Method, apparatus, device, and storage medium for evaluating battery health status

Abstract

Embodiments of the present disclosure relate to a method, apparatus, device, and storage medium for evaluating the battery health state. The method proposed herein includes: determining a set of charging characteristics associated with the target device based on the historical charging information of the target device; determining a first evaluation value of the battery health state of the target device using a first evaluation model based on the set of charging characteristics; determining a first set of input characteristics of a second evaluation model based on the first evaluation value, the first set of input characteristics including at least one high-order feature, and the at least one high-order feature being used to characterize the change in the dependence of the battery health state on the set of charging characteristics; and determining a target value of the battery health state of the target device based on the second evaluation value output by the second evaluation model. In this way, by utilizing these high-order features, the second evaluation model can more effectively embed physical constraint information, thereby improving the accuracy of the constraints.

Description

Technical Field

[0001] Example embodiments of the present disclosure generally relate to the technical field of battery management, and more particularly, to a method, apparatus, device, and storage medium for evaluating the state of health of a battery. Background Art

[0002] The evaluation of the state of health of a battery is a key technology to ensure the performance, safety, and service life of the battery. Evaluating the state of health of a battery is crucial for timely detecting potential problems, taking maintenance measures, and avoiding safety accidents and performance degradation. Specific evaluation methods include charge-discharge tests, internal resistance detection, using a battery capacity tester, a battery management system, and cycle life assessment, etc. These methods can comprehensively and accurately reflect the health status of the battery and provide strong support for the maintenance, management, and optimization of the battery. Summary of the Invention

[0003] In a first aspect of the present disclosure, there is provided a method for evaluating the state of health of a battery. The method includes: determining a set of charging characteristics associated with a target device based on historical charging information of the target device; determining a first evaluation value of the state of health of the battery of the target device using a first evaluation model based on the set of charging characteristics; determining a first set of input characteristics of a second evaluation model based on the first evaluation value, the first set of input characteristics including at least one high-order feature, the at least one high-order feature being used to characterize the change in the degree of dependence of the state of health of the battery on the set of charging characteristics; and determining a target value of the state of health of the battery of the target device based on a second evaluation value output by the second evaluation model.

[0004] In a second aspect of the present disclosure, there is provided an apparatus for evaluating the state of health of a battery. The apparatus includes: a feature determination module configured to determine a set of charging characteristics associated with a target device based on historical charging information of the target device; a first evaluation value determination module configured to determine a first evaluation value of the state of health of the battery of the target device using a first evaluation model based on the set of charging characteristics; an input characteristic determination module configured to determine a first set of input characteristics of a second evaluation model based on the first evaluation value, the first set of input characteristics including at least one high-order feature, the at least one high-order feature being used to characterize the change in the degree of dependence of the state of health of the battery on the set of charging characteristics; and a target value determination module configured to determine a target value of the state of health of the battery of the target device based on a second evaluation value output by the second evaluation model.

[0005] In a third aspect of the present disclosure, there is provided an electronic device. The device includes at least one processing unit; and at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit. The instructions, when executed by the at least one processing unit, cause the device to execute the method for evaluating the state of health of a battery according to the first aspect.

[0006] In a fourth aspect of the present disclosure, a computer-readable storage medium is provided. A computer program is stored on the computer-readable storage medium, and the computer program can be executed by a processor to implement the method for evaluating the battery health state in the first aspect.

[0007] In a fifth aspect of the present disclosure, a computer program product is provided. The computer program product includes computer-executable instructions that, when executed by a processor, implement the method for evaluating the battery health state according to the first aspect of the present disclosure.

[0008] The solution of the present disclosure aims to overcome the limitations in traditional methods, especially to improve the deficiencies in physical information constraint modeling. Specifically, the solution of the present disclosure introduces a second evaluation model that can embed physical constraint information when evaluating the battery health state. The solution of the present disclosure enables the input of the second evaluation model to include at least one high-order feature, and these high-order features can characterize the change in the degree of dependence of the battery health state on the set of charging features. By utilizing these high-order features, the second evaluation model can more effectively embed physical constraint information, thereby improving the accuracy of the constraints and ultimately enhancing the accuracy of the battery health state evaluation.

[0009] It should be understood that the content described in this part is not intended to define the key features or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In combination with the accompanying drawings and with reference to the following detailed description, the above and other features, advantages, and aspects of the embodiments of the present disclosure will become more apparent. In the drawings, the same or similar reference numerals denote the same or similar elements, where:

[0011] Figure 1 A schematic diagram of an example environment according to an embodiment of the present disclosure is shown;

[0012] Figure 2 A flowchart of an example process of a method for evaluating the battery health state according to some embodiments of the present disclosure is shown;

[0013] Figure 3 A schematic diagram of an example of a time window according to some embodiments of the present disclosure is shown;

[0014] Figure 4 and Figure 5 A schematic diagram of an example of a physical information neural network according to the present disclosure is shown;

[0015] Figure 6A schematic structural block diagram of a device for evaluating the state of health of a battery according to some embodiments of the present disclosure is shown; and

[0016] Figure 7 A block diagram of an electronic device in which one or more embodiments of the present disclosure can be implemented is shown. Detailed implementation manners

[0017] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0018] It should be noted that the titles of any sections / subsections provided herein are not restrictive. Various embodiments are described throughout this document, and any type of embodiment can be included under any section / subsection. In addition, the embodiments described in any section / subsection can be combined with any other embodiments described in the same section / subsection and / or different sections / subsections in any manner.

[0019] In the description of the embodiments of the present disclosure, the term "including" and its like terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". There may also be other explicit and implicit definitions hereinafter. The terms "first", "second", etc. may refer to different or the same objects. There may also be other explicit and implicit definitions hereinafter.

[0020] The embodiments of the present disclosure may involve the user's data, data acquisition and / or use, etc. These aspects all comply with the corresponding laws, regulations and related provisions. In the embodiments of the present disclosure, all data collection, acquisition, processing, processing, forwarding, use, etc. are carried out on the premise that the user is aware and confirms. Correspondingly, when implementing the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the data or information that may be involved should be informed to the user and the user's authorization should be obtained through appropriate means according to the relevant laws and regulations. The specific informing and / or authorization methods may vary according to the actual situation and application scenarios, and the scope of the present disclosure is not limited in this regard.

