Battery state of health prediction method, device, battery cell, medium and product

Abstract

The application discloses a battery health state prediction method and device, a battery monomer, a medium and a product, and belongs to the technical field of batteries. The battery health state prediction method comprises the following steps: acquiring charge-discharge curve data of a to-be-tested battery monomer in an initial use stage; extracting an electric signal feature corresponding to the charge-discharge curve data based on the charge-discharge curve data, wherein the electric signal feature comprises at least one of a peak value feature of a first peak of the charge-discharge curve, a feature of a first charging platform segment voltage curve and a feature of a charge-discharge starting segment energy curve; and performing prediction based on the electric signal feature by using a service life prediction model to obtain health state attenuation information. The battery health state prediction method can improve the accuracy, authenticity and accuracy of the prediction result, and can reduce the dependence on time sequence data such as cycle time which is strongly related to the cycle number.

Description

Technical Field

[0001] This application relates to the field of battery technology, and more specifically, to a method, apparatus, battery cell, dielectric, and product for predicting the health status of a battery. Background Technology

[0002] In related technologies, the remaining lifespan of lithium-ion batteries is often predicted based on data such as usage time, terminal voltage, output current, load voltage and current, and battery temperature. However, this method relies on a large amount of time-series data after the battery begins cycling, making it unsuitable for scenarios requiring rapid lifespan prediction (such as battery factory quality inspection and pre-installation evaluation of new batteries), and its prediction accuracy is not high. Summary of the Invention

[0003] This application provides a method, apparatus, battery cell, medium, and product for predicting the health status of a battery, to improve the accuracy, realism, and precision of the prediction results, and to reduce the dependence on time-series data such as cycle time that are strongly correlated with the number of cycles.

[0004] In a first aspect, embodiments of this application provide a method for predicting the health status of a battery, including:

[0005] Acquire charge-discharge curve data of the battery cell under test during the initial use phase;

[0006] Based on the charge-discharge curve data, the electrical signal features corresponding to the charge-discharge curve data are extracted. The electrical signal features include at least one of the peak features of the capacity versus voltage rate curve, the target features of the charge-discharge plateau segment electrical signal curve, and the target features of the charge-discharge initiation segment electrical signal curve.

[0007] Based on the characteristics of the electrical signal, a lifetime prediction model is used to predict and obtain information on the decline of health status;

[0008] The target feature includes at least one of sample entropy and kurtosis.

[0009] In the above technical solution, by using the electrical signal characteristics extracted from the charge-discharge curve data of the battery cell under test during the initial use stage for lifetime prediction, and introducing intrinsic features that are highly correlated with the aging mechanism, combined with the guidance of electrochemical mechanism, the accuracy, authenticity and accuracy of the prediction results can be improved, and the dependence on time-series data that are strongly correlated with the number of cycles, such as cycle time, can be reduced. It is suitable for a variety of application scenarios and has high universality.

[0010] In some embodiments, the target features of the charge / discharge plateau segment electrical signal curve include at least one of the sample entropy of the voltage curve of the first charging plateau segment and the kurtosis of the energy curve of the first charging plateau segment.

[0011] The target features of the electrical signal curve at the start of the charge / discharge phase include at least one of the sample entropy of the energy curve at the start of the discharge phase and the kurtosis of the voltage curve at the start of the charge phase, wherein the kurtosis is used to characterize the flatness of the curve.

[0012] In the above technical solution, by selecting electrical signal characteristics directly related to the internal electrochemical process of the battery, such as the peak value of the first peak of the charge-discharge curve, the sample entropy of the voltage curve of the first charging plateau segment, the kurtosis of the energy curve of the first charging plateau segment, the sample entropy of the energy curve of the discharge initiation segment, and the kurtosis of the voltage curve of the charging initiation segment, aging prediction can be performed in conjunction with electrochemical mechanism guidance, thereby improving the accuracy, authenticity, and reliability of the prediction results.

[0013] In some embodiments, the electrical signal features further include:

[0014] The peak and valley characteristics of the capacity versus voltage rate curve, the potential characteristics of the charge and discharge curve, the difference between the charging voltage and the discharging voltage, the voltage seasonality of the charging curve, the trend of the charge and discharge voltage curve, the coulombic efficiency, the energy efficiency, the internal resistance within the target duration of the discharge start, the rebound voltage within the target duration of the discharge end, the discharge start voltage, the curve characteristics of the electrical signal curve of the charge and discharge start segment, the curve characteristics of the electrical signal curve of at least one plateau segment of the charge and discharge, the minimum value of the charging capacity difference curve, the minimum value of the discharging capacity difference curve, the voltage position corresponding to the minimum value of the charging capacity difference curve, and the voltage position corresponding to the minimum value of the discharging capacity difference curve.

[0015] In some embodiments, the step of using a lifetime prediction model to predict health status decline information based on the electrical signal characteristics includes:

[0016] The electrical signal features are input into the lifetime prediction model, which learns attenuation features based on the electrical signal features and predicts the health status attenuation information based on the attenuation features.

[0017] In the above technical solution, by learning high-dimensional semantic features that can reflect attenuation characteristics through model learning, it is possible to learn the complex nonlinear mapping relationship between various electrical signal features, thereby improving the accuracy of prediction results.

[0018] In some embodiments, after obtaining health status decline information by using a lifetime prediction model based on the electrical signal characteristics, the method further includes:

[0019] The health status decay information is fitted to obtain the health status decay curve corresponding to the battery cell under test.

[0020] In the above technical solution, by fitting multiple predicted health state degradation information to obtain a health state degradation curve, the life degradation of the battery cell under test can be displayed more intuitively, improving the accuracy of the prediction results; and it is also convenient for users to customize the life of the battery cell under test according to the health state degradation curve, improving the user experience.

[0021] In some embodiments, before using a lifetime prediction model to predict based on the electrical signal characteristics to obtain health status decline information, the method further includes:

[0022] Obtain multiple training datasets;

[0023] With the goal of outputting health status decline information, sub-models corresponding to the training dataset are trained in the lifespan prediction model based on the training dataset.

[0024] In some embodiments, training sub-models corresponding to the training dataset in the lifespan prediction model based on the training dataset, with the goal of outputting health status decline information, includes:

[0025] The training dataset is input into the sub-model to obtain the predicted value output by the sub-model;

[0026] The target number of predicted values ​​with the largest offsets are removed from each of the predicted values, and the sub-model is optimized based on the average of the remaining predicted values.

[0027] In some embodiments, obtaining multiple sets of training datasets includes:

[0028] Obtain sample charge-discharge curve data of sample batteries during the initial use phase, and at least one sample health status corresponding to the sample charge-discharge curve data;

[0029] The sample charge-discharge curve data of the same sample battery, and the health status of a sample corresponding to the sample charge-discharge curve data, are determined as a training sample, and multiple training samples are obtained.

[0030] A first number of training samples are randomly extracted from the plurality of training samples to form a training dataset.

[0031] In the above technical solution, the corresponding sub-models are trained separately using different training datasets, and the sub-models are selected based on the prediction results of each sub-model. The prediction results of the retained sub-models are then translated into lifetime prediction results, which can alleviate the overfitting effect, reduce the impact of outliers on the overall effect, and improve the overall prediction effect of the lifetime prediction model.

[0032] Secondly, embodiments of this application provide a battery health status prediction device.

[0033] Thirdly, embodiments of this application provide a single battery cell;

[0034] Fourthly, embodiments of this application provide a non-transitory computer-readable storage medium;

[0035] Fifthly, embodiments of this application provide a computer program product. Attached Figure Description

[0036] Figure 1 This is one of the flowcharts illustrating a battery health status prediction method provided in some embodiments of this application;

[0037] Figure 2 This is a second schematic flowchart illustrating a battery health status prediction method provided in some embodiments of this application.

[0038] Figure 3 This is the third flowchart illustrating a battery health status prediction method provided in some embodiments of this application.

[0039] Figure 4 This is one of the schematic diagrams illustrating intermediate results of the battery health status prediction method provided in some embodiments of this application;

[0040] Figure 5 This is a second schematic diagram illustrating intermediate results of a battery health status prediction method provided in some embodiments of this application.

