Method, apparatus, device, medium and program product for predicting battery health status

Abstract

The present application relates to a method, apparatus, device, medium, and program product for predicting the health state of a battery. The method includes: obtaining design data, manufacturing process data, and test and characterization data of a target battery to be predicted; inputting the design data, manufacturing process data, and test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the health state curve of the target battery; and determining the health state curve corresponding to the target battery according to a pre-constructed health state decay relational expression and at least one of the key features, where the health state decay relational expression is used to characterize the correlation between the health state and the usage duration. Using the present application can accurately predict the health state of the battery and accelerate the battery R & D speed.

Description

Technical Field

[0001] The present application relates to the technical field of batteries, and particularly to a method, device, equipment, medium and program product for predicting the state of health of a battery. Background Art

[0002] With the development of new energy technologies, secondary batteries represented by lithium-ion batteries have gradually gained market favor due to their high energy density and low cost, and have been widely used in fields such as vehicle power and energy storage.

[0003] Currently, due to the long test cycle and wide adjustment dimension of the battery SOH (State of Health), the R & D cycle has been severely extended. Therefore, there is an urgent need for a method that can accurately predict the SOH to accelerate the battery R & D speed. Summary of the Invention

[0004] Based on the above problems, the present application provides a method, device, equipment, medium and program product for predicting the state of health of a battery, which can accurately predict the state of health of the battery and accelerate the battery R & D speed.

[0005] In a first aspect, the present application provides a method for predicting the state of health of a battery. The method includes: obtaining the design data, manufacturing process data, test and characterization data of a target battery to be predicted; inputting the design data, manufacturing process data, test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the state of health curve of the target battery; determining the state of health curve corresponding to the target battery according to a pre-constructed state of health decay relational expression and at least one key feature; wherein, the state of health decay relational expression is used to characterize the correlation between the state of health and the usage duration.

[0006] The technical solution provided by the embodiments of the present application is that the feature prediction model can predict the state of health of the battery based on the design data, manufacturing process data, test and characterization data, fully considering the influences in aspects such as materials and processes; and, the feature prediction model focuses on the core decay mechanism and outputs key features related to the change trend of the state of health, which is beneficial to improving the prediction accuracy of the state of health, thereby accelerating the battery R & D speed.

[0007] In some embodiments, determining the health state curve corresponding to the target battery according to a pre-constructed health state decay relationship and at least one key feature includes: substituting at least one key feature as a parameter into the health state decay relationship; calculating the health state corresponding to different usage durations according to the health state decay relationship after substituting the parameter, so as to obtain the health state curve corresponding to the target battery. In the technical solution provided by the embodiments of the present application, the feature prediction model obtains key features by learning the dynamic relationship between features through data, and the health state decay relationship provides prior knowledge of battery decay (such as the non-linear law of degradation). The combination of the two can accurately predict the health state of the target battery and reduce the dependence on massive data, especially performing better in scenarios with insufficient data or high noise.

[0008] In some embodiments, the method further includes: obtaining the actual curves of the health states of multiple sample batteries, and determining the actual curves belonging to the same design type according to the design data of each sample battery; performing curve fitting on multiple actual curves belonging to the same design type and an initial decay formula, and determining the target values of each parameter term in the initial decay formula according to the fitting situation; substituting the target values of each parameter term into the initial decay formula to obtain the health state decay relationship. In the technical solution provided by the embodiments of the present application, the actual curves of the health states of the sample batteries are used for fitting the decay formula, automatically capturing the core law of the battery decay process. The extracted features are directly related to the SOH dynamics, ensuring the physical meaning and effectiveness of the features. Moreover, it provides a basis for using the health state decay relationship to predict the SOH.

[0009] In some embodiments, the parameter term in the health state decay relationship includes an exponential term. Performing curve fitting on multiple actual curves belonging to the same design type and an initial decay formula, and determining the target values of each parameter term in the initial decay formula according to the fitting situation includes: obtaining a candidate value set of the exponential term; in each round of iteration, substituting a candidate value in the candidate value set into the initial decay formula to obtain an intermediate decay formula, respectively performing fitting processing on multiple actual curves belonging to the same design type and the curve corresponding to the intermediate decay formula to obtain an average fitting error; determining the target value of the exponential term according to the average fitting errors of multiple rounds of iteration. In the technical solution of the embodiments of the present application, the target value of the exponential term is determined through iteration, which can avoid overfitting and ensure that the target value of the exponential term is more stable and reliable.

[0010] In some embodiments, the health state degradation relationship includes one of the Arrhenius exponential formula, the battery degradation power function formula, and the inflection point fitting formula; the parameter terms in the Arrhenius exponential formula include the battery degradation rate and the constant term; the parameter terms in the battery degradation power function formula include the coefficients of the square root, linear, and second-order power terms, as well as the constant term; the parameter terms in the inflection point fitting formula include the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point. The health state degradation relationship in the embodiments of the present application provides the key features of the health state curve (such as slope, inflection point, attenuation rate, etc.), can automatically capture the core law of the battery degradation process, and the extracted features are directly related to the SOH dynamics, ensuring the physical meaning and effectiveness of the features.

[0011] In some embodiments, the method further includes: obtaining the design data, process data, test and characterization data of a plurality of sample batteries, and the target values of the parameter terms obtained by fitting according to the actual curves of the health states of the plurality of sample batteries; using the design data, process data, test and characterization data as training samples, and using the target values of the parameter terms as labels for model training to obtain a feature prediction model. In the technical solution of the embodiments of the present application, the target values of the parameter values are the key features of the health state curve. Directly modeling the key features can focus on the core degradation mechanism, reduce the model complexity, and at the same time, the interpretability of the features also provides a physical basis for subsequent analysis. Moreover, taking the key features as the direct output target of the machine learning model can weaken the influence of noise and the overfitting problem of machine learning, and improve the accuracy and stability of model training. Further, the trained model only needs to predict a small number of key features (such as 2-3 parameters), rather than the entire SOH curve, reducing the computational burden of real-time prediction.