[0021] In the solutions described in this specification and the embodiments, if personal information processing is involved, it will be processed on the premise of having a legal basis (such as obtaining the consent of the personal information subject, or being necessary for performing a contract), and will only be processed within the specified or agreed scope. If a user refuses to process personal information other than the necessary information required for basic functions, it will not affect the user's use of the basic functions.

[0022] With the continuous growth of the demand for sustainable energy solutions globally, new energy vehicles, as a key technological means to reduce greenhouse gas emissions and improve urban air quality, are becoming increasingly important. As the core component of new energy vehicles, the power battery not only determines the vehicle's endurance but also deeply affects the vehicle's overall performance, reliability, and safety. The state of health (SOH) of the power battery gradually deteriorates over time and use, directly affecting the endurance performance and service life of new energy vehicles. The industry generally takes the SOH dropping to 80% as the standard for battery retirement. Therefore, accurately evaluating the SOH of power batteries is of crucial significance for optimizing the design of the battery management system (BMS) and ensuring the safe operation of vehicles, and is an important topic in the research field of power batteries.

[0023] Currently, the evaluation methods of power battery SOH are mainly divided into two categories: physics-based models and data-driven models. Although physics models are stable and accurate, due to the influence of differences in battery chemical composition, the model parameters need to be adjusted specifically, resulting in limited generality; while data-driven models are efficient and have high accuracy, but their generality and stability are limited by feature extraction and model design.

[0024] The physics-informed neural network (PINN) combines the advantages of data-driven and physics models by integrating physical information constraints into the neural network and using the automatic differentiation mechanism. However, the features used in the physical information constraint modeling of this method are relatively simple, which limits the model's ability to obtain more accurate SOH estimates.

[0025] In view of this, embodiments of the present disclosure provide a solution for evaluating the battery health state. According to this solution, first, based on the historical charging information of the target device, a set of charging characteristics associated with the target device is determined. Then, based on the set of charging characteristics, a first evaluation value of the battery health state of the target device is determined using a first evaluation model. Next, based on the first evaluation value, a first set of input characteristics of a second evaluation model is determined. The first set of input characteristics includes at least one high-order characteristic, and the at least one high-order characteristic is used to characterize the change in the dependence degree of the battery health state on the set of charging characteristics. Subsequently, based on the second evaluation value output by the second evaluation model, the target value of the battery health state of the target device is determined.

[0026] As will be more clearly understood from the following description, the solution of the present disclosure aims to overcome the limitations in traditional methods, especially to improve the deficiency in physical information constraint modeling. Specifically, the solution of the present disclosure introduces a second evaluation model, which can embed physical constraint information when evaluating the battery health state. The solution of the present disclosure makes the input of the second evaluation model include at least one high-order characteristic, and these high-order characteristics can characterize the change in the dependence degree of the battery health state on the set of charging characteristics. By using these high-order characteristics, the second evaluation model can more effectively embed physical constraint information, thereby improving the accuracy of the constraints and ultimately achieving an improvement in the accuracy of battery health state evaluation.

[0027] Various example implementations of this solution will be further described in detail below with reference to the accompanying drawings.

[0028] Figure 1 FIG. shows a schematic diagram of an example environment 100 according to an embodiment of the present disclosure. Referring to Figure 1 FIG., the example environment 100 may include an electronic device 120 and a battery 130.

[0029] As an example, the electronic device 120 may be applied to a scenario of evaluating the battery health state based on a Physics-Informed Neural Network (PINN). The electronic device 120 may evaluate the battery health state of the battery 130 in response to an instruction sent by the user 140 through the terminal device 110. In addition, the terminal device 110 may present an interface 150. The target value of the battery health state determined by the electronic device 120 may be presented to the user 140 through the interface 150.

[0030] In some embodiments, the terminal device 110 can be any type of mobile terminal, fixed terminal, or portable terminal, including mobile phones, desktop computers, laptop computers, notebook computers, netbook computers, tablet computers, media computers, multimedia tablets, personal communication system (PCS) devices, personal navigation devices, personal digital assistants (PDAs), audio / video players, digital cameras / camcorders, positioning devices, television receivers, radio broadcast receivers, e-book devices, gaming devices, or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof. In some embodiments, the terminal device 110 can also support any type of user interface (such as a "wearable" circuit, etc.).

[0031] The electronic device 120 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks, and big data and artificial intelligence platforms. The electronic device 120 can include, for example, a computing system / server, such as a mainframe, an edge computing node, a computing device in a cloud environment, and so on. The electronic device 120 can provide background services for the application 130 that supports content presentation in the terminal device 110.

[0032] A communication connection can be established between the electronic device 120 and the terminal device 110. The communication connection can be established by wired or wireless means. The communication connection can include, but is not limited to, a Bluetooth connection, a mobile network connection, a universal serial bus connection, a Wi-Fi connection, etc., and the embodiments of the present disclosure are not limited in this regard. In the embodiments of the present disclosure, the electronic device 120 and the terminal device 110 can perform signaling interaction through the communication connection therebetween.

[0033] It should be understood that the structures and functions of the various elements in the environment 100 are described only for exemplary purposes, without implying any limitation to the scope of the present disclosure.

[0034] Figure 2 A flowchart of an example process 200 for evaluating the battery health status according to some embodiments of the present disclosure is shown. The process 200 can be implemented at the electronic device 120.

[0035] Referring to Figure 2 , at block 210, the electronic device 120 determines a set of charging characteristics associated with the target device based on the historical charging information of the target device.

[0036] As an example, the electronic device 120 can access and analyze the historical charging information stored or transmitted in real time by the target device. The target device may include a battery, and the corresponding historical charging information may include detailed data records of the battery during multiple charging processes, such as timestamps, voltages, currents, etc.