[0041] Figure 6 A schematic diagram illustrating the results of a battery health status prediction method provided in some embodiments of this application;

[0042] Figure 7 Fourth of a flowchart illustrating a battery health status prediction method provided in some embodiments of this application;

[0043] Figure 8 The third schematic diagram illustrates the intermediate results of the battery health status prediction method provided in some embodiments of this application;

[0044] Figure 9 Fourth of four schematic diagrams illustrating intermediate results of the battery health status prediction method provided in some embodiments of this application;

[0045] Figure 10 A schematic diagram of the structure of a battery health status prediction device provided in some embodiments of this application;

[0046] Figure 11 The diagram shows the structure of an electronic device provided in some embodiments of this application. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0048] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used in the description of this application is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms "comprising" and "having," and any variations thereof, in the description, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the description, claims, or accompanying drawings of this application are used to distinguish different objects, not to describe a specific order or hierarchy.

[0049] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0050] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "attachment" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0051] In this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, in this application, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0052] In this application, "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0053] The following description, in conjunction with the accompanying drawings, details the battery health state prediction method, battery health state prediction device, electronic device, and readable storage medium provided in this application through specific embodiments and application scenarios.

[0054] Among them, the battery health status prediction method can be applied to the terminal, specifically executed by the hardware or software in the terminal.

[0055] The battery health status prediction method provided in this application embodiment can be executed by an electronic device or a functional module or entity in an electronic device that can implement the battery health status prediction method. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablet computers, computers, etc. The following uses an electronic device as the execution subject to describe the battery health status prediction method provided in this application embodiment.

[0056] In related technologies, battery life prediction typically involves using data such as battery usage time, terminal voltage, output current, load voltage and current, and battery temperature. A temporal convolutional neural network model with structures like extended causal convolution and residual blocks is employed to establish a mapping relationship between the input sequence and the output sequence of the remaining lifespan of the lithium-ion battery. However, this method relies on a large amount of time-series data after the battery begins cycling, making it unsuitable for scenarios requiring rapid lifespan prediction (such as battery factory quality inspection and pre-installation evaluation of new batteries), and its prediction accuracy is not high.

[0057] Based on the above considerations, in order to address the problems of high dependence on time-series data in lifetime prediction, insufficient applicability in scenarios requiring rapid lifetime prediction, and low prediction accuracy, a battery health state prediction method is designed, including: acquiring charge-discharge curve data of the battery cell under test during the initial use stage; extracting electrical signal features corresponding to the charge-discharge curve data based on the charge-discharge curve data, including at least one of the peak value features of the first peak of the charge-discharge curve, the features of the voltage curve of the first charging plateau segment, and the features of the energy curve of the initial charging and discharging segment; and using a lifetime prediction model to predict based on the electrical signal features to obtain health state decay information.

[0058] This battery health status prediction method uses electrical signal features extracted from the charge-discharge curve data of the battery cell under test during the initial use stage to predict its lifetime. It introduces intrinsic features that are highly correlated with the aging mechanism and combines them with electrochemical mechanism guidance. This can improve the accuracy, realism and precision of the prediction results, and reduce the dependence on time-series data that are strongly correlated with the number of cycles, such as cycle time. It is suitable for a variety of application scenarios and has high universality.

[0059] like Figure 1 As shown, the battery health status prediction method includes steps 110, 120 and 130.

[0060] Step 110: Obtain the charge-discharge curve data of the battery cell under test during the initial use phase;

[0061] In this step, the battery cell can be a secondary battery, which refers to a battery cell that can be recharged to activate the active materials and continue to be used after the battery cell has been discharged.

[0062] The battery cell can be a lithium-ion battery, sodium-ion battery, sodium-lithium-ion battery, lithium metal battery, sodium metal battery, lithium-sulfur battery, magnesium-ion battery, nickel-metal hydride battery, nickel-cadmium battery, lead-acid battery, etc., and the embodiments of this application are not limited to this.

[0063] The battery cell can be cylindrical, flat, cuboid or other shapes, and the embodiments of this application are not limited to this.

[0064] The initial use phase refers to the initial stage of use of the battery cell under test, which may include the first charge-discharge cycle period or the period of the first preset number of charge-discharge cycles, such as the charge-discharge curve data within the first 10 or 50 charge-discharge cycles. This application does not limit this.

[0065] The horizontal axis of the charge-discharge curve represents time, and the vertical axis represents the voltage, current, and energy values ​​corresponding to each time value. The charge-discharge curve is a curve that can characterize the change of electrical signal over time in the initial charge-discharge stage of the battery cell under test.

[0066] Charge-discharge curve data can include: charging curve data and / or discharging curve data. Charge-discharge curve data represents the current, voltage, and energy values ​​at multiple time points during the charging and discharging process of the battery cell under test at the target ambient temperature. Energy is the product of current and voltage. Figure 4 As shown.

[0067] The target ambient temperature can be the ambient temperature under normal use conditions of the battery cell under test, such as 45°C or 60°C; in some embodiments, the target ambient temperature can also be adaptively set according to user needs, which is not limited here.

[0068] like Figure 2 As shown, in some embodiments, acquiring charge-discharge curve data of the battery cell under test during the initial use phase may include:

[0069] Receive the user's first input;

[0070] In response to the first input, charge / discharge curve data is acquired.

[0071] In this embodiment, the first input is used to input the charge-discharge curve data of the battery cell under test during the initial use phase.

[0072] The first input can be in at least one of the following ways:

[0073] Firstly, the first input can be a touch operation, including but not limited to click, swipe, and press operations.

[0074] In this embodiment, receiving the user's first input can be receiving the user's touch operation on the display area of ​​the terminal screen.

[0075] For example, when displaying a health status prediction interface, a target control can be displayed on the current interface. Touching the target control will enable the first input; or the first input can be set to multiple consecutive taps on the display area within a target time interval.

[0076] For example, users can upload files containing charge / discharge curve data, such as txt or excel files, by clicking the target control displayed on the health status prediction interface.

[0077] Secondly, the first input can be a physical button input.

[0078] In this embodiment, receiving the user's first input can be receiving the user's operation of pressing a corresponding physical button; the first input can also be a combination operation of pressing multiple physical buttons simultaneously.

[0079] Of course, in other embodiments, the first input may also be in other forms, including but not limited to character input, etc., which can be determined according to actual needs, and this application embodiment does not limit it.

[0080] In some embodiments, charge-discharge curve data can also be extracted from the historical operating parameters of the battery cell under test stored in the present application, which is not limited herein.

[0081] Step 120: Based on the charge-discharge curve data, extract the electrical signal features corresponding to the charge-discharge curve data;

[0082] In this step, electrical signal features are used to characterize the curve features of the charge-discharge curve data, which contain intrinsic features that are highly correlated with the aging mechanism of individual battery cells.

[0083] In some embodiments, the electrical signal characteristics may include at least one of the following: peak characteristics of the capacity versus voltage rate of change curve, target characteristics of the electrical signal curve of the charge / discharge plateau segment, and target characteristics of the electrical signal curve of the charge / discharge initiation segment.

[0084] In this embodiment, the rate of change of capacity with respect to voltage is dQ / dV, where Q represents the capacity of a single battery cell and V represents the voltage of the battery. By analyzing the curve of voltage (V) versus capacity (Q), the "inflection point" on the smooth slope can be converted into an easily identifiable peak.

[0085] Peak characteristics include, but are not limited to, the position, amplitude, width, area, and symmetry of the curve's peaks.

[0086] In some embodiments, the target feature includes at least one of sample entropy and kurtosis, where kurtosis is used to characterize the flatness of the curve.

[0087] Sample entropy is a feature value used to measure the complexity of a signal sequence; the more regular the sequence, the smaller the entropy. Kurtosis is a feature value used to measure the distribution of a signal sequence, describing the flatness of the signal; the flatter the signal, the smaller the kurtosis value.

[0088] In some embodiments, the peak characteristics of the rate of change of capacity with respect to voltage may include the peak characteristics corresponding to the first peak of the rate of change of capacity with respect to charge and discharge voltage.

[0089] In some embodiments, the peak characteristics of the rate of change of capacity versus voltage curve may include the peak height (DCdQdV_peakValues1) of the first peak of the rate of change of capacity versus charging voltage curve.

[0090] like Figure 9 As shown, based on the charge-discharge curves of multiple battery cells, the peak height and position of the first peak of the charge-discharge curve are approximately 3.15V. This first peak is the positive electrode peak and is related to positive electrode lithium loss and material loss. With the decrease in active lithium content, there is a trend of the peak height of the first peak of the charge-discharge curve gradually decreasing. This indicates a strong correlation between the peak height of the first peak of the charge-discharge curve and the actual active lithium content in the initial state of the battery cell.