[0012] In some embodiments, obtaining the design data, process data, test and characterization data of the target battery to be predicted includes: obtaining the design information, process information, and test and characterization information of the target battery; performing data preprocessing on the design information, process information, and test and characterization information to obtain the design data, process data, and test and characterization data of the target battery. In the technical solution of the embodiments of the present application, the impacts of design, process, test and other factors on the health state are fully considered, enabling the prediction of SOH for the batteries under research. Moreover, considering physical and chemical parameters, process adjustments, etc. in the prediction is beneficial to improving the prediction accuracy of the health state.

[0013] In a second aspect, the present application also provides a prediction device for the health state of a battery, and the device includes:

[0014] A data acquisition module, configured to obtain the design data, process data, test and characterization data of the target battery to be predicted;

[0015] A feature prediction module, configured to input design data, manufacturing process data, and test and characterization data of a target battery into a pre-trained feature prediction model for prediction processing, so as to obtain at least one key feature of the health state curve of the target battery;

[0016] A health state determination module, configured to determine a health state curve corresponding to the target battery according to a pre-constructed health state decay relational expression and at least one key feature; wherein, the health state decay relational expression is used to characterize the association relationship between the health state and the usage duration.

[0017] In a third aspect, the present application further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the method according to any one of the first aspect is implemented.

[0018] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method according to any one of the first aspect is implemented.

[0019] In a fifth aspect, the present application further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method according to any one of the first aspect is implemented. Description of the Drawings

[0020] By reading the detailed description of the optional embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the optional embodiments, and are not considered to be a limitation of the present application. Moreover, in all the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0021] Figure 1 is a schematic diagram of an application environment of a method for predicting the health state of a battery according to an embodiment of the present application;

[0022] Figure 2 is a schematic flowchart of a method for predicting the health state of a battery according to an embodiment of the present application;

[0023] Figure 3 is a schematic flowchart of a step of determining a health state curve corresponding to a target battery according to an embodiment of the present application;

[0024] Figure 4 is a schematic flowchart of a step of constructing a health state decay relational expression according to an embodiment of the present application;

[0025] Figure 5 is a schematic flowchart of a step of determining target values of each parameter item according to an embodiment of the present application;

[0026] Figure 6It is a schematic flowchart of the training process of the feature prediction model according to an embodiment of the present application;

[0027] Figure 7 It is a schematic flowchart of the steps of obtaining the design data, manufacturing process data, and test and characterization data of the target battery according to an embodiment of the present application;

[0028] Figure 8 It is one of the structural block diagrams of the prediction device for the battery health state according to an embodiment of the present application;

[0029] Figure 9 It is the second structural block diagram of the prediction device for the battery health state according to an embodiment of the present application;

[0030] Figure 10 It is the third structural block diagram of the prediction device for the battery health state according to an embodiment of the present application;

[0031] Figure 11 It is the internal structure diagram of a computer device according to an embodiment of the present application. Detailed implementation manners

[0032] Next, embodiments of the technical solutions of the present application will be described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application, so they are only examples and cannot be used to limit the protection scope of the present application.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above accompanying drawing descriptions are intended to cover non-exclusive inclusion.

[0034] In the description of the embodiments of this application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order, or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, "a plurality" means more than two unless otherwise specifically defined.

[0035] Referring to "embodiment" herein means that a specific feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0036] In the description of the embodiments of the present application, the term "and / or" is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0037] In the description of the embodiments of the present application, the term "plurality" refers to two or more (including two). Similarly, "multiple groups" refers to two or more groups (including two groups), and "multiple pieces" refers to two or more pieces (including two pieces).

[0038] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installation", "connection", "connection", and "fixation" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or integrated; it may also be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to specific situations.

[0039] With the development of new energy technologies, secondary batteries represented by lithium-ion batteries have gradually gained market favor due to their high energy density and low cost, and have been widely used in fields such as vehicle power and energy storage. During the use and storage of batteries, capacity attenuation, internal resistance increase, and performance degradation occur due to cyclic charge and discharge, chemical reaction aging, and environmental factors (such as temperature, deep discharge, etc.). Therefore, it is necessary to evaluate the SOH of the battery and quantify the degree of battery aging or damage by comparing the deviation between its current performance and the initial design performance.

[0040] Currently, due to the long test cycle and wide adjustment dimension of battery SOH, the R & D cycle has been severely extended. Therefore, there is an urgent need for a method that can accurately predict SOH to accelerate the R & D speed of batteries. However, most of the existing SOH prediction methods are for commercially available batteries and do not consider the impact of adjustments at the material and process levels on SOH. Therefore, the prediction effect for R & D batteries is not ideal. Moreover, the relationship between material and process adjustments and SOH is non-linear, and some existing prediction models are also difficult to accurately predict SOH. Further, existing SOH prediction methods are difficult to associate with the battery degradation mechanism and thus difficult to be used to generalize the battery degradation mechanism.

[0041] To study the above problems, an embodiment of the present application provides a method for predicting the state of health of a battery, which includes obtaining design data, manufacturing process data, and test and characterization data of a target battery to be predicted; inputting the design data, manufacturing process data, and test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the state of health curve of the target battery; and determining the state of health curve corresponding to the target battery according to a pre-constructed state of health degradation relationship and at least one key feature. In the technical solution provided by the embodiment of the present application, the feature prediction model can predict the state of health of the battery based on the design data, manufacturing process data, and test and characterization data, fully considering the influences of aspects such as materials and processes; moreover, the feature prediction model focuses on the core degradation mechanism and outputs key features related to the changing trend of the state of health, which is beneficial to improving the prediction accuracy of the state of health and thus accelerating the battery R & D speed.