[0037] In some embodiments, a set of charging characteristics includes at least one of the following categories: the first type of charging characteristics, which are determined based on the current parameters indicated by the historical charging information; the second type of charging characteristics, which are determined based on the voltage parameters indicated by the historical charging information; or the third type of charging characteristics, which are determined based on the number of historical charging times indicated by the historical charging information.

[0038] As an example, the first type of charging characteristics reflects the variation of current during the charging process. The first type of charging characteristics may include the mean, variance, kurtosis, skewness, etc. of the current. The second type of charging characteristics reflects the variation of voltage during the charging process. The second type of charging characteristics may include the mean, variance, kurtosis, skewness, etc. of the voltage. The third type of charging characteristics reflects the number of charging times experienced by the battery and is an important indicator for evaluating the battery life and health status.

[0039] As an example, in order to better obtain a set of charging characteristics, the electronic device 120 can adopt a specific feature extraction framework. This feature extraction framework involves the extraction of multiple (e.g., 16) charging characteristics including the mean, variance, kurtosis, skewness, etc. of voltage and current. These charging characteristics can reflect the variation characteristics of voltage and current from different angles and provide a rich information basis for subsequent battery health status evaluation.

[0040] It should be noted that the above description of the types of charging characteristics is only an exemplary description, which does not constitute a limitation on the embodiments of the present disclosure. According to actual needs, the embodiments of the present disclosure can also extract more charging characteristics.

[0041] Figure 3 FIG. shows a schematic diagram of an example 300 of a time window according to some embodiments of the present disclosure. Referring to Figure 3 , in some embodiments, the electronic device 120 determines a charging sequence 310 corresponding to the historical charging process based on the historical charging information. Then, the electronic device 120 divides the charging sequence 310 into multiple groups of sequence segments based on multiple time windows 320 of different lengths, where each group of sequence segments corresponds to a time window 320 of the same length. Subsequently, the electronic device 120 determines a set of charging characteristics associated with the target device based on the multiple groups of sequence segments.

[0042] As an example, during the determination of the charging sequence 310, the electronic device 120 focuses on two key stages in the battery charging process: one is to extract detailed voltage data within a certain range before the power battery is fully charged; the other is to extract detailed current data during the small current (e.g., 0.5A - 1A) stage of constant voltage charging. The data in these two stages is particularly important for reflecting the health status of the battery. As an example, the charging sequence can be at least one of a voltage sequence or a current sequence. Taking the charging sequence as a voltage sequence as an example, the charging sequence can include multiple voltage values arranged in time - , where t represents the time when the battery reaches voltage .

[0043] In some embodiments, the electronic device 120 can determine the charging sequence 310 corresponding to the historical charging process in the following manner. First, the electronic device 120 determines the charging start voltage and the charging end voltage of the historical charging process corresponding to the charging sequence 310. Then, the electronic device 120 determines the starting position of the charging sequence 310 based on the difference between the charging end voltage and the charging start voltage and a predetermined coefficient

[0044] As an example, the charging start voltage can refer to the voltage value when the battery starts charging, that is, the voltage level in the initial stage of the charging process. It is usually lower than the rated voltage of the battery and marks the start of the charging cycle. The charging end voltage can refer to the voltage value when the battery is charged to the full charge state or a predetermined charging level, that is, the voltage upper limit at the end of the charging process. It is usually close to or equal to the rated voltage of the battery (but should not exceed the maximum allowable voltage of the battery) to prevent the battery from being damaged due to overcharging. The electronic device 120 can use the difference between the charging end voltage and the charging start voltage, and a predetermined coefficient, to calculate the starting position of the charging sequence 310

[0045] As an example, the electronic device 120 can calculate the difference between the charging end voltage minus the charging start voltage, and multiply this difference by the predetermined coefficient. The result obtained in this way can indicate the offset of the starting position of the charging sequence 310 relative to the charging end voltage. Then, the electronic device 120 can find the voltage value corresponding to this offset in the charging sequence 310, and thus use this voltage value as the starting point of the charging sequence 310

[0046] As an example, the predetermined coefficient is editable and can be adjusted according to actual needs. As an example, the predetermined coefficient can be set to 40%, which means that the electronic device 120 will determine the starting position of intercepting the charging sequence 310 based on 40% of the difference between the charging start voltage and the charging end voltage

[0047] In this way, embodiments of the present disclosure can adjust the starting position of the charging sequence 310 according to different battery types and charging rates, ensuring that the intercepted charging sequence 310 is both representative and adaptable to various situations.

[0048] It should be noted that during the processing of the charging sequence 310, the electronic device 120 can clean and preprocess the original data. For example, for null value data, the electronic device 120 can directly delete the entire row to ensure data integrity. For data with abnormal value ranges, the electronic device 120 can use methods such as 3-sigma for processing, that is, remove data whose voltage or current values exceed the range of the mean plus or minus three standard deviations, thereby restricting the data to fluctuate within a reasonable range.

[0049] After extracting the charging sequence 310, the electronic device 120 further extracts the above-mentioned set of charging characteristics on the charging sequence 310 by using the time window 320. As an example, a part of a set of charging characteristics is used to describe the change characteristics of voltage, and another part is used to describe the change characteristics of current.

[0050] As an example first, the electronic device 120 can divide the charging sequence 310 based on multiple time windows 320 of different lengths. The lengths W and step sizes S of these time windows 320 are preset, and the window length W is less than the total length L of the charging sequence 310. By sliding these time windows 320, the electronic device 120 can intercept multiple sequence segments of the same length from the charging sequence 310.

[0051] As an example, multiple time windows 320 correspond one-to-one with multiple sets of sequence segments, and different time windows 320 correspond to different sets of sequence segments. Taking a time window 320 with a length of w in multiple time windows 320 of different lengths as an example. n The electronic device 120 can start from the starting position of the charging sequence 310 and extract the first w 1 -w n voltage values as the first sequence segment of a set of sequence segments corresponding to this time window 320. Then, according to the set step size, move the time window 320 backward, and intercept the next w 10 -w n+9 voltage values as the next sequence segment. This process will be repeated until this time window 320 reaches the end of the charging sequence 310, or the remaining length of the charging sequence 310 is not sufficient to form a complete sequence segment.