[0091] In some embodiments, the electrical signal curve of the charge / discharge plateau segment may include: the voltage curve of the charge / discharge plateau segment and / or the energy curve of the charge / discharge plateau segment.

[0092] In some embodiments, the charge / discharge initiation segment electrical signal curve may include: the charge / discharge initiation segment voltage curve and / or the charge / discharge initiation segment energy curve.

[0093] In this embodiment, the initial segment is the initial stage of the charge-discharge curve. Taking the charging curve as an example, the initial segment is generally set near the 0%-10% battery state of charge (SOC). Within the initial segment, the voltage shows a steep upward trend during charging and a sharp downward trend during discharging. The electrical signal characteristics of the initial segment are greatly affected by ohmic polarization and activation polarization.

[0094] The plateau segment is the main reaction region following the initial stage, within which the charge and discharge voltage remains relatively stable. The number of plateau segments can be determined based on the reaction mechanism of the electrode material.

[0095] The electrical signal curve of the charge / discharge plateau segment is used to characterize the changes in electrical signals of one or more plateau segments during the charge / discharge process, including but not limited to changes in voltage, energy, and SOC within the corresponding plateau segment.

[0096] In some embodiments, the target features of the electrical signal curve of the charging and discharging plateau segment may include: target features of the electrical signal curve of the first charging plateau segment, including but not limited to one or more of the sample entropy (CC_V2_entropy) of the voltage curve of the first charging plateau segment and the kurtosis (CCEn1_kurtosis) of the energy curve of the first charging plateau segment.

[0097] Among them, the first charging platform segment is the first negative electrode platform. The sample entropy of the voltage curve of the first charging platform segment reflects the change of the negative electrode material loss in the time domain with the complexity of aging, that is, the voltage curve of the first charging platform segment is highly correlated with the negative electrode.

[0098] In an ideal and perfect battery cell, the energy of its charging plateau segment should be relatively stable, that is, its voltage and current should be relatively stable. Correspondingly, the flatness of the energy curve of the first charging plateau segment is higher. Through multiple experiments and verifications by the inventors, battery cells with a higher flatness of the energy curve of the first charging plateau segment have a relatively longer lifespan.

[0099] In some embodiments, the target feature of the charge / discharge initiation energy curve may include the sample entropy (DCEn4_entropy) of the discharge initiation energy curve. It is understood that the discharge initiation energy curve is highly correlated with the positive electrode. Extensive experimental data shows that a lower sample entropy in the discharge initiation energy curve corresponds to a battery cell whose positive electrode is closer to the ideal positive electrode.

[0100] In some embodiments, the target characteristic of the charge / discharge initiation voltage curve may include the kurtosis (CC_V1_kurtosis) of the charge initiation voltage curve. The smaller the kurtosis of the charge initiation voltage curve, the flatter the charge initiation voltage curve, and the longer the lifespan of the corresponding battery cell.

[0101] It should be noted that the peak characteristics of the first peak of the charge-discharge curve, the target characteristics of the electrical signal curve of the first plateau segment of charge-discharge, and the target characteristics of the electrical signal curve of the charge-discharge initiation segment are directly related to the internal electrochemical processes of the battery. They can directly reflect the key aging mechanisms such as SEI film growth and lithium dendrite formation inside the battery. By combining the guidance of electrochemical mechanisms and introducing intrinsic features that are strongly correlated with lifespan for lifespan prediction, the prediction error under complex operating conditions (such as extreme temperatures and dynamic charge-discharge) can be reduced, and the accuracy, authenticity, and reliability of the prediction results can be improved.

[0102] According to the battery health status prediction method provided in the embodiments of this application, aging prediction is performed by selecting electrical signal characteristics directly related to the internal electrochemical process of the battery, such as the peak value of the first peak of the charge-discharge curve, the sample entropy of the voltage curve of the first charging plateau segment, the kurtosis of the energy curve of the first charging plateau segment, the sample entropy of the energy curve of the discharge initiation segment, and the kurtosis of the voltage curve of the charging initiation segment. This method can be combined with electrochemical mechanism guidance to improve the accuracy, authenticity, and reliability of the prediction results.

[0103] In some embodiments, the electrical signal characteristics may further include at least one of the following: peak-valley characteristics of the capacity versus voltage rate curve, potential characteristics of the charge-discharge curve, difference between charging voltage and discharging voltage, voltage seasonality of the charging curve, trend change of the charge-discharge voltage curve, coulombic efficiency, energy efficiency, internal resistance within the target duration of discharge initiation, rebound voltage within the target duration of discharge termination, charge-discharge initiation voltage, curve characteristics of the charge-discharge initiation segment electrical signal curve, curve characteristics of the charge-discharge termination segment electrical signal curve, curve characteristics of the charge-discharge at least one plateau segment electrical signal curve, minimum value of the charge capacity difference curve, minimum value of the discharge capacity difference curve, voltage position corresponding to the minimum value of the charge capacity difference curve, and voltage position corresponding to the minimum value of the discharge capacity difference curve.

[0104] In this embodiment, the peak-valley characteristics of the capacity-voltage rate of change curve may include the peak-valley characteristics of the capacity-charging voltage rate of change curve (charging dQdV curve) and / or the peak-valley characteristics of the capacity-discharging voltage rate of change curve (discharging dQdV curve). The peak-valley characteristics include: peak characteristics and / or trough characteristics. These characteristics are used to characterize the position, peak height, peak width, and peak area of ​​the peaks or troughs.

[0105] like Figure 7 As shown, in some embodiments, the peak-valley characteristics of the capacity versus voltage rate curve may include: the peak position, peak height, peak width, and peak area of ​​the first peak of the charging dQdV curve; the peak position, peak height, peak width, and peak area of ​​the second peak; and so on.

[0106] The first peak of the charging dQdV curve is located at 3.31 ± 0.01 V. This first peak is the positive electrode peak and is associated with lithium loss and material loss in the positive electrode. The EOL curve shows that the position of the first peak tends to shift to the right as the lifetime decays.

[0107] The EOL curve represents the gradual degradation of battery performance over time as it is used, until it reaches the end of its life.

[0108] The peak height and position of the first peak in the charging dQdV curve are around 3.3V. This first peak is the positive electrode peak and is related to lithium loss and material loss in the positive electrode. The EOL curve shows that the first peak value tends to decrease as the lifespan degrades.

[0109] The peak width and position of the first peak in the charging dQdV curve are around 3.3V. The first peak is the positive electrode peak, which is related to the loss of lithium and material loss in the positive electrode.

[0110] The peak area of ​​the first peak in the charging dQdV curve is around 3.3V. This first peak is the positive electrode peak and is related to lithium loss and material loss in the positive electrode. The EOL curve shows that the area of ​​the first peak tends to decrease as the lifespan decreases.

[0111] The second peak of the charging dQdV curve is located near 3.4V, and this second peak is identified as the negative electrode peak, which is related to lithium loss and material loss in the negative electrode. The EOL curve shows that the position of the second peak also tends to shift to the right as the lifetime decays.

[0112] The peak height and position of the second peak in the charging dQdV curve are around 3.4V. This second peak is identified as the negative electrode peak and is related to lithium loss and material loss in the negative electrode. The EOL curve shows that the second peak also tends to decrease with the degradation of the battery life.

[0113] The peak width of the second peak in the charging dQdV curve indicates that this second peak is the negative electrode peak and is associated with negative electrode lithium loss and material loss. The EOL curve shows that the distance of the second peak tends to decrease with lifetime degradation.

[0114] The peak area of ​​the first peak in the charging dQdV curve is around 3.4V. The second peak is identified as the negative electrode peak, which is related to lithium loss and material loss in the negative electrode. The EOL curve shows that the area of ​​the second peak also tends to decrease with the degradation of the battery life.

[0115] In some embodiments, the peak-valley characteristics of the capacity versus voltage rate curve may include: the peak position, peak height, peak width, and peak area of ​​the first peak of the discharge dQdV curve; the peak position, peak height, peak width, and peak area of ​​the second peak; and so on.

[0116] The first peak of the discharge dQdV curve is located at 3.14 ± 0.01 V. This first peak is the positive electrode peak and is related to lithium loss and material loss in the positive electrode. Compared with the BOL curve, the first peak of the EOL curve shows a slight rightward shift, but the change is minimal.