[0042] The method for predicting the state of health of a battery provided by an embodiment of the present application can be applied to an application environment as Figure 1 shown. This application environment includes a terminal 102 and a server 104. Among them, the terminal 102 communicates with the server 104 through a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed in the cloud or other network servers. The data storage system can store various data in the battery R & D and production processes and various data related to the state of health of the battery. When it is necessary to predict the SOH of the target battery, the terminal 102 can instruct the server 104 about the target battery to be predicted. The server 104 can pre-train a feature prediction model and construct a state of health degradation relationship, and then obtain the design data, manufacturing process data, and test and characterization data of the target battery from the data storage system, and use the design data, manufacturing process data, and test and characterization data of the target battery, the feature prediction model, and the state of health degradation relationship to predict the state of health of the target battery. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0043] According to some embodiments of the present application, referring to Figure 2 , a method for predicting the state of health of a battery is provided. Taking the example that this method is applied to the Figure 1 server, it can include the following steps:

[0044] Step 201, obtain the design data, manufacturing process data, and test and characterization data of the target battery to be predicted.

[0045] Among them, the test and characterization data includes the SOH sequence data of the battery.

[0046] The terminal obtains the design data, process data, test and characterization data of the target battery to be predicted, and sends the design data, process data, test and characterization data to the server. The server can then obtain the design data, process data, test and characterization data of the target battery.

[0047] In some embodiments, the terminal obtains the battery information of the target battery to be predicted, such as battery identification, battery type, etc.; then, the terminal sends the battery information to the server. The server receives the battery information and obtains the design data, process data, test and characterization data of the target battery from the data storage system according to the battery information.

[0048] In some embodiments, the server can also trigger a prediction task at a preset time and automatically obtain the design data, process data, test and characterization data of the target battery from the data storage system.

[0049] It should be noted that the acquisition methods of the design data, process data, test and characterization data are not limited to the above examples and can be set according to the actual situation.

[0050] Step 202, input the design data, process data, test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the health state curve of the target battery.

[0051] Among them, the key feature is used to characterize the representative features of the health state curve during the change process. For example, the key feature can be the inflection point position of the health state curve, the curve slope before and after the inflection point, the decline rate of the health state, and so on. It should be noted that the key feature is not limited to the above examples and can be set according to the actual situation.

[0052] The server pre-trains a feature prediction model. The input of this feature prediction model is the design data, process data, test and characterization data, and the output is the key feature of the health state curve.

[0053] In practical applications, after obtaining the design data, process data, test and characterization data of the target battery, the server inputs the design data, process data, test and characterization data into the feature prediction model. The feature prediction model analyzes the internal reaction rate, physical and chemical changes, and the decline of the health state of the target battery according to the design data, process data, test and characterization data, and then outputs the key feature of the health state curve of the target battery.

[0054] Step 203, determine the health state curve corresponding to the target battery according to the pre-constructed health state decline relational expression and at least one key feature.

[0055] Among them, the health state decay relationship formula is used to characterize the correlation between the health state and the usage duration. For example, the health state decay relationship formula can be expressed as SOH = f(t), where SOH is the health state and t is the usage duration. The usage duration can be the date, the number of hours, the number of cycles, etc.

[0056] The health state curve includes the health states corresponding to different usage durations. For example, the health state curve includes the health state corresponding to January 1st, the health state corresponding to February 1st... the health state corresponding to December 1st; the health state curve can also include the health state corresponding to using t1 hours, the health state corresponding to using t2 hours... the health state corresponding to using tn hours; the health state curve can also include the health state corresponding to 10 cycles, the health state corresponding to 50 cycles... the health state corresponding to 100 cycles.

[0057] Based on the key features, the server knows the decay situation of the health state, determines the important parameters in the health state decay relationship formula; then, the server sets sampling points for multiple usage times, substitutes each sampling point into the health state decay relationship formula, obtains the health state corresponding to each sampling point, and forms the health state curve of the target battery from the health states corresponding to multiple sampling points.

[0058] In the above embodiments, the design data, manufacturing process data, test and characterization data of the target battery to be predicted are obtained; the design data, manufacturing process data, test and characterization data of the target battery are input into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the health state curve of the target battery; according to the pre-constructed health state decay relationship formula and at least one key feature, the health state curve corresponding to the target battery is determined. In the technical solution provided by the embodiments of the present application, the feature prediction model can predict the health state of the battery based on the design data, manufacturing process data, test and characterization data, fully considering the influences of aspects such as materials and processes; moreover, the feature prediction model focuses on the core decay mechanism and outputs key features related to the changing trend of the health state, which is beneficial to improving the prediction accuracy of the health state, thereby accelerating the battery R & D speed.

[0059] According to some embodiments of the present application, the health state decay relationship formula includes one of the Arrhenius exponential formula, the battery decay power function formula, and the inflection point fitting formula.

[0060] The Arrhenius exponential formula refers to formula (1), and the parameter terms in the Arrhenius exponential formula include the battery decay rate and the constant term.

[0061] ------------------------------------ (1)

[0062] Among them, x is the usage duration, y is the state of health SOH, b is the constant term, k is the battery degradation rate, and z is the exponential term. This formula is derived based on the Arrhenius formula. The parameter term k is greatly affected by temperature, and the parameter term z is related to the battery degradation reaction. The Arrhenius formula describes the relationship between the reaction rate constant of a chemical reaction and temperature.

