A multi-objective optimization lithium-ion battery charging design method, system and device

Abstract

This application discloses a multi-objective optimization method, system, and device for lithium-ion battery charging design. The method constructs an electrochemical-thermal-aging model and an enhanced Chebyshev function. Based on a first database and the enhanced Chebyshev function, each single-objective subproblem corresponding to each first charging protocol is scalarized to obtain a scalarized target dataset. Based on a second database and the enhanced Chebyshev function, each single-objective subproblem corresponding to each second charging protocol is scalarized to obtain a scalarized source dataset. Based on the scalarized target dataset, a Gaussian process model for the single-objective subproblems is constructed. Based on the scalarized source dataset, a migration acquisition function is designed using the Gaussian process model, the migration acquisition function is solved, and the solution is evaluated using the electrochemical-thermal-aging model to determine a set of target charging protocols. This application can improve the accuracy of charging protocols and reduce model simulation time.

Description

Technical Field

[0001] The present application relates to the field of battery charging technology, and in particular to a multi-objective optimized lithium-ion battery charging design method, system, and device. Background Art

[0002] Lithium-ion batteries, currently the most popular energy storage device, are widely used in a wide range of fields, including electric vehicles, portable electronic devices, and clean energy storage, due to their high energy density, high power density, long cycle life, excellent safety, and relative environmental friendliness. However, long charging times due to slow charging speeds and battery performance degradation caused by rapid charging remain key issues hindering the further promotion and development of lithium-ion batteries.

[0003] To accurately assess the impact of different charging protocols on battery performance, researchers generally rely on coupled electrochemical-thermal-aging models. These models, based on pseudo-two-dimensional (P2D) electrochemical equations, heat conduction equations, and solid electrolyte interface (SEI) nucleation and growth kinetics, accurately reproduce the internal ion diffusion, electrode reactions, and heat source distribution of the battery by solving a system of partial differential equations. However, simulations of full-order models are extremely time-consuming, making them unsuitable for online optimization and large-scale sample screening. Therefore, equivalent circuit models (ECMs) and reduced-order models (ROMs) have been proposed as alternatives to the highly complex P2D models. ECMs use a series of RC networks to approximate battery polarization and transport processes, enabling real-time estimation of terminal voltage and internal impedance. However, their characterization of temperature fields and aging mechanisms is relatively crude. ROMs (such as the single-particle model (SPM) and its coupled thermal-aging extension) simplify and make assumptions about key modes of the P2D model, reducing computational complexity while retaining some physical interpretability. However, they can still suffer from accuracy errors at high rates, low temperatures, or under inhomogeneous boundary conditions.

[0004] Therefore, the charging protocol accuracy in existing related battery charging protocol design methods is relatively low, and the model simulation is time-consuming. Summary of the Invention

[0005] This application aims to propose a multi-objective optimization lithium-ion battery charging design method, system and equipment, which can improve the accuracy of the charging protocol and reduce the model simulation time.

[0006] In a first aspect, an embodiment of the present application provides a multi-objective optimization lithium-ion battery charging design method, the method comprising:

[0007] Constructing a constrained multi-objective optimization problem comprising a plurality of single-objective sub-problems and constraints, wherein each single-objective sub-problem satisfies the constraints;

[0008] Using historical data corresponding to the current battery health state, a first database is constructed containing multiple first charging protocols and total target values ​​corresponding to the first charging protocols. Furthermore, using empirical data corresponding to other battery health states, a second database is constructed containing multiple second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is the total optimization objective in the constrained multi-objective optimization problem corresponding to the charging protocols.

[0009] Constructing electrochemical-thermal-aging models and enhanced Chebyshev functions;

[0010] scalarizing the constrained multi-objective optimization problem corresponding to each of the first charging protocols based on the first database and the enhanced Chebyshev function to obtain a scalarized target data set;

[0011] scalarizing the constrained multi-objective optimization problem corresponding to each second charging protocol based on the second database and the enhanced Chebyshev function to obtain a scalarized source data set;

[0012] Based on the scalarized target data set, a Gaussian process model of a single target subproblem is constructed;

[0013] Based on the scalarized source data set, the Gaussian process model is used to design a migration acquisition function, the migration acquisition function is solved, and the solution is evaluated using the electrochemical-thermal-aging model to determine a set of target charging protocols.