[0117] Among them, the BOL curve represents the battery's performance at the beginning of its lifespan, and is used to characterize the battery's performance benchmark in a brand-new state.

[0118] The peak height and position of the first peak in the discharge dQdV curve are around 3.15V. This first peak is the positive electrode peak and is related to lithium loss and material loss in the positive electrode. The EOL curve shows that the first peak value has a slight decreasing trend with the decay of the lifetime.

[0119] The peak width and position of the first peak in the discharge dQdV curve are approximately 3.3V. This first peak is the positive electrode peak and is related to lithium loss and material loss in the positive electrode.

[0120] The peak area of ​​the first peak in the discharge dQdV curve is around 3.3V. This first peak is the positive electrode peak and is related to lithium loss and material loss in the positive electrode. The EOL curve shows that the area of ​​the first peak tends to decrease with the decay of the lifetime.

[0121] The second peak of the discharge dQdV curve is located near 3.4V, and is identified as the negative electrode peak, which is related to lithium loss and material loss in the negative electrode. The EOL curve shows that the position of the second peak also tends to shift to the left as the lifetime decays.

[0122] The peak width of the second peak in the discharge dQdV curve indicates that this second peak is the negative electrode peak, which is associated with lithium loss and material loss in the negative electrode. The EOL curve shows that the distance of the second peak tends to shift to the left and decrease as the lifetime decays.

[0123] The peak area of ​​the second peak in the discharge dQdV curve, with a peak position around 3.4V, indicates that the second peak is the negative electrode peak, which is related to lithium loss and material loss in the negative electrode. The EOL curve shows that the area of ​​the second peak also tends to decrease with the decay of the lifetime.

[0124] In some embodiments, the peak-valley characteristics of the capacity versus voltage rate curve may include: the peak-valley position of the first valley of the charging dQdV curve, the peak-valley height of the first valley, the peak-valley width of the first valley, the peak-valley area of ​​the first valley, the peak-valley position of the second valley, the peak-valley height of the second valley, the peak-valley width of the second valley, the peak-valley area of ​​the second valley, etc.

[0125] The first valley of the charging dVdQ curve, located around 20-30 Ah, is associated with the typical phase transition behavior during lithium insertion / extraction. The EOL curve shows that the first valley tends to increase with lifetime decay. The area of ​​the first peak-valley of the charging dVdQ curve, also located around 20-30 Ah, is associated with the typical phase transition behavior during lithium insertion / extraction.

[0126] The first peak and valley value of the charging dVdQ curve, with the peak position around 45Ah~55Ah, is associated with the general phase transition behavior of the lithium insertion / extraction process.

[0127] The second peak-valley area of ​​the charging dVdQ curve, with the peak position around 45Ah~55Ah, is associated with the general phase transition behavior during the lithium insertion / extraction process.

[0128] The second valley of the charging dVdQ curve, located at approximately 95Ah~120Ah, is associated with the general phase transition behavior during the lithium insertion / extraction process.

[0129] The third peak-valley area of ​​the charging dVdQ curve, with the peak position around 45Ah~55Ah, is associated with the general phase transition behavior during the lithium insertion / extraction process.

[0130] In some embodiments, the peak-valley characteristics of the capacity versus voltage rate curve may include: the peak-valley position of the first valley of the discharge dQdV curve, the peak-valley height of the first valley, the peak-valley width of the first valley, the peak-valley area of ​​the first valley, the peak-valley position of the second valley, the peak-valley height of the second valley, the peak-valley width of the second valley, the peak-valley area of ​​the second valley, etc.

[0131] The first valley of the discharge dVdQ curve, located around 200-260 Ah, is associated with the typical phase transition behavior during lithium insertion / extraction. The EOL curve shows that the first valley tends to increase with lifetime decay.

[0132] The area of ​​the first peak and valley of the discharge dVdQ curve, with the valley position around 200Ah~260Ah, is related to the general phase transition behavior of the lithium insertion / extraction process.

[0133] The first peak and valley value of the discharge dVdQ curve, with the peak position around 45Ah~55Ah, is associated with the general phase transition behavior of the lithium insertion / extraction process.

[0134] The second peak-valley area of ​​the discharge dVdQ curve, with the peak position around 45Ah~55Ah, is associated with the general phase transition behavior during the lithium insertion / extraction process.

[0135] The second valley of the discharge dVdQ curve, located at approximately 95Ah~120Ah, is associated with the general phase transition behavior during the lithium insertion / extraction process.

[0136] The third peak-valley area of ​​the discharge dVdQ curve, with a peak position of approximately 45Ah~55Ah, is associated with the general phase transition behavior during the lithium insertion / extraction process.

[0137] Potential characteristics can include median potential and average potential. The potential characteristics of a charge-discharge curve can include at least one of the median potential of the charging curve, the average potential of the charging curve, the median potential of the discharging curve, and the average potential of the discharging curve.

[0138] The median potential of the charging curve reflects the change in the voltage plateau of the charging curve and is related to lithium loss and material loss. Similarly, the median potential of the discharging curve reflects the change in the voltage plateau of the discharging curve and is related to lithium loss and material loss.

[0139] The average potential of the charging curve reflects the change in the cumulative polarization of the early stage of the charging curve and is related to the increase in cell polarization; the average potential of the discharging curve reflects the change in the cumulative polarization of the early stage of the discharging curve and is related to the increase in cell polarization.

[0140] The difference between charging voltage and discharging voltage may include, but is not limited to: the difference between charging voltage and discharging voltage at 50% SOC, and the voltage difference between average charging voltage and average discharging voltage.

[0141] The seasonal variation of voltage in the charging curve and the trend of voltage variation in the charging and discharging curves can be obtained by seasonal-trend decomposition using LOESS (STL) based on the charging and discharging curve data during the initial use phase. In this method, a complete time series (i.e., charging and discharging curve data) can be separated into three parts: trend variation, seasonal variation, and residual term.

[0142] Trend changes correspond to smooth changes over a long-term dimension, stripping away short-term fluctuations and residual terms, and can express fundamental patterns of change, such as the trend of voltage changing over time during charging, which is always upward.

[0143] Seasonal changes correspond to recurring fluctuation patterns. The amount of this periodic change is usually related to the nature of the equipment and the battery cell itself; the value of this item is generally lower than that of trend changes.

[0144] In some embodiments, the voltage seasonality of the charging curve may include the voltage seasonality of the charging curve, the time-series variation characteristics of the voltage (charging S-area), and the voltage seasonality of the discharging curve, the time-series variation characteristics of the voltage (discharging S-area).

[0145] In some embodiments, the trend change of the charge and discharge voltage curves may include: the trend change of the charging voltage curve (charging T-mean) and / or the trend change of the discharging voltage curve (discharging T-mean).

[0146] Figure 8 This example demonstrates an STL decomposition. Figure 8 The data, from top to bottom, consists of "raw data", "trend change", "seasonal change" and "residual term". It can be seen that although the trend change may be a monotonic function, its monotonicity disappears after adding the seasonal change term and the residual term.

[0147] Coulomb efficiency (CE) is related to the difference in capacity between charge and discharge.

[0148] Energy efficiency (RTE) is related to the polarization difference between charge and discharge.

[0149] The target duration can be customized by the user, such as: the duration of 1 second at the start of discharge (100% SOC), the duration of 5 seconds at the start of discharge (100% SOC), and the duration of 30 seconds at the start of discharge (100% SOC), as well as the duration of 1 second, 5 seconds, and 30 seconds at the end of discharge.

[0150] The charge / discharge initiation voltage may include: the charge initiation voltage (initCCVol) and / or the discharge initiation voltage (ΔU_mean).

[0151] Electrical signal curves can include voltage curves and energy curves.

[0152] Curve characteristics include, but are not limited to: kurtosis, skewness, slope, range, and sample entropy.

[0153] Skewness is a characteristic value used to measure the distribution of a signal sequence and to describe the symmetry of the signal. The more symmetrical the signal, the smaller the skewness value.

[0154] The slope is used to characterize how fast the signal frequency changes with the state, reflecting how fast the internal reaction of the battery changes at the first charging plateau.

[0155] Range is a characteristic value used to measure the range of variation of a signal sequence.