[0063] The battery degradation power function formula refers to Formula (2). The parameter terms in the battery degradation power function formula include the coefficients of the square root, linear, and second-order power terms, as well as the constant term.

[0064] ----------------------------- (2)

[0065] Among them, x is the usage duration, y is the state of health SOH, a is the constant term, and b, c, and d are the coefficients of the square root, linear, and second-order power terms respectively. This formula is easy to fit and has good generalization.

[0066] The inflection point fitting formula refers to Formula (3). The parameter terms in the inflection point fitting formula include the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point.

[0067] --- (3)

[0068] Among them, x is the usage duration; y is the state of health SOH; a is the abscissa of the slope change point; b is the starting slope; c is the ordinate of the starting time, that is, the ordinate when x = 0; d is the slope difference before and after the inflection point.

[0069] It should be noted that Formula (3) has good fitting ability for the SOH curve with a significant inflection point, and the formula parameters are related to the geometric information of the SOH curve, but it is not applicable to fitting the SOH curve without an inflection point.

[0070] The state of health degradation relationship in the embodiments of the present application provides the key features (such as slope, inflection point, decay rate, etc.) of the state of health curve, can automatically capture the core law of the battery degradation process, and the extracted features are directly related to the SOH dynamics, ensuring the physical meaning and effectiveness of the features.

[0071] According to some embodiments of the present application, referring to Figure 3 , "determining the state of health curve corresponding to the target battery according to the pre-constructed state of health degradation relationship and at least one key feature" in the above embodiments may include the following steps:

[0072] Step 301, substituting at least one key feature as a parameter into the state of health degradation relationship.

[0073] Taking the relationship formula for the decline in health state as the exponential formula of Arrhenius as an example, the key features include the battery decline rate and the constant term. That is, the design data, process data, test and characterization data of the target battery are input into the feature prediction model, and the feature prediction model outputs the target values of the battery decline rate and the constant term. When calculating the SOH of the target battery, the battery decline rate and the constant term are substituted into the relationship formula for the decline in health state.

[0074] Taking the relationship formula for the decline in health state as the power function formula for battery decline as an example, the key features include the coefficient terms of the square root, linear, and second-order power, as well as the constant term. That is, the design data, process data, test and characterization data of the target battery are input into the feature prediction model, and the feature prediction model outputs the coefficient terms of the square root, linear, and second-order power, as well as the constant term. When calculating the SOH of the target battery, the coefficient terms of the square root, linear, and second-order power, as well as the constant term are substituted into the relationship formula for the decline in health state.

[0075] Taking the relationship formula for the decline in health state as the inflection point fitting formula as an example, the key features include the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point. That is, the design data, process data, test and characterization data of the target battery are input into the feature prediction model, and the feature prediction model outputs the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point. When calculating the SOH of the target battery, the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point are substituted into the relationship formula for the decline in health state.

[0076] Step 302: Calculate the health state corresponding to different usage durations according to the relationship formula for the decline in health state after substituting the parameters, and obtain the health state curve corresponding to the target battery.

[0077] After substituting the parameters into the relationship formula for the decline in health state, a plurality of pre-selected sampling points (usage duration x) are sequentially substituted into the relationship formula for the decline in health state, and a plurality of health states y are calculated, and the plurality of health states y form the health state curve corresponding to the target battery.

[0078] For example, the plurality of sampling points are respectively x1 = 50 hours, x2 = 100 hours, x3 = 150 hours...; substituting x1 into the relationship formula for the decline in health state, the health state y1 is calculated; substituting x2 into the relationship formula for the decline in health state, the health state y2 is calculated; substituting x3 into the relationship formula for the decline in health state, the health state y3 is calculated. And so on, a plurality of health states y1, y2, y3... are obtained, and the plurality of health states form the health state curve corresponding to the target battery.

[0079] In some embodiments, boundary conditions can be set, and the health state curve of the target battery can be determined according to the boundary conditions. For example, if the boundary condition is SOH≥80%, when forming the health state curve, the part with a health state less than 80% is discarded, and the part with a health state greater than or equal to 80% is retained.

[0080] In the above embodiments, at least one key feature is substituted as a parameter into the health state decay relationship; the health states corresponding to different usage durations are calculated according to the health state decay relationship after substituting the parameters, and the health state curve corresponding to the target battery is obtained. In the technical solution provided by the embodiments of the present application, the feature prediction model obtains key features by learning the dynamic relationship between features through data, and the health state decay relationship provides prior knowledge of battery decay (such as the non-linear law of degradation). The combination of the two can accurately predict the health state of the target battery and reduce the dependence on a large amount of data, especially performing better in scenarios where data is insufficient or noisy.

[0081] According to some embodiments of the present application, referring to Figure 4 , the process of constructing the health state decay relationship may include the following steps:

[0082] Step 401, obtain the actual curves of the health states of multiple sample batteries, and determine the actual curves belonging to the same design type according to the design data of each sample battery.

[0083] The multiple sample batteries can be batteries tested by traditional testing methods, or batteries in use or that have been used.

[0084] For each sample battery, during the life cycle of the sample battery, the health state of the sample battery is recorded multiple times, and the actual curve of the health state of the sample battery is constructed according to the recorded data.

[0085] In some embodiments, the sample battery is set in a new energy vehicle, and the new energy vehicle uploads the health state of the sample battery to the server at preset time intervals; or, the new energy vehicle uploads the health state of the sample battery to the server at preset cycle numbers. The server can establish the actual curve of the health state of the sample battery with time or cycle number as the horizontal axis and health state as the vertical axis according to the data uploaded by the new energy vehicle.