[0014] Compared with the prior art, the first aspect of the present application has the following beneficial effects:

[0015] The method constructs a constrained multi-objective optimization problem including multiple single-objective sub-problems and constraints, wherein each single-objective sub-problem satisfies the constraints; uses historical data corresponding to the current battery health state to construct a first database including multiple first charging protocols and total target values ​​corresponding to the first charging protocols, and uses empirical data corresponding to other battery health states to construct a second database including multiple second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is the total optimization target in the constrained multi-objective optimization problem corresponding to the charging protocol; constructs an electrochemical-thermal-aging model, and constructs an enhanced Chebyshev model. function; based on the first database and the enhanced Chebyshev function, the constrained multi-objective optimization problem corresponding to each first charging protocol is scalarized to obtain a scalarized target data set; based on the second database and the enhanced Chebyshev function, the constrained multi-objective optimization problem corresponding to each second charging protocol is scalarized to obtain a scalarized source data set; based on the scalarized target data set, a Gaussian process model of a single-objective subproblem is constructed; based on the scalarized source data set, a migration acquisition function is designed using the Gaussian process model, the migration acquisition function is solved, and the solution results are evaluated using the electrochemical-thermal-aging model to determine a set of target charging protocols. In this way, the charging protocol is scalarized by adopting the enhanced Chebyshev function, and a Gaussian process model of a single-target subproblem is constructed based on the scalarized target data set. The Gaussian process model is trained with limited real simulation results to achieve rapid prediction and uncertainty assessment of the scalarized target of the charging protocol, which can greatly reduce expensive simulation calls, that is, reduce the model simulation time; then the Gaussian process model is used to design the migration acquisition function, which can effectively combine historical data and empirical data, dynamically adjust the optimization direction, avoid repeated exploration, and quickly generate a new charging protocol, which improves the accuracy of the charging protocol and improves the efficiency of charging protocol design.

[0016] In some embodiments, the electrochemical-thermal-aging model includes an electrochemical model, a thermal model, and an aging model, and constructing the electrochemical-thermal-aging model includes:

[0017] Build an electrochemical model:

[0018] ;

[0019] Build the thermal model:

[0020] ;

[0021] Building an aging model:

[0022] ;

[0023] in, express The terminal voltage at the moment, represents the solid phase potential, represents the spatial coordinates, Indicates the length of the negative electrode, represents the length of the diaphragm, Indicates the length of the positive electrode, Indicates the battery density, represents the specific heat capacity, Indicates temperature, represents thermal conductivity, represents the heat generation rate determined by the electrochemical model, Indicates lithium ion loss, represents the reaction current formed by the solid electrolyte interface layer, represents the reaction current associated with the continuous formation of the surface solid electrolyte interface, Represents the reaction current of the formation of a new solid electrolyte interface layer on the cracks of graphite particles.

[0024] In some embodiments, constructing an enhanced Chebyshev function includes:

[0025] ;

[0026] in, represents the scalarized target generated by the augmented Chebyshev function, represents the number of single-objective subproblems in the constrained multi-objective optimization problem, Represents the weight vector No. A quantity, represents the protocol parameter vector, represents the first A normalized single-objective subproblem, Indicates a positive weight coefficient.

[0027] In some embodiments, constructing a Gaussian process model for a single-target subproblem based on the scalarized target dataset includes:

[0028] A Gaussian kernel function is used to quantify the correlation between any two charging protocols in the scalarized target data set;

[0029] Constructing a correlation matrix based on the correlation between any two charging protocols;

[0030] Calculating a predicted value and an uncertainty of the predicted value of a charging protocol based on the correlation matrix and the scalarized target data set;

[0031] A Gaussian process model of the single-objective subproblem is constructed according to the predicted value and the uncertainty of the predicted value.

[0032] In some embodiments, calculating the predicted value and the uncertainty of the predicted value of the charging protocol based on the correlation matrix and the scalarized target dataset includes:

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] in, represents the predicted value of the charging protocol, Represents N-dimensional full vector, represents transpose, represents the correlation matrix, represents the inverse transformation of the correlation matrix, represents the target value set in the scalar target dataset, represents an N-dimensional vector, represents the uncertainty of the predicted value, represents N dimensions, Represents the protocol parameter vector.