[0156] The curve characteristics of the electrical signal curve at the beginning of charging and discharging may include, but are not limited to: the curve characteristics of the voltage curve at the beginning of charging, the curve characteristics of the energy curve at the beginning of charging, the curve characteristics of the voltage curve at the beginning of discharging, and the curve characteristics of the energy curve at the beginning of discharging.

[0157] In some embodiments, the curve characteristics of the electrical signal curve at the start of charging and discharging may include: kurtosis of the voltage curve at the start of charging, skewness of the voltage curve at the start of charging, slope of the voltage curve at the start of charging, sample entropy of the voltage curve at the start of discharging, kurtosis of the voltage curve at the start of discharging, skewness of the voltage curve at the start of discharging, and slope of the voltage curve at the start of discharging.

[0158] The curve characteristics of the electrical signal curve at the end of the charge and discharge phases may include, but are not limited to: the curve characteristics of the voltage curve at the end of the charge phase, the curve characteristics of the energy curve at the end of the charge phase, the curve characteristics of the voltage curve at the end of the discharge phase, and the curve characteristics of the energy curve at the end of the discharge phase.

[0159] In some embodiments, the curve characteristics of the voltage curve at the end of charging include, but are not limited to: the sample entropy of the voltage curve at the end of charging, the kurtosis of the voltage curve at the end of charging, the skewness of the voltage curve at the end of charging, and the slope of the voltage curve at the end of charging.

[0160] The voltage curve at the end of the charging process is closely related to liquid-phase and solid-phase diffusion. The entropy energy of this voltage curve changes with the complexity of aging in the time domain due to the change in liquid-phase and solid-phase diffusion reactions.

[0161] In some embodiments, the curve characteristics of the discharge terminal voltage curve may include: the sample entropy of the discharge terminal voltage curve, the kurtosis of the discharge terminal voltage curve, the skewness of the discharge terminal voltage curve, and the slope of the discharge terminal voltage curve, etc.

[0162] At least one platform segment may include, but is not limited to, a first platform segment and a second platform segment.

[0163] In some embodiments, the curve characteristics of the electrical signal curve of at least one plateau segment during charging and discharging may include: the sample entropy of the voltage curve of at least one plateau segment during charging, the kurtosis of the voltage curve of at least one plateau segment during charging, the skewness of the voltage curve of at least one plateau segment during charging, and the slope of the voltage curve of at least one plateau segment during charging.

[0164] In some embodiments, the curve characteristics of the electrical signal curves of at least one plateau segment during charging and discharging may include: the sample entropy of the energy curve of the first plateau segment during charging, the kurtosis of the energy curve of the first plateau segment during charging, the range of the energy curve of the first plateau segment during charging, the skewness of the energy curve of the first plateau segment during charging, the slope of the energy curve of the first plateau segment during charging, the sample entropy of the energy curve of the second plateau segment during charging, the kurtosis of the energy curve of the second plateau segment during charging, the range of the energy curve of the second plateau segment during charging, the skewness of the energy curve of the second plateau segment during charging, and the slope of the energy curve of the second plateau segment during charging. In some embodiments, the curve characteristics of the electrical signal curves of at least one plateau segment during charging and discharging may include: the sample entropy of the energy curve of at least one plateau segment during discharging, the kurtosis of the energy curve of at least one plateau segment during discharging, the range of the energy curve of at least one plateau segment during discharging, the skewness of the energy curve of at least one plateau segment during discharging, and the slope of the energy curve of at least one plateau segment during discharging.

[0165] In some embodiments, the curve characteristics of the electrical signal curve of at least one plateau segment during charging and discharging may include: the sample entropy of the voltage curve of the first plateau segment during discharge, the kurtosis of the voltage curve of the first plateau segment during discharge, the range of the voltage curve of the first plateau segment during discharge, the skewness of the voltage curve of the first plateau segment during discharge, the slope of the voltage curve of the first plateau segment during discharge, the sample entropy of the voltage curve of the second plateau segment during discharge, the kurtosis of the voltage curve of the second plateau segment during discharge, the range of the voltage curve of the second plateau segment during discharge, the skewness of the voltage curve of the second plateau segment during discharge, and the slope of the voltage curve of the second plateau segment during discharge, etc.

[0166] In some embodiments, the electrical signal characteristics may further include: the curve characteristics of the discharge terminal voltage curve, including but not limited to: the sample entropy of the discharge terminal voltage curve, the kurtosis of the discharge terminal voltage curve, the skewness of the discharge terminal voltage curve, and the slope of the discharge terminal voltage curve.

[0167] The minimum value of the discharge capacity difference curve is the EOL voltage-capacity curve minus the BOL voltage-capacity curve.

[0168] The minimum value of the charging capacity difference curve is the EOL voltage-capacity curve minus the BOL voltage-capacity curve.

[0169] Among them, the EOL voltage-capacity curve is the curve showing how battery performance gradually declines with use or time until it reaches the end of life; the BOL voltage-capacity curve is the performance of the battery at the beginning of life, and the BOL voltage-capacity curve is used to characterize the performance benchmark of the battery in a brand new state.

[0170] In actual implementation, any feasible method can be used to extract the corresponding electrical signal features. The specific category of electrical signal features selected can be based on user-defined criteria or selected according to requirements. This application does not limit this.

[0171] Step 130: Use a lifetime prediction model to predict based on electrical signal characteristics to obtain health status decline information.

[0172] In this step, the lifespan prediction model can be an artificial intelligence model, including but not limited to: neural network models, machine learning models, and deep learning models. The input features of the lifespan prediction model are electrical signal features extracted from charge-discharge curve data, and the output value is health status decay information.

[0173] The lifespan prediction model can be pre-trained. The specific training method will be described in the following examples and will not be elaborated here.

[0174] State of Health (SOH) information refers to the predicted changes in the state of health over time. SOH is the percentage of a battery cell's current capacity relative to its factory capacity, used to measure battery performance degradation and remaining lifespan, expressed as a percentage (%), with a value range of [0, 100].

[0175] Understandably, the health of a battery is affected by storage time decay and cycle decay.

[0176] Storage time decay refers to the process by which the health status of a battery cell declines only over time (e.g., days, months) in scenarios of "static storage" or "low-intensity cycling" when it is not used for charging or discharging, which will reduce the actual usable capacity of the battery cell.

[0177] The health status corresponding to storage time decay is the remaining capacity of a battery cell after the capacity decay caused by changes in storage time, that is, the remaining capacity affected by storage time decay.

[0178] Storage time decay corresponding to a healthy state is the percentage decrease in the battery's current remaining capacity compared to its nominal capacity when stored under specific conditions, such as a specific temperature or no current.

[0179] Cyclic degradation refers to the process by which the health of a battery cell gradually declines with the number of cycles due to irreversible chemical changes such as electrode material wear and electrolyte aging during repeated charge and discharge cycles. This results in a reduction in the usable capacity of the battery cell and a decrease in charge and discharge efficiency.

[0180] The health state corresponding to cycle degradation is the remaining capacity of a battery cell after the charge and discharge cycles have caused the energy to decrease.

[0181] Cycle degradation corresponding to a healthy state is the percentage decrease in the current remaining capacity of a battery cell compared to its nominal capacity after repeated charging and discharging under specific charging and discharging conditions, such as specific temperature, depth of charge and discharge, and charge and discharge rate.

[0182] In some embodiments, health status decay information may include the number of cycles required to reach a certain health status and the storage time, such as TTF_0.99, TTF_0.98, ..., TTF_0.90, where TTF represents the number of cycles or the storage time. TTF_0.97 represents the storage time / number of cycles required to reach a 97% health status. Figure 5 As shown.

[0183] Among related technologies, there is another method that uses features extracted during the first charge-discharge cycle of a lithium-ion battery, such as fast charging time, standard deviation of charging voltage, standard deviation of charging current, standard deviation of constant current discharge voltage, constant voltage discharge time, maximum capacity, and the ratio of maximum capacity to average fast charging current, to predict the remaining life of the battery using a model. However, the features used in this method do not directly reflect the internal aging mechanism of the battery, and the prediction accuracy is prone to fluctuations under complex operating conditions (such as extreme temperatures and dynamic charge-discharge), thus affecting the authenticity and accuracy of the prediction results.