[0086] In some embodiments, the test device tests the sample battery and uploads the health state of the sample battery to the server at preset time intervals; or, at preset cycle numbers. The server can establish the actual curve of the health state of the sample battery with time or cycle number as the horizontal axis and health state as the vertical axis according to the data uploaded by the test device.

[0087] The actual curves of the health states of multiple sample batteries form a curve sequence Batt = {batt1, batt2, batt3,..., battn}, where battn is the actual curve corresponding to the nth sample battery.

[0088] It should be noted that traditional methods may use batteries of different design types for fitting, which is likely to lead to inaccurate fitting. In this application, first, the actual curves of the health states of multiple sample batteries are obtained, and then the actual curves are grouped according to the design data, so as to select the actual curves belonging to the same design type, and then fit the batteries of a specific design, which can avoid cross-design interference and improve the accuracy of the fitting result.

[0089] Step 402: Perform curve fitting on multiple actual curves belonging to the same design type and the initial decay formula, and determine the target values of each parameter term in the initial decay formula according to the fitting situation.

[0090] Among them, the initial decay formula is a health state decay relationship formula with unknown parameter terms. For example, the initial decay formula is the Arrhenius exponential formula with unknown battery decay rate and constant term; or, the initial decay formula is the battery decay power function formula with unknown coefficients of square root, linear, and second-order power and constant term; or, the initial decay formula is the inflection point fitting formula with unknown abscissa of the slope change point, starting slope, ordinate of the starting time, and slope difference before and after the inflection point.

[0091] Perform curve fitting on multiple actual curves belonging to the same design type and the initial decay formula. For example, perform curve fitting on the actual curve batt1 and the Arrhenius exponential formula with unknown parameters, perform curve fitting on the actual curve batt2 and the Arrhenius exponential formula with unknown parameters... perform curve fitting on the actual curve battn and the Arrhenius exponential formula with unknown parameters. According to the fitting situation, the target values of the battery decay rate and constant term in the Arrhenius exponential formula can be determined.

[0092] For the process of performing fitting on multiple actual curves belonging to the same design type and the battery decay power function formula and the inflection point fitting formula to obtain the target values of each parameter term, referring to the above example, the embodiments of this application will not be elaborated here.

[0093] Step 403: Substitute the target values of each parameter term into the initial decay formula to obtain the health state decay relationship formula.

[0094] After obtaining the target values of each parameter item and substituting the target values into the initial degradation formula, a health state degradation relationship can be obtained. For example, substituting the battery degradation rate and the constant term into the exponential formula of Arrhenius with unknown parameters, an exponential formula of Arrhenius with known parameters is obtained, that is, a health state degradation relationship is obtained.

[0095] In the above embodiments, the actual curves of the health states of multiple sample batteries are obtained, and the actual curves belonging to the same design type are determined according to the design data of each sample battery; the actual curves of multiple samples belonging to the same design type are curve-fitted with the initial degradation formula, and the target values of each parameter item in the initial degradation formula are determined according to the fitting situation; the target values of each parameter item are substituted into the initial degradation formula to obtain a health state degradation relationship. The technical solution provided by the embodiments of the present application uses the actual curves of the health states of sample batteries to fit the degradation formula, automatically captures the core law of the battery degradation process, and the extracted features are directly related to the SOH dynamics, ensuring the physical meaning and effectiveness of the features. Moreover, it provides a basis for predicting SOH using the health state degradation relationship.

[0096] According to some embodiments of the present application, the parameter items in the health state degradation relationship include an exponential term. Referring to Figure 5 , in the above embodiments, "the actual curves of multiple samples belonging to the same design type are curve-fitted with the initial degradation formula, and the target values of each parameter item in the initial degradation formula are determined according to the fitting situation" may include the following steps:

[0097] Step 501, obtain a set of candidate values for the exponential term.

[0098] Taking the exponential formula of Arrhenius as the health state degradation relationship as an example, in the fitting process, first obtain a set of candidate values Z = {z1, z2, z3... zm} for the exponential term z, where zm is the candidate value used in the m-th iteration, and m is a positive integer.

[0099] Step 502, in each round of iteration, substitute a candidate value in the set of candidate values into the initial degradation formula to obtain an intermediate degradation formula, and respectively perform fitting processing on the actual curves of multiple samples belonging to the same design type and the curves corresponding to the intermediate degradation formula to obtain an average fitting error.

[0100] Taking the initial decay formula as an Arrhenius exponential formula with unknown parameters as an example, in the first round of iteration, the first candidate value z1 in the candidate value set is substituted into the initial decay formula to obtain an intermediate decay formula. Then, multiple actual curves belonging to the same design type are successively fitted with the curve corresponding to the intermediate decay formula to obtain the constant term b and the battery decay rate k in the Arrhenius exponential formula. After each actual curve is fitted with the curve corresponding to the intermediate decay formula, a fitting error can be obtained. The fitting errors corresponding to multiple actual curves belonging to the same design type are averaged to obtain the average fitting error E1 of the first round of iteration.

[0101] According to the above iteration process, the second round of iteration, the third round of iteration until the end of the mth round of iteration can be calculated to obtain multiple average fitting errors E1, E2... Em.

[0102] Step 503, determine the target value of the exponential term according to the average fitting errors of multiple rounds of iteration.

[0103] Determine the minimum average fitting error Emin in multiple rounds of iteration, and determine the candidate value zmin corresponding to the minimum average fitting error Emin as the target value of the exponential term.