[0038] In some embodiments, the step of designing a migration acquisition function based on the scalarized source data set using the Gaussian process model, solving the migration acquisition function, and evaluating the solution using the electrochemical-thermal-aging model to determine a set of target charging protocols includes:

[0039] Based on the scalarized source data set, a migration acquisition function is designed using the predicted values ​​and the uncertainty of the predicted values ​​in the Gaussian process model;

[0040] An evolutionary algorithm is used to solve the migration acquisition function to obtain a new charging protocol;

[0041] Adding the new charging protocol to the first database to obtain an updated database;

[0042] Evaluating the new charging protocol using the electrochemical-thermal-aging model to obtain an evaluation result;

[0043] A set of target charging protocols is selected from the updated database according to the evaluation result.

[0044] In some embodiments, designing a migration acquisition function based on the scalarized source data set and using the predicted value and the uncertainty of the predicted value in the Gaussian process model includes:

[0045] ;

[0046] ;

[0047] in, represents the migration acquisition function, represents the predicted value of the charging protocol, represents the balance coefficient, represents the uncertainty of the predicted value, represents the weight term, represents a Gaussian distribution derived from a scalarized source dataset, represents the maximum number of iterations, Indicates the current iteration number, Indicates a preset constant, represents the KL divergence, represents the Gaussian distribution derived from the scalarized target dataset, Represents the protocol parameter vector.

[0048] In a second aspect, an embodiment of the present application further provides a multi-objective optimization lithium-ion battery charging design system, the system comprising:

[0049] A first construction unit is configured to construct a constrained multi-objective optimization problem comprising a plurality of single-objective sub-problems and constraint conditions, wherein each single-objective sub-problem satisfies the constraint conditions;

[0050] a second construction unit, configured to construct, using historical data corresponding to the current battery health state, a first database comprising a plurality of first charging protocols and total target values ​​corresponding to the first charging protocols, and to construct, using empirical data corresponding to other battery health states, a second database comprising a plurality of second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is a total optimization objective in a constrained multi-objective optimization problem corresponding to the charging protocols;

[0051] A third building unit is used to build an electrochemical-thermal-aging model and an enhanced Chebyshev function;

[0052] a first scalarization unit, configured to scalarize the constrained multi-objective optimization problem corresponding to each of the first charging protocols based on the first database and the enhanced Chebyshev function to obtain a scalarized target data set;

[0053] a second scalarization unit, configured to scalarize the constrained multi-objective optimization problem corresponding to each second charging protocol based on the second database and the enhanced Chebyshev function to obtain a scalarized source data set;

[0054] A model building unit, configured to build a Gaussian process model of a single-objective subproblem based on the scalarized target data set;

[0055] A charging protocol determination unit is configured to design a migration acquisition function based on the scalarized source data set using the Gaussian process model, solve the migration acquisition function, evaluate the solution using the electrochemical-thermal-aging model, and determine a set of target charging protocols.

[0056] In a third aspect, an embodiment of the present application further provides an electronic device comprising at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute a multi-objective optimization lithium-ion battery charging design method as described above.

[0057] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the multi-objective optimization lithium-ion battery charging design method as described above.

[0058] It can be understood that the beneficial effects of the above-mentioned second to fourth aspects compared with the relevant technologies are the same as the beneficial effects of the above-mentioned first aspect compared with the relevant technologies. Please refer to the relevant description in the above-mentioned first aspect and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0060] Figure 1 This is a flow chart of an embodiment of a multi-objective optimization lithium-ion battery charging design method provided by the present application;

[0061] Figure 2 This is a schematic diagram of determining a charging protocol in a preferred embodiment of the multi-objective optimization lithium-ion battery charging design method provided by the present application;

[0062] Figure 3 Schematic diagram of a multi-stage constant current protocol in a preferred embodiment of a multi-objective optimized lithium-ion battery charging design method provided by the present application;

[0063] Figure 4 This is a schematic diagram of the electrochemical-thermal-aging model in the best embodiment of the multi-objective optimization lithium-ion battery charging design method provided by the present application;

[0064] Figure 5 This is a schematic diagram of using Bayesian optimization for charging design in a preferred embodiment of the multi-objective optimization lithium-ion battery charging design method provided by the present application;

[0065] Figure 6 1. It is a schematic diagram of the overall framework of the experience-driven multi-objective Bayesian optimization in the best embodiment of the multi-objective optimization lithium-ion battery charging design method provided by the present application;

[0066] Figure 7 1 is a schematic structural diagram of an embodiment of a multi-objective optimization lithium-ion battery charging design system provided by the present application;

[0067] Figure 8 It is a structural diagram of an embodiment of the electronic device provided by this application. DETAILED DESCRIPTION

[0068] The following describes in detail embodiments of the present application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and are not to be construed as limiting the present application.