[0184] In this application, by using electrical signal features extracted from the charge-discharge curve data of the battery cell under test during the initial use stage for lifetime prediction, the dependence on time series data that are strongly correlated with the number of cycles, such as cycle time, can be reduced. The remaining lifetime can be quickly assessed during the initial use stage of the battery (such as after the first charge-discharge cycle). This method is suitable for scenarios that require rapid lifetime prediction, such as lifetime prediction in scenarios such as battery factory quality inspection and new battery pre-installation evaluation.

[0185] In addition, lifetime prediction is performed by extracting electrical signal characteristics such as the peak characteristics of the capacity versus voltage rate curve, the characteristics of the voltage curve at the charging plateau, and the characteristics of the energy curve at the start of charging and discharging. These characteristic parameters are directly related to the internal electrochemical processes of the battery and can directly reflect key aging mechanisms such as SEI film growth and lithium dendrite formation. By combining electrochemical mechanism guidance and introducing intrinsic characteristics strongly correlated with lifetime for lifetime prediction, prediction errors under complex operating conditions (such as extreme temperatures and dynamic charging and discharging) can be reduced, and the accuracy, realism, and precision of the prediction results can be improved.

[0186] The battery health status prediction method provided in this application uses electrical signal features extracted from the charge-discharge curve data of the battery cell under test during the initial use stage to predict its lifetime. It introduces intrinsic features that are highly correlated with the aging mechanism and combines them with electrochemical mechanism guidance. This method can improve the accuracy, authenticity and precision of the prediction results, and can reduce the dependence on time-series data that are strongly correlated with the number of cycles, such as cycle time. It is applicable to a variety of application scenarios and has high universality.

[0187] In some embodiments, step 130 may include:

[0188] The electrical signal features are input into the lifetime prediction model, which learns the decay features based on the electrical signal features and predicts the decay information of the health status based on the decay features.

[0189] In this embodiment, the attenuation feature is a high-dimensional semantic feature that is correlated with the health status of the battery cell.

[0190] In some embodiments, the lifetime prediction model may include: a feature extraction module and a prediction module, wherein electrical signal features are input into the lifetime prediction model, the lifetime prediction model learns decay features based on the electrical signal features, and predicts health status decay information based on the decay features, and may include:

[0191] The electrical signal features are input into the feature extraction module to obtain the attenuation features output by the feature extraction module;

[0192] The decay characteristics are input into the prediction module to obtain the health status decay information output by the prediction module.

[0193] In this embodiment, the feature extraction module is used to learn electrical signal features to extract high-dimensional semantic features associated with the lifespan of a battery cell, thereby obtaining attenuation features; then, the prediction module predicts the health status attenuation information based on the extracted attenuation features.

[0194] In actual training, the feature extraction module and the prediction module can be trained separately or as a whole.

[0195] According to the battery health status prediction method provided in the embodiments of this application, the model learns high-dimensional semantic features that can reflect the attenuation characteristics, and can learn the complex nonlinear mapping relationship between various electrical signal features, thereby improving the accuracy of the prediction results.

[0196] In some embodiments, after step 130, the method further includes:

[0197] The health status degradation information is fitted to obtain the health status degradation curve of the battery cell under test.

[0198] In this embodiment, the health state curve is used to characterize the change in the health state of the tested battery cell over time. The horizontal axis of the health state curve represents time, and the vertical axis represents the health state, as shown below. Figure 6 As shown. In some embodiments, the health state curve may include at least one of a thermodynamic decay curve and a kinetic decay curve.

[0199] The health state decay information predicted through steps 110 to 130 can be multiple discrete data points. In actual execution, the health state decay information can also be fitted to obtain a smooth curve. In some embodiments, fitting the health state decay information may include interpolating each piece of health state decay information to obtain a health state curve. In some embodiments, the interpolation algorithm may include, but is not limited to, linear interpolation, polynomial interpolation, and cubic spline interpolation.

[0200] According to the battery health state prediction method provided in the embodiments of this application, by fitting multiple predicted health state decay information to obtain a health state decay curve, the life decay of the battery cell under test can be displayed more intuitively, improving the accuracy of the prediction results; and it is also convenient for users to customize the life of the battery cell under test according to the health state decay curve, improving the user experience.

[0201] like Figure 2 As shown, in some embodiments, after step 130, the method further includes: displaying health status decay information and a lifespan analysis report determined based on the health status decay information.

[0202] The display may include, but is not limited to, displaying on a web platform or displaying on an app.

[0203] The training method for the lifespan prediction model will be explained below.

[0204] like Figure 3 As shown, in some embodiments, before inputting the electrical signal features into the lifetime prediction model and obtaining the health status decay information output by the lifetime prediction model, the method further includes:

[0205] Obtain multiple training datasets;

[0206] With the goal of outputting health status decay information, sub-models corresponding to the training dataset are trained in the lifespan prediction model based on the training dataset.

[0207] In this embodiment, the lifespan prediction model may include multiple sub-models with the same architecture. The sub-models may be neural network models, machine learning models, or deep learning models, etc. The number of training datasets is the same as the number of sub-models, such as n training datasets, where n is a positive integer. Each training dataset is used to train one sub-model. Each training dataset may include multiple training samples.

[0208] In some embodiments, obtaining multiple sets of training datasets includes:

[0209] Obtain sample charge-discharge curve data of sample batteries during the initial use phase, and at least one sample health status corresponding to the sample charge-discharge curve data;

[0210] The sample charge-discharge curve data of the same sample battery, and the health status of a sample corresponding to the sample charge-discharge curve data, are determined as a training sample, and multiple training samples are obtained.

[0211] The first number of training samples are randomly extracted from multiple training samples to form a training dataset.

[0212] In this embodiment, the sample battery can be a number of different types of batteries.

[0213] The initial use phase refers to the initial stage of use of the sample battery, which may include the charge and discharge curve data within the time period of the first charge and discharge or the first preset number of charge and discharge cycles, such as the first 10 or 50 charge and discharge cycles.

[0214] The sample charge-discharge curve data refers to the actual charge-discharge curve data of the sample battery during the initial use phase, including but not limited to the current and voltage values ​​at multiple time points during charge-discharge processes at one or more ambient temperatures. The sample health status refers to the actual health status at multiple time points corresponding to the sample charge-discharge curve data.

[0215] The ambient temperature can be customized by the user, such as 20℃, 45℃, and 60℃.

[0216] The first quantity is less than the total number of training samples. The first quantity can be based on a user-defined setting, such as setting it to m, where m is a positive integer.

[0217] A training sample is constructed by taking the charge-discharge curve data of the same sample battery and the corresponding health status of the sample charge-discharge curve data, thus obtaining multiple training samples. A first set of a certain number of training samples is randomly extracted from these multiple training samples to form a training dataset, thereby obtaining multiple training datasets containing different training samples.

[0218] During training, the training dataset can be input into the corresponding sub-model, with the goal of obtaining health status decay information from the sub-model's output. The sub-model is then trained using a loss function. Similarly, each sub-model can be trained independently.

[0219] It is understandable that the training datasets for different sub-models may differ, and the training results may also vary. In some embodiments, the sub-model with the best training performance can be selected from among the multiple trained sub-models as the lifespan prediction model for subsequent applications.

[0220] In some embodiments, with the goal of outputting health status decay information, sub-models corresponding to the training dataset are trained in the lifespan prediction model based on the training dataset, including:

[0221] Input the training dataset into the sub-model and obtain the predicted values ​​output by the sub-model;

[0222] Remove the target number of predictions with the largest offset from each prediction, and optimize the sub-model based on the average of the remaining predictions.

[0223] In this embodiment, the target number is less than the total number of sub-models. The specific value can be user-defined and is not limited in this application.

[0224] The offset is used to characterize the difference between the prediction result of each sub-model and the average of all prediction results. The greater the difference, the greater the offset of the prediction result of that sub-model. By deleting some sub-models with large offsets and retaining the sub-models with smaller offsets, the average of the prediction results of the retained sub-models is used as the final result to optimize the sub-models.

[0225] According to the battery health status prediction method provided in the embodiments of this application, corresponding sub-models are trained separately using different training datasets. Sub-models are selected based on the prediction results of each sub-model, and the prediction results of the retained sub-models are translated into lifespan prediction results. This can alleviate the overfitting effect, reduce the impact of outliers on the overall effect, and improve the overall prediction effect of the lifespan prediction model.

[0226] The battery health status prediction method provided in this application can be executed by a battery health status prediction device. This application uses the example of a battery health status prediction device executing the battery health status prediction method to illustrate the battery health status prediction device provided in this application.