[0052] As an example, in order to avoid dividing the charging sequence into too many sequence segments, which may lead to excessive input for the subsequent evaluation model and potential feature redundancy issues, the number of time windows 320 can be adjusted according to actual needs. For example, embodiments of the present disclosure can set two time windows 320 of different sizes. One time window 320 has a size of L / 2 and a step size of L / 2, which can split the original charging sequence 310 into two sequence segments. Another window has a size of L / 3 and a step size of L / 3, which can split the original charging sequence 310 into three sequence segments.

[0053] Of course, according to actual needs, embodiments of the present disclosure can also adopt more different time windows 320. As an example, embodiments of the present disclosure can try different combinations of time windows 320 by means of cross-validation or introducing dynamic programming, etc., so as to select an appropriate length and step size of the time window 320.

[0054] After obtaining multiple sets of sequence segments, the electronic device 120 can extract multiple charging features including the standard deviation, mean, skewness, kurtosis, slope, information entropy of voltage / current, as well as the duration and energy of constant voltage charging and the duration and energy of constant current charging, etc. based on these original charging sequences 310 and multiple sets of sequence segments by using the feature extraction framework described above. Among them, the feature extraction framework will be applied to each sequence segment divided by the time window 320. In other words, for each charging feature, this charging feature is determined based on both the original charging sequence 310 and multiple sets of sequence segments.

[0055] Referring to Figure 3 , for the charging feature 330 in a set of charging features, this charging feature can be obtained by concatenating the first feature 341 determined based on the original charging sequence 310 and multiple second features 342-1 and 342-2 determined by the time window 320.

[0056] In this way, the charging features will be extended. According to the number of sets of sequence segments, the dimension of the charging features will be extended multiple times accordingly. In this way, the charging features can have a feature space with more dimensions, so as to more comprehensively focus on the local data features of the charging sequence 310 and overcome the situation of insufficient focus on local information of the time series.

[0057] In some embodiments, the electronic device 120 determines multiple sets of window features regarding a target charging attribute based on multiple sets of sequence segments, where each set of window features corresponds to a time window 320 of the same length. Then, for a set of window features among the multiple sets of window features, the electronic device 120 deletes a target window feature from the set of window features in response to the dispersion degree of the target window feature relative to the set of window features being lower than a threshold, so as to update the multiple sets of window features. Subsequently, the electronic device 120 determines a set of charging features associated with the target device based on the updated multiple sets of window features. As an example, the electronic device 120 determines multiple sets of window features regarding a target charging attribute based on multiple sets of sequence segments. Each set of window features corresponds to a time window 320 of a specific length and contains a series of feature values extracted from within the window.

[0058] As an example, the charging sequence 310 can be understood as the raw data for battery health state assessment. The sequence segments can be understood as data segments obtained by slicing the raw data through the time window 320, and a set of sequence segments can be understood as all the data segments sliced by the same time window 320.

[0059] As an example, the window features can be understood as the features obtained after performing feature extraction on the sequence segments (this process can be achieved by calculating the mean, average value, etc. with the help of the feature extraction framework described above), and a set of window features can be understood as the set of all features obtained after performing feature extraction on all the sequence segments sliced by the same time window 320. For example, a set of window features can be as Figure 3 shown by the second features 342-1 and 342-2 in

[0060] As an example, the window features will be used to construct the corresponding charging features. For example, assuming that the first feature 341, the second features 342-1 and 342-2 are all features related to the voltage mean, then the electronic device 120 can form a charging feature related to the voltage mean by splicing the first feature 341, the second features 342-1 and 342-2.

[0061] As an example, in order to reduce the redundancy of the features (i.e., window features) extracted through the time window 320 and reduce the computational amount, the electronic device 120 may perform feature screening on each group of window features. As an example, the embodiments of the present disclosure may adopt the variance method. As an example, for each group of window features among multiple groups of window features, the electronic device 120 may calculate the variance value of each window feature within the group of window features. Subsequently, the electronic device 120 sorts these variance values from large to small and calculates a reference quantile (e.g., the median 50), and this reference quantile may be used as a threshold for determining whether a window feature is valid. If the variance value of a certain window feature is lower than this reference quantile, it means that its degree of dispersion is relatively low and the information it may contain is also less. Therefore, the electronic device 120 will delete this window feature from the group of window features. This process is performed for each group of window features, so as to update multiple groups of window features and remove those window features with lower variance values (i.e., lower degrees of dispersion).

[0062] It should be noted that the reference quantile may be a fixed value or a dynamic value that can change according to the variance size, and can be specifically determined according to actual needs. In addition, the embodiments of the present disclosure are not limited to the variance method as the feature screening method. According to actual needs, the embodiments of the present disclosure may also select other methods (such as correlation coefficient analysis, etc.) to perform feature screening.

[0063] After completing the feature screening, the electronic device 120 may determine a final set of charging features based on the updated multiple groups of window features. These features are screened, retaining sufficient information while reducing redundancy and computational complexity.

[0064] In block 220, the electronic device 120 determines a first evaluation value of the battery health state of the target device based on a set of charging features by using a first evaluation model.

[0065] As an example, the first evaluation model may be used to learn the mapping relationship between the charging features and the battery health state. As an example, the first evaluation model may include at least one of the following: a machine learning model based on a Multilayer Perceptron (MLP), a machine learning model based on a Long Short-Term Memory network, a machine learning model based on a Recurrent Neural Network, or a machine learning model based on a Gated Recurrent Unit.

[0066] As an example, the first evaluation model adopts a machine learning model based on a Long Short-Term Memory network. The Long Short-Term Memory network is a neural network structure specifically used to process time series data. It can control the flow of information through a forget gate, an input gate, and an output gate, so as to effectively capture the long-term and short-term dependencies in the time series.

[0067] As an example, to reduce the computational complexity, embodiments of the present disclosure reduce the number of neurons in each layer in the case of a machine learning model employing a long short-term memory network. For example, embodiments of the present disclosure reduce the number of neurons in each layer to one-third of the original architecture. It should be noted that this value is only for illustrative purposes and does not constitute a limitation on the embodiments of the present disclosure. According to actual needs, other values can also be adopted for the number of neurons in each layer. In this way, the first evaluation model can accurately learn the mapping relationship between the charging characteristics and the battery health state while considering the time factor.