[0104] It can be understood that if the target value z of the exponential term is selected only based on a single battery, it may be interfered by experimental noise; while in this application, the average fitting error is calculated for sample batteries of the same design type, which can smooth out individual differences through population data and ensure that the selected target value z of the exponential term is more stable and reliable. Moreover, if direct fitting is performed on a single battery, it may overfit its individual characteristics; while the embodiment of this application adopts the minimum value of the average fitting error, which can improve the general applicability of the parameter term to batteries of the same type as the sample battery and avoid overfitting.

[0105] In the above embodiment, a candidate value set of the exponential term is obtained; in each round of iteration, a candidate value in the candidate value set is substituted into the initial decay formula to obtain an intermediate decay formula, and multiple actual curves belonging to the same design type are respectively fitted with the curve corresponding to the intermediate decay formula to obtain an average fitting error; the target value of the exponential term is determined according to the average fitting errors of multiple rounds of iteration. In the technical solution of the embodiment of this application, the target value of the exponential term is determined through iteration, which can avoid overfitting and ensure that the target value of the exponential term is more stable and reliable.

[0106] According to some embodiments of the present application, referring to Figure 6 , the training process of the feature prediction model may include the following steps:

[0107] Step 601: Obtain the design data, manufacturing process data, test and characterization data of multiple sample batteries, and the target values of the parameter terms obtained by fitting according to the actual curves of the health states of the multiple sample batteries.

[0108] The server can obtain the design data, manufacturing process data, test and characterization data of multiple sample batteries from the data storage system; and according to the method of the above embodiments, fit according to the actual curves of the health states of the multiple sample batteries to obtain the target values of the parameter terms in the health state decay relational expression.

[0109] Step 602: Use the design data, manufacturing process data, test and characterization data of the sample batteries as training samples, and use the target values of the parameter terms as labels to train the model to obtain a feature prediction model.

[0110] The overall process of training the feature prediction model includes steps such as model construction, model training, hyperparameter optimization, and statistical model error.

[0111] For model construction, multiple applicable machine learning models can be built first, and the model structure can be adjusted to make each machine learning model applicable to the design data, manufacturing process data, test and characterization data as training samples, and the target values of the parameter terms as labels.

[0112] For model training, an applicable training method should be selected according to the machine learning model.

[0113] For hyperparameter optimization, the main hyperparameters of the machine learning model should be adjusted. Hyperparameter automatic optimization methods such as Bayesian optimization can be used to make the model achieve the best effect on the current training samples and labels. Among them, hyperparameters are parameters that need to be manually set before training the machine learning model. Hyperparameters will not be adjusted through the learning of the optimization algorithm during the training process, but are determined by developers or researchers according to experience, experiments, or automatic tuning methods. Bayesian hyperparameter optimization is a method used in the field of machine learning to find the optimal combination of hyperparameters.

[0114] For statistical model error, after obtaining the model output, the average difference between the model output and the corresponding label should be calculated as the model estimation error. Finally, select the model with the smallest estimation error as the feature prediction model.

[0115] In the above embodiments, design data, process data, test and characterization data of multiple sample batteries are obtained, as well as the target values of parameter items obtained by fitting according to the actual curves of the health states of the multiple sample batteries; the design data, process data, test and characterization data are used as training samples, and the target values of the parameter items are used as labels for model training to obtain a feature prediction model. In the technical solution of the embodiments of the present application, the target values of the parameter values are the key features of the health state curve. Directly modeling the key features can focus on the core degradation mechanism, reduce the model complexity, and at the same time, the interpretability of the features provides a physical basis for subsequent analysis. Moreover, taking the key features as the direct output target of the machine learning model can weaken the influence of noise and the overfitting problem of machine learning, and improve the accuracy and stability of model training. Further, the trained model only needs to predict a small number of key features (such as 2-3 parameters), rather than the entire SOH curve, reducing the computational burden of real-time prediction.

[0116] According to some embodiments of the present application, with reference to Figure 7 , in the above embodiments, "obtaining the design data, process data, test and characterization data of the target battery to be predicted" may include the following steps:

[0117] Step 701, obtain the design information, process information, and test and characterization information of the target battery.

[0118] Among them, the design information includes component dimensions, cathode material and ratio, anode material and ratio, electrolyte material and ratio, separator selection, binder material and ratio, etc.; the process information includes first efficiency, compaction density, substrate thickness, coating weight, cold pressing density, cold pressing thickness, active film area size, film sheet resistance, electrode adhesion, OH, hot pressing pressure / time, etc.; the test and characterization information includes voltage, current, capacity, temperature of charge and discharge tests, electrochemical impedance spectroscopy data, X-ray diffraction data, scanning electron microscopy data, etc. The test and characterization information should include the SOH sequence data of the battery.

[0119] The server can obtain the design information, process information, and test and characterization information of the target battery from the terminal or from the data storage system.

[0120] Step 702, perform data preprocessing on the design information, process information, and test and characterization information to obtain the design data, process data, and test and characterization data of the target battery.

[0121] Among them, data preprocessing may include one or more of data cleaning, data transformation, feature engineering, categorical variable transformation, and data splitting. Data cleaning can remove abnormal data and perform data filling, etc. Data transformation can perform format conversion and normalization processing on data, etc.; Feature engineering can calculate new information from known information; Categorical variable transformation can perform conversion of data dimensions and physical units, and data splitting can group data according to a preset time window.

[0122] It should be noted that the data preprocessing method is not limited to the above examples and can be set according to the actual situation.

[0123] The server performs data preprocessing on the obtained design information, process information, and test and characterization information to obtain the design data, process data, and test and characterization data of the target battery.

