second condition; A test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window is obtained and input into the performance screening model to perform long-term battery performance screening on the test set battery samples.

2. A battery long-term performance screening method as claimed in claim 1, characterized in that: The selecting of training set battery samples and test set battery samples to obtain the training set charge-discharge capacity increment curve and the test set charge-discharge capacity increment curve includes: Selecting a number of battery samples, and dividing the battery samples into training set battery samples and test set battery samples; By performing cyclic testing experiments on the training set battery samples and the test set battery samples, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are obtained.

3. A battery long-term performance screening method as claimed in claim 2, characterized in that: The method of performing a cyclic test experiment on the training set battery samples and the test set battery samples to obtain the training set charge-discharge capacity increment curve and the test set charge-discharge capacity increment curve comprises: Performing a cyclic testing experiment on the training set battery samples and the test set battery samples; Whenever the cyclic test experiment meets the preset cycle number interval, the current training set battery sample and the current test set battery sample corresponding to the preset cycle number interval are obtained; Performing charge and discharge tests on the current training set battery samples and the current test set battery samples corresponding to the preset cycle number intervals to obtain the current training set charge and discharge curve and the current test set charge and discharge curve; When the cycle test experiment is finished, based on several current training set charge and discharge curves and current test set charge and discharge curves, the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve are calculated.

4. A battery long-term performance screening method as claimed in claim 1, characterized in that: The extracting of training sequence fragment data and test sequence fragment data in different first voltage attention windows based on the preset first voltage attention window, the training set charge-discharge capacity increment curve and the test set charge-discharge capacity increment curve includes: Based on a preset first voltage attention window, obtaining a first starting voltage and a first voltage interval; Based on the first starting voltage and the first voltage interval, sequence extraction is performed on the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve respectively to obtain training sequence fragment data and test sequence fragment data in different first voltage attention windows.

5. A battery long-term performance screening method as claimed in claim 1, characterized in that: The extracting features of the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set includes: Respectively extracting the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the training sequence segment data and the test sequence segment data; Based on the maximum value, minimum value, average value, kurtosis, skewness, upper and lower limits of the voltage attention window and the average value of the voltage attention window of the training sequence segment data and the test sequence segment data, a training sequence feature set and a test sequence feature set are obtained.

6. A battery long-term performance screening method according to any one of claims 1 to 5, characterized in that: The training sequence feature set and the test sequence feature set also include: training set battery sample labels and test set battery sample labels; The training set battery sample labels and the test set battery sample labels are obtained by obtaining the attenuation trends of the training set battery samples and the test set battery samples based on the training set charge and discharge capacity increment curve and the test set charge and discharge capacity increment curve, and labeling the training set battery samples and the test set battery samples based on the attenuation trends of the training set battery samples and the test set battery samples.

7. A battery long-term performance screening method as claimed in claim 6, characterized in that: The method of training a preset classification model based on a training sequence feature set and selecting a second voltage attention window that meets a preset first condition includes: Input the training sequence feature set and training set battery sample labels extracted from the training set battery samples into a preset classification model for classification training and scoring, and output the scoring values ​​of the training sequence segment data in different first voltage attention windows; The score values ​​of the training sequence segment data in different first voltage attention windows are sorted, and the first voltage attention windows corresponding to the training sequence segment data that meet the preset first condition are selected as the second voltage attention windows.

8. A battery long-term performance screening method as claimed in claim 6, characterized in that: The step of acquiring a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window and training a preset classification model to obtain a performance screening model that meets the preset second condition includes: Based on the second voltage attention window, acquiring a second starting voltage and a second voltage interval; extracting a training sequence feature set corresponding to the training sequence segment data within the second voltage attention window based on the second starting voltage and the second voltage interval; Based on the training sequence feature set and the training set battery sample labels corresponding to the training sequence segment data in the second voltage attention window, the preset classification model is trained to obtain a performance screening model that meets the preset second condition.

9. A battery long-term performance screening method as claimed in claim 6, characterized in that: The step of obtaining a test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window and inputting the feature set into the performance screening model to perform long-term battery performance screening on the test set battery samples includes: extracting a test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window based on the second voltage attention window and the test sequence fragment data in the second voltage attention window; Based on the test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window, the long-term performance of the battery samples in the test set is screened and judged through the performance screening model to obtain the long-term performance judgment results of each battery; wherein the long-term performance judgment results of each battery include: normally aged batteries and accelerated aged batteries.

10. A battery long-term performance screening system, characterized in that: include: Incremental curve acquisition module, sequence fragment extraction module, sequence feature set extraction module, second voltage attention window acquisition module, performance screening model training module and battery long-term performance screening module; The incremental curve acquisition module is used to select training set battery samples and test set battery samples to obtain the training set charge and discharge capacity incremental curve and the test set charge and discharge capacity incremental curve; The sequence segment extraction module is used to extract training sequence segment data and test sequence segment data in different first voltage attention windows based on a preset first voltage attention window, a training set charge-discharge capacity increment curve, and a test set charge-discharge capacity increment curve; The sequence feature set extraction module is used to extract features from the training sequence segment data and the test sequence segment data to obtain a training sequence feature set and a test sequence feature set; The second voltage attention window acquisition module is used to train a preset classification model based on a training sequence feature set, and screen a second voltage attention window that meets a preset first condition; The performance screening model training module is used to obtain a training sequence feature set corresponding to the training sequence segment data in the second voltage attention window, and train a preset classification model to obtain a performance screening model that meets the preset second condition; The battery long-term performance screening module is used to obtain the test sequence feature set corresponding to the test sequence fragment data in the second voltage attention window, and input it into the performance screening model to perform battery long-term performance screening on the test set battery samples.

Images