[0068] As an example, in the case of training the first evaluation model, to ensure the coherence of the charging characteristic training samples and the sensitivity of the long short-term memory network-based machine learning model to the time sequence, embodiments of the present disclosure divide the charging characteristic training samples in a specific ratio in an orderly manner. Specifically, after sorting the charging characteristic training samples by time, they are divided into a training set, a validation set, and a test set in the ratio of 0.7, 0.1, and 0.2 in sequence. Among them, the first 70% of the data is used as the training set for training the first evaluation model; the next 10% is used as the validation set to determine the convergence of the first evaluation model during the training process; and the remaining 20% is used as the test set to determine the generalization ability of the first evaluation model. This orderly division method avoids the problem of time sequence chaos that may be brought about by random division.

[0069] As an example, to improve the training effect, embodiments of the present disclosure can also perform standardization processing on the charging characteristic training samples. As an example, this process can be represented by formula (1).

[0070] ; (1)

[0071] Wherein, represents the standardized charging characteristic training samples, represents the original charging characteristic training samples, and respectively represent the mean and standard deviation of the charging characteristic training samples. It should be noted that the standardization processing is performed on the training set, and the obtained mean and standard deviation are recorded to apply these parameters to the validation set and the test set to ensure the consistency of data processing.

[0072] In block 230, the electronic device 120 determines a first set of input features of the second evaluation model based on the first evaluation value, and the first set of input features includes at least one high-order feature, and the at least one high-order feature is used to characterize the change in the degree of dependence of the battery health state on a set of charging characteristics.

[0073] In some embodiments, the second evaluation model is implemented based on a physics-informed neural network to embed physical constraint information corresponding to at least one high-order feature.

[0074] As an example, Physics-Informed Neural Networks (PINN) is a technology that combines neural networks with physical models. It can not only perform feature learning like traditional neural networks but also make full use of existing physical information, such as the dynamic characteristics of the system, data distribution characteristics, etc., to provide new possibilities for solving complex systems. PINN encapsulates this physical information in the loss function of the neural network to guide the training process of the network, thereby achieving accurate modeling and prediction of complex systems.

[0075] Figure 4 A schematic diagram of an example 400 of a physics-informed neural network according to the present disclosure is shown. Referring to Figure 4 , as an example, the second evaluation model 412 can model the attenuation rate of the battery by embedding physical constraint information. Embodiments of the present disclosure can more accurately capture the laws and characteristics of battery attenuation with the help of the second evaluation model 412 by embedding physical constraint information corresponding to at least one high-order feature, thereby contributing to improving the accuracy of battery health state evaluation.

[0076] In some embodiments, the electronic device 120 determines a second set of input features 422 of the third evaluation model 413 based on the first evaluation value SOH1. The second set of input features 422 includes a first set of low-order features, and the first set of low-order features is used to characterize the change of the battery health state with a set of charging features. Then, the electronic device 120 determines at least one high-order feature in the first set of input features 421 based on the third evaluation value SOH3 determined by the third evaluation model 413.

[0077] As an example, the third evaluation model 413 is implemented based on a physics-informed neural network to embed physical constraint information corresponding to at least one low-order feature. As an example, the third evaluation model 413 can model the attenuation rate of the battery by embedding physical constraint information. Embodiments of the present disclosure can more comprehensively capture the laws and characteristics of battery attenuation with the help of the third evaluation model 413 and the second evaluation model 412 by respectively embedding physical constraint information corresponding to at least one high-order feature and physical constraint information corresponding to at least one low-order feature, thereby contributing to further improving the accuracy of battery health state evaluation.

[0078] As an example, the first set of low-order features may include first-order partial derivatives of the battery health state related to a set of charging features. As described above, a set of charging features may include a first type of charging feature X, a second type of charging feature (not shown in the figure), and a third type of charging feature T. Referring to Figure 4 , the first set of low-order features may include first-order partial derivatives of the battery health state related to the first type of charging feature X (or the second type of charging feature) calculated based on the first evaluation value SOH1 and first-order partial derivatives of the battery health state related to the third type of charging feature T calculated based on the first evaluation value SOH1 .

[0079] As an example, in addition to the first set of low-order features, the second set of input features 422 may further include a set of charging features and the first evaluation value SOH1. This information will be input into the third evaluation model 413 together with the first set of low-order features.

[0080] Referring to Figure 4 , the third evaluation model 413 may be connected in series before the second evaluation model 412. In this case, after determining the third evaluation value SOH3 based on the third evaluation model 413, the electronic device 120 will further determine at least one high-order feature in the first set of input features 421 based on the third evaluation value SOH3. As an example, at least one high-order feature may include second-order partial derivatives of the battery health state related to the first type of charging feature X (or the second type of charging feature) calculated based on the third evaluation value SOH3 . And second-order partial derivatives of the battery health state related to the third type of charging feature T calculated based on the third evaluation value SOH3 .

[0081] It should be noted that in addition to second-order partial derivatives, at least one high-order feature may further include higher-order partial derivatives, so as to further embed the accuracy of physical constraint information.

[0082] The third evaluation model 413 and the second evaluation model 412 are connected in series, so that the second evaluation model 412 further constrains the output of the third evaluation model 413, thereby achieving a more refined evaluation.

[0083] In some embodiments, the first set of input features 421 further includes at least one of the following: the first evaluation value SOH1, the third evaluation value SOH3, and the first set of low-order features. This information will be input into the second evaluation model 412 together with the high-order features described above, so that the second evaluation model 412 can comprehensively consider the state of health of the battery and its low-order and high-order features related to a set of charging features (such as first-order and second-order partial derivatives), thereby more accurately modeling the battery degradation and providing more accurate support for the evaluation of the state of health of the battery.

[0084] In block 240, the electronic device 120 determines a target value of the state of health of the battery of the target device based on the second evaluation value output by the second evaluation model 412.