[0227] This application also provides a battery health status prediction device.

[0228] like Figure 10As shown, the battery health status prediction device includes: a first processing module 1010, a second processing module 1020 and a third processing module 1030.

[0229] The first processing module 1010 is used to acquire charge and discharge curve data of the battery cell under test during the initial use stage.

[0230] The second processing module 1020 is used to extract electrical signal features corresponding to the charge and discharge curve data based on the charge and discharge curve data. The electrical signal features include at least one of the peak features of the capacity versus voltage rate curve, the target features of the charge and discharge plateau segment electrical signal curve, and the target features of the charge and discharge start segment electrical signal curve.

[0231] The third processing module 1030 is used to use a life prediction model to predict based on electrical signal characteristics to obtain health status decay information;

[0232] The target features include at least one of sample entropy and kurtosis.

[0233] The battery health status prediction device provided in the embodiments of this application uses electrical signal features extracted from the charge-discharge curve data of the battery cell under test during the initial use stage to predict its lifespan. It introduces intrinsic features that are highly correlated with the aging mechanism and combines them with electrochemical mechanism guidance, which can improve the accuracy, authenticity and precision of the prediction results. It can also reduce the dependence on time-series data that are strongly correlated with the number of cycles, such as cycle time. It is suitable for a variety of application scenarios and has high universality.

[0234] In some embodiments, the third processing module 1030 is configured to:

[0235] The electrical signal features are input into the lifetime prediction model, which learns the decay features based on the electrical signal features and predicts the decay information of the health status based on the decay features.

[0236] In some embodiments, the device further includes a fourth processing module for:

[0237] After using a lifetime prediction model to predict based on electrical signal characteristics and obtaining health state degradation information, the health state degradation information is fitted to obtain the health state degradation curve corresponding to the battery cell under test.

[0238] In some embodiments, the device further includes a fifth processing module for:

[0239] Before using a lifespan prediction model based on electrical signal characteristics to predict and obtain information on health status decline, multiple sets of training datasets are obtained.

[0240] With the goal of outputting health status decay information, sub-models corresponding to the training dataset are trained in the lifespan prediction model based on the training dataset.

[0241] In some embodiments, the fifth processing module is configured to:

[0242] Input the training dataset into the sub-model and obtain the predicted values ​​output by the sub-model;

[0243] Remove the target number of predictions with the largest offset from each prediction, and optimize the sub-model based on the average of the remaining predictions.

[0244] In some embodiments, the fifth processing module is configured to:

[0245] Obtain charge-discharge curve data of the sample battery during the initial use phase, and at least one health state degradation value corresponding to the charge-discharge curve data;

[0246] The charge-discharge curve data of the same sample battery, and a health state decay value corresponding to the charge-discharge curve data, are determined as a training sample, and multiple training samples are obtained.

[0247] The first number of training samples are randomly extracted from multiple training samples to form a training dataset.

[0248] The battery health status prediction device in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the specific device.

[0249] The battery health status prediction device in this application embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit it.

[0250] The battery health status prediction device provided in this application embodiment can achieve... Figures 1 to 7 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0251] This application also provides a single battery cell. The battery cell performs health state prediction based on the battery health state prediction method described in any of the above embodiments.

[0252] Among them, the battery cell can be a secondary battery, which refers to a battery cell that can be recharged to activate the active materials and continue to be used after the battery cell has been discharged.

[0253] The battery cell can be a lithium-ion battery, sodium-ion battery, sodium-lithium-ion battery, lithium metal battery, sodium metal battery, lithium-sulfur battery, magnesium-ion battery, nickel-metal hydride battery, nickel-cadmium battery, lead-acid battery, etc., and the embodiments of this application are not limited to this.

[0254] Battery cells can be cylindrical, flat, cuboid, or other shapes, and this application embodiment is not limited to any of these. Battery cells are generally classified into three types according to their packaging method: cylindrical battery cells, square battery cells, and pouch battery cells, and this application embodiment is not limited to any of these types either.

[0255] A battery cell includes a casing, electrode components, and electrolyte. The casing houses the electrode components and electrolyte. The electrode components consist of a positive electrode, a negative electrode, and a separator. The battery cell primarily functions by the movement of metal ions between the positive and negative electrode components. The positive electrode includes a positive current collector and a positive active material layer. The positive current collector includes a current collector body and a positive electrode tab. The positive active material layer is coated on the surface of the current collector body, while the positive electrode tab is not coated with the positive active material layer and protrudes from the current collector body. Taking a lithium-ion battery as an example, the material of the positive current collector can be aluminum, and the positive active material can be lithium cobalt oxide, lithium iron phosphate, ternary lithium, or lithium manganese oxide, etc. The negative electrode includes a negative current collector and a negative active material layer. The negative current collector includes a current collector body and a negative electrode tab. The negative active material layer is coated on the surface of the current collector body, while the negative electrode tab is not coated with the negative active material layer and protrudes from the current collector body. The negative electrode current collector can be made of copper, and the negative electrode active material can be carbon or silicon, etc. To ensure that a large current can be passed without melting, there are multiple positive electrode tabs stacked together, and there are multiple negative electrode tabs stacked together.

[0256] The separator can be made of PP (polypropylene) or PE (polyethylene), etc. Furthermore, the electrode assembly can be a wound structure or a stacked structure; the embodiments of this application are not limited to these.

[0257] The technical solutions described in the embodiments of this application are applicable to various electrical devices that use individual battery cells, such as mobile phones, portable devices, laptops, electric vehicles, electric toys, power tools, vehicles, ships, and spacecraft, including aircraft, rockets, space shuttles, and spacecraft. Individual battery cells are used to store or provide electrical energy.

[0258] According to the battery cell provided in the embodiments of this application, by using the electrical signal features extracted from the charge-discharge curve data of the battery cell under test in the initial use stage for lifetime prediction, the intrinsic features with high correlation to aging mechanism are introduced. Combined with the guidance of electrochemical mechanism, the accuracy, authenticity and accuracy of prediction results can be improved, and the dependence on time series data such as cycle time which are strongly correlated with the number of cycles can be reduced. It is suitable for a variety of application scenarios and has high universality.

[0259] This application also provides a battery device. The battery device mentioned in the embodiments of this application may include one or more battery cell assemblies for providing voltage and capacity. The battery cell assembly may include multiple battery cells, which are connected in series, parallel, or mixed connection through a busbar.

[0260] In some embodiments, a battery cell assembly is typically formed by arranging multiple battery cells.

[0261] As an example, a battery cell assembly can be a battery module, which is formed by arranging and fixing multiple battery cells together to form an independent module. As another example, a battery module can be formed by bundling multiple battery cells together with cable ties.

[0262] In some embodiments, the battery device may be a battery pack, which includes a housing and one or more individual battery cells housed within the housing.

[0263] As an example, the battery cell assembly can be a battery module, which can be housed in a housing by fixing the battery module in the housing.

[0264] As an example, battery cell assemblies can also be housed in a housing by directly fixing multiple battery cells to the housing.

[0265] As an example, the enclosure may include a first enclosure and a second enclosure. The first enclosure and the second enclosure are fastened together to form a closed space inside the enclosure to house the individual battery cells. Here, "closed" refers to covering or closing, and can be either sealed or unsealed. The first enclosure may be a top cover or a bottom plate.

[0266] As an example, the enclosure may include a top cover, a frame, and a bottom plate. The top cover and bottom plate are connected to the frame, creating an enclosed space inside the enclosure to house the individual battery cells.

[0267] In some embodiments, the housing may be part of the vehicle's chassis structure. For example, a portion of the housing may be at least a part of the vehicle's floor, or a portion of the housing may be at least a part of the vehicle's crossbeams and longitudinal beams.

[0268] The technical solutions described in the embodiments of this application are applicable to various electrical devices that use battery devices, such as mobile phones, portable devices, laptops, electric vehicles, electric toys, power tools, vehicles, ships, and spacecraft, etc. For example, spacecraft include airplanes, rockets, space shuttles, and spacecraft. The battery device is used to store or provide electrical energy.

[0269] This application provides an energy storage device including one or more battery clusters to increase the voltage and capacity of the energy storage device. The battery cluster may include multiple battery devices, which are connected in series via a busbar to increase the voltage of the energy storage device. When the energy storage device includes multiple battery clusters, the multiple battery clusters are connected in parallel to increase the capacity of the energy storage device.