[0124] In the above embodiments, the design information, process information, and test and characterization information of the target battery are obtained; data preprocessing is performed on the design information, process information, and test and characterization information to obtain the design data, process data, and test and characterization data of the target battery. In the technical solution of the embodiments of the present application, the impacts of factors such as design, process, and test on the health state are fully considered, enabling the prediction of SOH for the battery under research. Moreover, considering physical and chemical parameters, process adjustments, etc. in the prediction is beneficial to improving the prediction accuracy of the health state.

[0125] According to some embodiments of the present application, a method for predicting the health state of a battery is provided. Taking the computer device in which this method is applied as an example, it may include the following steps: Figure 1 is described as an example, and may include the following steps:

[0126] Step 1, obtain the actual curves of the health states of multiple sample batteries; among them, the multiple sample batteries belong to the same design type.

[0127] Step 2, obtain a set of candidate values for the exponential term.

[0128] Step 3, in each round of iteration, substitute a candidate value in the set of candidate values into the initial decay formula to obtain an intermediate decay formula, and respectively perform fitting processing on the actual curves of multiple batteries belonging to the same design type and the curve corresponding to the intermediate decay formula to obtain the average fitting error.

[0129] Step 4, determine the target value of the exponential term according to the average fitting error of multiple rounds of iteration.

[0130] Step 5, substitute the target values of each parameter term into the initial decay formula to obtain a health state decay relationship formula.

[0131] Step 6: Obtain the design data, manufacturing process data, test and characterization data of multiple sample batteries, and the target values of the parameter terms obtained by fitting according to the actual curves of the health states of the multiple sample batteries.

[0132] Step 7: Use the design data, manufacturing process data, test and characterization data as training samples, and use the target values of the parameter terms as labels for model training to obtain a feature prediction model.

[0133] Step 8: Obtain the design information, manufacturing process information, and test and characterization information of the target battery.

[0134] Step 9: Perform data preprocessing on the design information, manufacturing process information, and test and characterization information to obtain the design data, manufacturing process data, and test and characterization data of the target battery.

[0135] Step 10: Input the design data, manufacturing process data, and test and characterization data of the target battery into the pre-trained feature prediction model for prediction processing to obtain at least one key feature of the health state curve of the target battery.

[0136] Step 11: Substitute at least one key feature as a parameter into the health state degradation relationship formula; calculate the health state corresponding to different usage durations according to the health state degradation relationship formula after substituting the parameter to obtain the health state curve corresponding to the target battery.

[0137] For the technical solution provided by the embodiment of the present application, the feature prediction model can predict the health state of the battery based on the design data, manufacturing process data, and test and characterization data, fully considering the influences in aspects such as materials and processes; moreover, the feature prediction model focuses on the core degradation mechanism and outputs key features related to the change trend of the health state, which is beneficial to improving the prediction accuracy of the health state, thereby accelerating the R & D speed of the battery.

[0138] It should be understood that although the steps in the above flowchart are shown in sequence according to the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the above flowchart may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same moment, but can be executed at different moments. The execution order of these steps or stages does not necessarily need to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0139] Based on the same inventive concept, an embodiment of the present application further provides a prediction device for the battery health state for implementing the prediction method of the battery health state involved above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the prediction device for the battery health state provided below can refer to the limitations on the prediction method of the battery health state in the above text, and will not be repeated here.

[0140] According to some embodiments of the present application, referring to Figure 8 , a prediction device for the battery health state is provided, and the device includes:

[0141] A data acquisition module 801, configured to acquire design data, manufacturing process data, and test and characterization data of a target battery to be predicted;

[0142] A feature prediction module 802, configured to input the design data, manufacturing process data, and test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the health state curve of the target battery;

[0143] A health state determination module 803, configured to determine the health state curve corresponding to the target battery according to a pre-constructed health state decay relational expression and at least one key feature; wherein, the health state decay relational expression is used to characterize the correlation between the health state and the usage duration.

[0144] In some embodiments, the health state determination module 803 is specifically configured to substitute at least one key feature as a parameter into the health state decay relational expression; calculate the health state corresponding to different usage durations according to the health state decay relational expression after substituting the parameter to obtain the health state curve corresponding to the target battery.

[0145] In some embodiments, referring to Figure 9 , the device further includes:

[0146] An actual curve acquisition module 804, configured to acquire the actual curves of the health states of multiple sample batteries, and determine the actual curves belonging to the same design type according to the design data of each sample battery;

[0147] A curve fitting module 805, configured to perform curve fitting on multiple actual curves belonging to the same design type and an initial decay formula, and determine the target values of each parameter item in the initial decay formula according to the fitting situation;

[0148] A relational expression determination module 806, configured to substitute the target values of each parameter item into the initial decay formula to obtain the health state decay relational expression.

[0149] In some embodiments, the parameter terms in the health state degradation relation include exponential terms. The curve fitting module 805 is specifically configured to obtain a set of candidate values for the exponential terms. In each iteration process, a candidate value in the set of candidate values is substituted into the initial degradation formula to obtain an intermediate degradation formula. The actual curves of multiple batteries belonging to the same design type are respectively fitted with the curve corresponding to the intermediate degradation formula to obtain an average fitting error. The target value of the exponential term is determined according to the average fitting errors of multiple iterations.

[0150] In some embodiments, the health state degradation relation includes one of Arrhenius' exponential formula, the battery degradation power function formula, and the inflection point fitting formula.

[0151] The parameter terms in Arrhenius' exponential formula include the battery degradation rate and a constant term.

[0152] The parameter terms in the battery degradation power function formula include the coefficients of the square root, linear, and second-order power terms, as well as a constant term.

[0153] The parameter terms in the inflection point fitting formula include the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point.

[0154] In some embodiments, referring to Figure 10 , the device further includes:

[0155] A sample acquisition module 807, configured to acquire the design data, manufacturing process data, test and characterization data of multiple sample batteries, and the target values of the parameter terms obtained by fitting according to the actual curves of the health states of the multiple sample batteries.