[0085] As an example, in Figure 4 the example shown, the target value can be determined based on the second evaluation value SOH2 of the second evaluation model 412. Figure 5 A schematic diagram of an example 500 of a physics-informed neural network according to the present disclosure is shown. In addition to Figure 4 the third evaluation model 413 shown, referring to Figure 5 , some other embodiments of the present disclosure further provide a fourth evaluation model 414. Different from the third evaluation model 413, the fourth evaluation model 414 is connected in parallel with the second evaluation model 412. In this case, the target value will be jointly determined based on the outputs of the second evaluation model 412 and the fourth evaluation model 414.

[0086] In some embodiments, the electronic device 120 determines a third set of input features 423 of the fourth evaluation model 414 based on the first evaluation value SOH1. The third set of input features 423 includes a second set of low-order features, and the second set of low-order features is used to characterize the change of the state of health of the battery with a set of charging features. Then, the electronic device 120 determines a target value of the state of health of the battery of the target device based on the fourth evaluation value SOH4 and the second evaluation value SOH2 determined by the fourth evaluation model 414.

[0087] As an example, the fourth evaluation model 414 is implemented based on a physics-informed neural network to embed physical constraint information corresponding to at least one low-order feature. As an example, the fourth evaluation model 414 can model the decay rate of the battery by embedding physical constraint information. Different from the third evaluation model 413, the fourth evaluation model 414 is connected in parallel with the second evaluation model 412.

[0088] As an example, the second set of low-order features may include first-order partial derivatives of the state of health of the battery related to a set of charging features. As described above, a set of charging features may include the first type of charging feature X, the second type of charging feature, and the third type of charging feature T. Referring toFigure 5 The second set of low-order features may include first-order partial derivatives of the state of health of the battery calculated based on the first evaluation value SOH1 and related to the first type of charging feature X (the second type of charging feature or the fused feature of the first type of feature and the second type of feature). And first-order partial derivatives of the state of health of the battery calculated based on the first evaluation value SOH1 and related to the third type of charging feature T. .

[0089] As an example, in addition to the second set of low-order features, the third set of input features 423 may further include a set of charging features and the first evaluation value SOH1. This information will be input into the fourth evaluation model 414 together with the second set of low-order features.

[0090] Referring to Figure 5 , the fourth evaluation model 414 is in parallel with the second evaluation model 412. In this case, the electronic device 120 may directly determine at least one high-order feature in the first set of input features 421 based on the first evaluation value SOH1. As an example, at least one high-order feature may include second-order partial derivatives of the state of health of the battery calculated based on the first evaluation value SOH1 and related to the first type of charging feature X (the second type of charging feature or the fused feature of the first type of feature and the second type of feature). And second-order partial derivatives of the state of health of the battery calculated based on the first evaluation value SOH1 and related to the third type of charging feature T. .

[0091] In some embodiments, the first set of input features 421 further includes at least one of the following: the first evaluation value SOH1, the second set of low-order features. This information will be input into the second evaluation model 412 together with the high-order features described above, so that the second evaluation model 412 can comprehensively consider the state of health of the battery and its low-order and high-order features (such as first-order and second-order partial derivatives) related to a set of charging features, so as to more accurately model the battery attenuation and provide more accurate support for the evaluation of the state of health of the battery.

[0092] As an example, after determining the fourth evaluation value SOH4 and the second evaluation value SOH2, the electronic device 120 may determine the target value based on the fusion of the fourth evaluation value SOH4 and the second evaluation value SOH2. For example, the electronic device 120 may splice or perform weighted averaging on the fourth evaluation value SOH4 and the second evaluation value SOH2 to achieve the fusion of the fourth evaluation value SOH4 and the second evaluation value SOH2.

[0093] In this way, the fourth evaluation model 414 and the second evaluation model 412 are connected in parallel, making the calculations of the fourth evaluation model 414 and the second evaluation model 412 relatively independent, which is beneficial to improving the processing speed.

[0094] As an example, when training the second evaluation model 412, in addition to considering the loss function related to low-order features, a loss function related to high-order features is also added, which helps to more comprehensively measure the learning situation of the second evaluation model 412 during the training process.

[0095] As an example, before training the first evaluation model, the second evaluation model 412, the third evaluation model 413, and the fourth evaluation model 414, their training hyperparameters and structural parameters need to be initialized. The training hyperparameters include but are not limited to: the number of training epochs, batch size, learning rate and its decay coefficient, the weight of the PINN loss function, the number of early stop rounds, etc. In addition, as an example, AdamOptimizer can be selected as the optimizer and its hyperparameters can be kept consistent with the original PINN to ensure the stability and efficiency of model training.

[0096] It can be clearly understood from the various embodiments described above that the embodiments of the present disclosure aim to overcome the limitations in traditional methods, especially to improve the deficiencies in physics-informed constraint modeling. Specifically, the embodiments of the present disclosure introduce a second evaluation model 412, which can embed physics constraint information when evaluating the battery health state. The embodiments of the present disclosure make the input of the second evaluation model 412 include at least one high-order feature, and these high-order features can characterize the change in the dependence degree of the battery health state on the set of charging features. By using these high-order features, the second evaluation model 412 can more effectively embed physics constraint information, thereby improving the accuracy of the constraints and ultimately achieving an improvement in the accuracy of battery health state evaluation.

[0097] The embodiments of the present disclosure also provide corresponding devices for implementing the above methods or processes. Figure 6 The schematic structural block diagram of a device 600 for evaluating the battery health state according to some embodiments of the present disclosure is shown. The device 600 can be implemented as or included in an electronic device 120. Each module / component in the device 600 can be implemented by hardware, software, firmware, or any combination thereof.

[0098] Refer to Figure 6, the device 600 includes a feature determination module 610, a first evaluation value determination module 620, an input feature determination module 630, and a target value determination module 640. In some embodiments, the feature determination module 610 is configured to determine a set of charging features associated with the target device based on the historical charging information of the target device. The first evaluation value determination module 620 is configured to determine a first evaluation value of the battery health state of the target device by using a first evaluation model based on the set of charging features. The input feature determination module 630 is configured to determine a first set of input features of a second evaluation model based on the first evaluation value, where the first set of input features includes at least one high-order feature, and the at least one high-order feature is used to characterize the change in the dependence degree of the battery health state on the set of charging features. The target value determination module 640 is configured to determine a target value of the battery health state of the target device based on a second evaluation value output by the second evaluation model.