[0270] Energy storage devices can be used in energy storage power stations, wind power generation systems, solar power generation systems, mobile power systems, or temporary power supply systems. Energy storage devices can store electrical energy as needed and output it when appropriate. For example, an energy storage device can store electrical energy during off-peak hours and provide power to relevant users or electrical devices during peak hours. The energy storage system provided in this application embodiment can be any power system that requires energy storage devices.

[0271] In some embodiments, the energy storage device is an energy storage container or an energy storage cabinet.

[0272] In some embodiments, the energy storage device may include a cabinet and one or more battery clusters housed within the cabinet.

[0273] In some embodiments, the energy storage device may include modules such as a thermal management module, a main control module, a central control module, a power distribution module, and a fire protection module.

[0274] As an example, the thermal management module may include a liquid cooling unit that supplies coolant to each battery device via piping to regulate the temperature of the individual battery cells.

[0275] As an example, the main control module can serve as the battery management unit for the battery cluster, used to monitor and manage the battery cluster. The main control module can monitor information such as the current, voltage, power, or temperature of the battery cluster. For instance, it can control the charging and discharging current and voltage of the battery cluster. The main control module includes modules such as an auxiliary battery management unit (SBMU) and a fusion switch.

[0276] As an example, the central control module can serve as the battery management unit for an energy storage device, used to monitor and manage the device. The central control module can monitor information such as the energy storage device's current, voltage, power, state of charge, or temperature. For instance, it can control the charging and discharging current and voltage of the energy storage device. As an example, the central control module includes modules such as an Insulation Monitoring Module (IMM), a Master Battery Management Unit (MBMU), an Ethernet (ETH) module, and a fiber optic conversion module.

[0277] As an example, a fire protection system includes control panels, detectors, alarm devices, etc., used to detect, alarm, or extinguish fires in energy storage systems.

[0278] As an example, the power distribution unit can be used to distribute power to the power modules of the energy storage device.

[0279] The technical solutions described in the embodiments of this application are applicable to various electrical devices that use energy storage devices, such as mobile phones, portable devices, laptops, electric vehicles, electric toys, power tools, vehicles, ships, and spacecraft, etc. For example, spacecraft include airplanes, rockets, space shuttles, and spacecraft. The energy storage device is used to store or provide electrical energy.

[0280] This application provides an electrical device that uses a single battery cell, battery device, energy storage device, or energy storage system as a power source. The electrical device can be, but is not limited to, mobile phones, tablets, laptops, electric toys, power tools, electric vehicles, electric cars, ships, and spacecraft. Electric toys can include stationary or mobile electric toys, such as game consoles, electric car toys, electric ship toys, and electric airplane toys. Spacecraft can include airplanes, rockets, space shuttles, and spacecraft.

[0281] In some embodiments, such as Figure 11As shown, this application embodiment also provides an electronic device 1100, including a processor 1101, a memory 1102, and a computer program stored in the memory 1102 and executable on the processor 1101. When the program is executed by the processor 1101, it implements the various processes of the above-described battery health state prediction method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0282] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0283] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described battery health state prediction method embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0284] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0285] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described battery health state prediction method.

[0286] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0287] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described battery health status prediction method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0288] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0289] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0290] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0291] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0292] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0293] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for predicting the health status of a battery, characterized in that, include: Acquire charge-discharge curve data of the battery cell under test during the initial use phase; Based on the charge-discharge curve data, the electrical signal features corresponding to the charge-discharge curve data are extracted. The electrical signal features include the peak features of the capacity versus voltage rate curve, the target features of the charge-discharge plateau segment electrical signal curve, and the target features of the charge-discharge initiation segment electrical signal curve. Based on the characteristics of the electrical signal, a lifetime prediction model is used to predict and obtain information on the decline of health status; The target feature includes at least one of sample entropy and kurtosis; The initial use phase includes: the first charge-discharge cycle time period or the time period of the first preset number of charge-discharge cycles; The target features of the electrical signal curve of the charging and discharging plateau segment include one or more of the sample entropy of the voltage curve of the first charging plateau segment and the kurtosis of the energy curve of the first charging plateau segment. The target features of the electrical signal curve at the start of the charge / discharge phase include one or more of the sample entropy of the energy curve at the start of the discharge phase and the kurtosis of the voltage curve at the start of the charge phase; the first charging plateau phase is the negative electrode first plateau.

2. The battery health status prediction method according to claim 1, characterized in that, The electrical signal characteristics also include: The peak-valley characteristics of the capacity versus voltage rate curve, the potential characteristics of the charge-discharge curve, the difference between the charging voltage and the discharging voltage, the voltage seasonality of the charging curve, the trend of the charge-discharge voltage curve, the coulombic efficiency, the energy efficiency, the internal resistance within the target duration of the discharge start, the rebound voltage within the target duration of the discharge end, the charge-discharge start voltage, the curve characteristics of the electrical signal curve at the end of the charge-discharge phase, the minimum value of the charge capacity difference curve, the minimum value of the discharge capacity difference curve, the voltage position corresponding to the minimum value of the charge capacity difference curve, and the voltage position corresponding to the minimum value of the discharge capacity difference curve.

3. The battery health status prediction method according to claim 1 or 2, characterized in that, The method of using a lifetime prediction model to predict health status decline based on the electrical signal characteristics includes: The electrical signal features are input into the lifetime prediction model, which learns attenuation features based on the electrical signal features and predicts the health status attenuation information based on the attenuation features.

4. The battery health status prediction method according to claim 1 or 2, characterized in that, After obtaining health status decline information by using a lifetime prediction model based on the electrical signal characteristics, the method further includes: The health status decay information is fitted to obtain the health status decay curve corresponding to the battery cell under test.

5. The battery health status prediction method according to claim 1 or 2, characterized in that, Before using a lifetime prediction model to predict health status decline information based on the electrical signal characteristics, the method further includes: Obtain multiple training datasets; With the goal of outputting health status decline information, sub-models corresponding to the training dataset are trained in the lifespan prediction model based on the training dataset.

6. The battery health status prediction method according to claim 5, characterized in that, The step of training sub-models corresponding to the training dataset in the lifespan prediction model based on the training dataset, with the goal of outputting health status decline information, includes: The training dataset is input into the sub-model to obtain the predicted value output by the sub-model; The target number of predicted values ​​with the largest offsets are removed from each of the predicted values, and the sub-model is optimized based on the average of the remaining predicted values.

7. The battery health status prediction method according to claim 5, characterized in that, The acquisition of multiple training datasets includes: Obtain sample charge-discharge curve data of sample batteries during the initial use phase, and at least one sample health status corresponding to the sample charge-discharge curve data; The sample charge-discharge curve data of the same sample battery, and the health status of a sample corresponding to the sample charge-discharge curve data, are determined as a training sample, and multiple training samples are obtained. A first number of training samples are randomly extracted from the plurality of training samples to form a training dataset.

8. A battery health status prediction device, characterized in that, include: The first processing module is used to acquire charge and discharge curve data of the battery cell under test during the initial use phase. The second processing module is used to extract electrical signal features corresponding to the charge-discharge curve data based on the charge-discharge curve data. The electrical signal features include the peak features of the capacity versus voltage rate curve, the target features of the charge-discharge plateau segment electrical signal curve, and the target features of the charge-discharge start segment electrical signal curve. The third processing module is used to use a lifetime prediction model to predict based on the electrical signal characteristics to obtain health status decay information; The target feature includes at least one of sample entropy and kurtosis; The initial use phase includes: the first charge-discharge cycle time period or the time period of the first preset number of charge-discharge cycles; The target features of the electrical signal curve of the charging and discharging plateau segment include one or more of the sample entropy of the voltage curve of the first charging plateau segment and the kurtosis of the energy curve of the first charging plateau segment. The target features of the electrical signal curve at the start of the charge / discharge phase include one or more of the sample entropy of the energy curve at the start of the discharge phase and the kurtosis of the voltage curve at the start of the charge phase; the first charging plateau phase is the negative electrode first plateau.

9. A single battery cell, characterized in that, The battery cell health status is predicted based on the battery health status prediction method as described in any one of claims 1-7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the battery health status prediction method as described in any one of claims 1-7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the battery health status prediction method as described in any one of claims 1-7.

Images