[0156] A model training module 808, configured to use the design data, manufacturing process data, test and characterization data as training samples, and use the target values of the parameter terms as labels for model training to obtain a feature prediction model.

[0157] In some embodiments, the data acquisition module 801 is specifically configured to acquire the design information, manufacturing process information, and test and characterization information of the target battery; perform data preprocessing on the design information, manufacturing process information, and test and characterization information to obtain the design data, manufacturing process data, and test and characterization data of the target battery.

[0158] Each module in the above battery health state prediction device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the electronic device in hardware form or be independent of it, or can be stored in the memory of the electronic device in software form, so as to facilitate the processor to call and execute the operations corresponding to the above respective modules.

[0159] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structural diagram may be as shown in Figure 11 . The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store prediction data of the battery health state. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for predicting the battery health state.

[0160] Those skilled in the art can understand that Figure 11 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0161] According to some embodiments of the present application, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory including instructions. The above instructions can be executed by the processor of an electronic device to complete the above method. For example, the non-transitory computer-readable storage medium may be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0162] According to some embodiments of the present application, a computer program product is also provided. When the computer program is executed by the processor, the above method can be implemented. The computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, the above method can be partially or fully implemented according to the process or function of the embodiments of the present application.

[0163] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above various methods. Among them, any reference to a memory, database, or other medium used in the various embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memories can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the various embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the various embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0164] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope described in this specification.

[0165] The above embodiments only illustrate several implementation manners of the present application, which are convenient for understanding the technical solutions of the present application specifically and in detail. However, it should not be construed as a limitation on the protection scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. It should be understood that the technical solutions obtained by those skilled in the art through logical analysis, reasoning or limited experiments based on the technical solutions provided by the present application are all within the protection scope of the appended claims of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the content of the appended claims, and the description and the drawings can be used to explain the content of the claims.

Claims

1. A method for predicting battery health status, characterized in that: The method comprises: Obtain design data, process data, test and characterization data of the target battery to be predicted; Inputting the design data, process data, test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing to obtain at least one key feature of the health state curve of the target battery; wherein the key feature is used to characterize the representative features of the health state curve during the change process; The health state curve corresponding to the target battery is determined according to a pre-constructed health state decay relationship and at least one of the key features; wherein the health state decay relationship is used to characterize the correlation between the health state and the usage time.

2. The method according to claim 1, characterized in that The determining, according to the pre-constructed health state decay relation and at least one of the key features, a health state curve corresponding to the target battery includes: Substituting at least one of the key features as a parameter into the health state decay relationship; The health states corresponding to different usage times are calculated according to the health state decay relationship after substituting the parameters, and the health state curve corresponding to the target battery is obtained.

3. The method according to claim 1 or 2, characterized in that: The method further comprises: Acquire actual curves of health status of multiple sample batteries, and determine actual curves belonging to the same design type according to design data of each of the sample batteries; Performing curve fitting on the multiple actual curves belonging to the same design type and the initial decay formula, and determining the target value of each parameter item in the initial decay formula according to the fitting situation; Substituting the target value of each parameter item into the initial decay formula, the health state decay relational expression is obtained.

4. The method according to claim 3, characterized in that: The parameter items in the health state decay relationship include exponential items, and the curve fitting of the multiple actual curves belonging to the same design type with the initial decay formula, and determining the target value of each parameter item in the initial decay formula according to the fitting situation, includes: Obtaining a set of candidate values ​​for the index item; In each round of iteration, a candidate value in the candidate value set is substituted into the initial decay formula to obtain an intermediate decay formula, and a plurality of actual curves belonging to the same design type are respectively fitted with the curve corresponding to the intermediate decay formula to obtain an average fitting error; The target value of the exponential term is determined according to the average fitting error of multiple rounds of iterations.

5. The method according to claim 4, characterized in that The health state decay relationship includes one of the Arrhenius exponential formula, the battery decay power function formula and the inflection point fitting formula; The parameter terms in the Arrhenius exponential formula include the battery degradation rate and the constant term; The parameter terms in the battery degradation power function formula include square root, linear, second-order power coefficient terms and constant terms; The parameters in the inflection point fitting formula include the abscissa of the slope change point, the starting slope, the ordinate of the starting time, and the slope difference before and after the inflection point.

6. The method according to claim 3, characterized in that The method further comprises: Acquire design data, process data, test and characterization data of the plurality of sample batteries, and target values ​​of parameter items obtained by fitting actual curves of health status of the plurality of sample batteries; The design data, process data, test and characterization data of the sample battery are used as training samples, and the target values ​​of the parameter items are used as labels for model training to obtain the feature prediction model.

7. The method according to claim 1, characterized in that The obtaining of design data, process data, test and characterization data of the target battery to be predicted includes: Obtaining design information, process information, and testing and characterization information of the target battery; The design information, the process information and the test and characterization information are preprocessed to obtain the design data, process data, test and characterization data of the target battery.

8. A battery health status prediction device, characterized in that: The device comprises: A data acquisition module, used to acquire design data, process data, test and characterization data of the target battery to be predicted; A feature prediction module, used to input the design data, process data, test and characterization data of the target battery into a pre-trained feature prediction model for prediction processing, and obtain at least one key feature of the health state curve of the target battery; wherein the key feature is used to characterize the representative features of the health state curve during the change process; A health status determination module is used to determine the health status curve corresponding to the target battery based on a pre-constructed health status decay relationship and at least one of the key features; wherein the health status decay relationship is used to characterize the correlation between the health status and the usage time.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When a processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images