[0099] In some embodiments, a set of charging features includes at least one of the following categories: a first category of charging features determined based on current parameters indicated by historical charging information; a second category of charging features determined based on voltage parameters indicated by historical charging information; or a third category of charging features determined based on the number of historical charging times indicated by historical charging information.

[0100] In some embodiments, the input feature determination module 630 is further configured to: determine a second set of input features of a third evaluation model based on the first evaluation value, where the second set of input features includes a first set of low-order features used to characterize the change of the battery health state with the set of charging features; and determine at least one high-order feature in the first set of input features based on a third evaluation value determined by the third evaluation model.

[0101] In some embodiments, the first set of input features further includes at least one of the following: the first evaluation value, the third evaluation value, the first set of low-order features.

[0102] In some embodiments, the target value determination module 640 is configured to: determine a third set of input features of a fourth evaluation model based on the first evaluation value, where the third set of input features includes a second set of low-order features used to characterize the change of the battery health state with the set of charging features; and determine a target value of the battery health state of the target device based on a fourth evaluation value determined by the fourth evaluation model and the second evaluation value.

[0103] In some embodiments, the first set of input features further includes at least one of the following: the first evaluation value, the second set of low-order features.

[0104] In some embodiments, the second evaluation model is implemented based on a physics-informed neural network to embed physical constraint information corresponding to at least one high-order feature.

[0105] In some embodiments, the feature determination module 610 is further configured to: determine a charging sequence corresponding to a historical charging process based on historical charging information; divide the charging sequence into multiple groups of sequence segments based on multiple time windows of different lengths, where each group of sequence segments corresponds to a time window of the same length; and determine a group of charging features associated with the target device based on the multiple groups of sequence segments.

[0106] In some embodiments, the feature determination module 610 is further configured to: determine multiple groups of window features regarding a target charging attribute based on the multiple groups of sequence segments, where each group of window features corresponds to a time window of the same length; for a group of window features among the multiple groups of window features, in response to the discrete degree of a target window feature relative to the group of window features being lower than a threshold, delete the target window feature from the group of window features to update the multiple groups of window features; and determine a group of charging features associated with the target device based on the updated multiple groups of window features.

[0107] In some embodiments, the feature determination module 610 is further configured to: determine a charging start voltage and a charging cut-off voltage of a historical charging process corresponding to the charging sequence; and determine a starting position of the charging sequence based on the difference between the charging cut-off voltage and the charging start voltage and a predetermined coefficient.

[0108] Figure 7 The block diagram of an electronic device 700 in which one or more embodiments of the present disclosure can be implemented is shown. The electronic device 700 can be used, for example, to implement the electronic device 120 as shown in Figure 1 It should be understood that Figure 7 The electronic device 700 shown is merely exemplary and should not constitute any limitation to the functions and scopes of the embodiments described herein.

[0109] Referring to Figure 7 , the electronic device 700 is in the form of a general-purpose electronic device. The components of the electronic device 700 may include, but are not limited to, one or more processors or processing units 710, a memory 720, a storage device 730, one or more communication units 740, one or more input devices 750, and one or more output devices 760. The processing unit 710 can be an actual or virtual processor and can execute various processes according to the programs stored in the memory 720. In a multi-processor system, multiple processing units execute computer-executable instructions in parallel to improve the parallel processing ability of the electronic device 700.

[0110] The electronic device 700 generally includes multiple computer storage media. Such media can be any available media accessible to the electronic device 700, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory 720 can be volatile memory (such as registers, caches, random access memory (RAM)), non-volatile memory (such as read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. The storage device 730 can be removable or non-removable media and can include machine-readable media, such as a flash drive, a magnetic disk, or any other media that can be capable of storing information and / or data and can be accessed within the electronic device 700.

[0111] The electronic device 700 can further include additional removable / non-removable, volatile / non-volatile storage media. Although not shown in Figure 7 it, a disk drive for reading from or writing to a removable, non-volatile magnetic disk (such as a "floppy disk") and an optical disk drive for reading from or writing to a removable, non-volatile optical disk can be provided. In these cases, each drive can be connected to a bus (not shown) by one or more data media interfaces. The memory 720 can include a computer program product 725 having one or more program modules that are configured to perform the various methods or actions of the various embodiments of the present disclosure.

[0112] The communication unit 740 enables communication with other electronic devices through a communication medium. Additionally, the functions of the components of the electronic device 700 can be implemented by a single computing cluster or multiple computer machines that are capable of communicating through a communication connection. Thus, the electronic device 700 can operate in a networked environment using a logical connection with one or more other servers, network personal computers (PCs), or another network node.

[0113] The input device 750 can be one or more input devices, such as a mouse, a keyboard, a trackball, etc. The output device 760 can be one or more output devices, such as a display, a speaker, a printer, etc. The electronic device 700 can also communicate with one or more external devices (not shown) as needed through the communication unit 740, such as a storage device, a display device, etc., communicate with one or more devices that enable a user to interact with the electronic device 700, or communicate with any device that enables the electronic device 700 to communicate with one or more other electronic devices (such as a network card, a modem, etc.). Such communication can be performed via an input / output (I / O) interface (not shown).

[0114] According to an exemplary implementation of the present disclosure, there is provided a computer-readable storage medium having computer-executable instructions stored thereon, wherein the computer-executable instructions are executed by a processor to implement the method described above. According to an exemplary implementation of the present disclosure, there is also provided a computer program product, the computer program product being tangibly stored on a non-transitory computer-readable medium and including computer-executable instructions, and the computer-executable instructions being executed by a processor to implement the method described above.

[0115] Aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of methods, apparatuses, devices, and computer program products according to the present disclosure. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0116] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine such that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device is produced that implements the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause a computer, a programmable data processing device, and / or other devices to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured article, which includes instructions for implementing various aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0117] The computer-readable program instructions can be loaded onto a computer, other programmable data processing device, or other device, such that a series of operation steps are performed on the computer, other programmable data processing device, or other device to produce a computer-implemented process, so that the instructions executed on the computer, other programmable data processing device, or other device implement the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0118] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various implementations of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks may in fact be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0119] The various implementations of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art in the field without departing from the scope and spirit of the described implementations. The determination of the terms used herein is intended to best explain the principles of the implementations, the practical application, or the improvement of the technology in the market, or to enable other ordinary skilled persons in the art in the field to understand the various implementation manners disclosed herein.

Claims

1. A method for evaluating a battery health state, characterized in that: include: determining a set of charging characteristics associated with the target device based on historical charging information of the target device; Determine, based on the set of charging characteristics, a first evaluation value of a battery health state of the target device using a first evaluation model; Based on the first evaluation value, determining a first set of input features of a second evaluation model, the first set of input features comprising at least one high-order feature, the at least one high-order feature being used to characterize a change in the degree of dependence of the battery health state on the set of charging features; as well as determining a target value of the battery health state of the target device based on a second evaluation value output by the second evaluation model, Wherein, based on the first evaluation value, determining a first set of input features of the second evaluation model includes: Based on the first evaluation value, determining a second set of input features of a third evaluation model, the second set of input features comprising a first set of low-order features, the set of charging features, and the first evaluation value, the first set of low-order features being used to characterize changes in the battery health state as a function of the set of charging features; as well as Based on a third evaluation value determined by the third evaluation model, at least one high-order feature in the first group of input features is determined, wherein the at least one high-order feature includes a high-order partial differential of the battery health state related to the group of charging features determined based on the third evaluation value.

2. The method for evaluating the battery health status according to claim 1, characterized in that: The set of charging characteristics includes at least one of the following: a first type of charging feature, wherein the first type of charging feature is determined based on a current parameter indicated by the historical charging information; a second type of charging characteristic, wherein the second type of charging characteristic is determined based on a voltage parameter indicated by the historical charging information; or The third type of charging characteristics is determined based on the number of historical charging times indicated by the historical charging information.

3. The method for evaluating the battery health status according to claim 1, characterized in that: The first group of input features also includes at least one of the following: the first evaluation value, the third evaluation value, and the first group of low-order features.

4. The method for evaluating battery health status according to claim 1, characterized in that: Determining a target value of the battery health state of the target device based on a second evaluation value output by the second evaluation model includes: Based on the first evaluation value, determining a third set of input features of a fourth evaluation model, the third set of input features comprising a second set of low-order features, the second set of low-order features being used to characterize changes in the battery health state as a function of the set of charging features; and The target value of the battery health state of the target device is determined based on a fourth evaluation value determined by the fourth evaluation model and the second evaluation value.

5. The method for evaluating the battery health status according to claim 4, characterized in that: The first group of input features also includes at least one of the following: the first evaluation value and the second group of low-order features.

6. The method for evaluating the battery health status according to claim 1, characterized in that: The second evaluation model is implemented based on a physical information neural network to embed physical constraint information corresponding to the at least one high-order feature.

7. The method for evaluating the battery health status according to claim 1, characterized in that: Based on historical charging information of the target device, a set of charging characteristics associated with the target device is determined, including: Based on the historical charging information, determining a charging sequence corresponding to the historical charging process; Based on a plurality of time windows of different lengths, the charging sequence is divided into a plurality of groups of sequence segments, wherein each group of sequence segments corresponds to a time window of the same length; and The set of charging characteristics associated with the target device is determined based on the plurality of sets of sequence segments.

8. The method for evaluating the battery health status according to claim 7, characterized in that: Determining the set of charging characteristics associated with the target device based on the plurality of sets of sequence segments includes: Based on the multiple groups of sequence segments, determining multiple groups of window features about target charging attributes, wherein each group of window features corresponds to a time window of the same length; For a set of window features in the plurality of sets of window features, in response to a discrete degree of a target window feature relative to the set of window features being lower than a threshold, deleting the target window feature from the set of window features to update the plurality of sets of window features; and The set of charging characteristics associated with the target device is determined based on the updated sets of window characteristics.

9. The method for evaluating the battery health status according to claim 7, characterized in that: Determining the charging sequence corresponding to the historical charging process includes: Determining a charging start voltage and a charging end voltage of the historical charging process corresponding to the charging sequence; and The starting position of the charging sequence is determined based on the difference between the charging cut-off voltage and the charging start voltage and a predetermined coefficient.

10. A device for evaluating the health status of a battery, characterized in that: include: a characteristic determination module configured to determine a set of charging characteristics associated with the target device based on historical charging information of the target device; A first evaluation value determination module is configured to determine a first evaluation value of the battery health state of the target device using a first evaluation model based on the set of charging characteristics; an input feature determination module, configured to determine a first set of input features of a second evaluation model based on the first evaluation value, the first set of input features comprising at least one high-order feature, the at least one high-order feature being used to characterize a change in the degree of dependence of the battery health state on the set of charging features; as well as a target value determination module, configured to determine a target value of the battery health state of the target device based on a second evaluation value output by the second evaluation model, The input feature determination module is further configured as follows: Based on the first evaluation value, determining a second set of input features of a third evaluation model, the second set of input features comprising a first set of low-order features, the set of charging features, and the first evaluation value, the first set of low-order features being used to characterize changes in the battery health state as a function of the set of charging features; as well as Based on a third evaluation value determined by the third evaluation model, at least one high-order feature in the first group of input features is determined, wherein the at least one high-order feature includes a high-order partial differential of the battery health state related to the group of charging features determined based on the third evaluation value.

11. An electronic device, characterized in that: include: at least one processing unit; as well as At least one memory, the at least one memory is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit, and the instructions, when executed by the at least one processing unit, cause the electronic device to perform the method for evaluating the battery health status according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program can be executed by a processor to implement the method for evaluating the battery health state according to any one of claims 1 to 9.

13. A computer program product comprising computer executable instructions, characterized in that: When the computer executable instructions are executed by a processor, the method for evaluating the battery health state according to any one of claims 1 to 9 is implemented.

Images