[0069] In the description of this application, if there is a description of first, second, etc., it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0070] In the description of this application, it should be understood that descriptions involving orientation, such as the orientation or positional relationship indicated by up, down, etc., are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application.

[0071] In the description of this application, it should be noted that, unless otherwise clearly defined, terms such as setting, installing, and connecting should be understood in a broad sense, and technical personnel in the relevant technical field can reasonably determine the specific meaning of the above terms in this application based on the specific content of the technical solution.

[0072] To accurately assess the impact of different charging protocols on battery performance, researchers generally rely on coupled electrochemical-thermal-aging models. These models, based on pseudo-two-dimensional (P2D) electrochemical equations, heat conduction equations, and solid electrolyte interface (SEI) nucleation and growth kinetics, accurately reproduce the internal ion diffusion, electrode reactions, and heat source distribution of the battery by solving a system of partial differential equations. However, simulations of full-order models are extremely time-consuming, making them unsuitable for online optimization and large-scale sample screening. Therefore, equivalent circuit models (ECMs) and reduced-order models (ROMs) have been proposed as alternatives to the highly complex P2D models. ECMs use a series of RC networks to approximate battery polarization and transport processes, enabling real-time estimation of terminal voltage and internal impedance. However, their characterization of temperature fields and aging mechanisms is relatively crude. ROMs (such as the single-particle model (SPM) and its coupled thermal-aging extension) simplify and make assumptions about key modes of the P2D model, reducing computational complexity while retaining some physical interpretability. However, they can still suffer from accuracy errors at high rates, low temperatures, or under inhomogeneous boundary conditions.

[0073] In order to solve the problems of low charging protocol accuracy and time-consuming model simulation in existing related battery charging protocol design methods, this application proposes a multi-objective optimization lithium-ion battery charging design method, system and device.

[0074] Reference Figure 1 , a flow chart of a multi-objective optimization lithium-ion battery charging design method provided in an embodiment of the present application. The multi-objective optimization lithium-ion battery charging design method is applied to an electronic device, which may be a server or a mobile terminal. Figure 1 As shown, the multi-objective optimization lithium-ion battery charging design method may include the following steps:

[0075] Step S100: constructing a constrained multi-objective optimization problem including multiple single-objective sub-problems and constraints, wherein each single-objective sub-problem satisfies the constraints;

[0076] Step S200: Using historical data corresponding to the current battery health state, construct a first database comprising a plurality of first charging protocols and total target values ​​corresponding to the first charging protocols. Furthermore, using empirical data corresponding to other battery health states, construct a second database comprising a plurality of second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is the total optimization objective in the constrained multi-objective optimization problem corresponding to the charging protocols.

[0077] Step S300: constructing an electrochemical-thermal-aging model and an enhanced Chebyshev function;

[0078] Step S400: scalarizing the constrained multi-objective optimization problem corresponding to each first charging protocol based on the first database and the enhanced Chebyshev function to obtain a scalarized target data set;

[0079] Step S500: scalarizing the constrained multi-objective optimization problem corresponding to each second charging protocol based on the second database and the enhanced Chebyshev function to obtain a scalarized source data set;

[0080] Step S600: constructing a Gaussian process model for a single-target subproblem based on the scalarized target data set;

[0081] Step S700: Based on the scalarized source data set, a Gaussian process model is used to design a migration acquisition function, the migration acquisition function is solved, and the solution is evaluated using an electrochemical-thermal-aging model to determine a set of target charging protocols.

[0082] In this embodiment, a constrained multi-objective optimization problem including multiple single-objective sub-problems and constraints is constructed, wherein each single-objective sub-problem satisfies the constraints; historical data corresponding to the current battery health state is used to construct a first database including multiple first charging protocols and total target values ​​corresponding to the first charging protocols, and empirical data corresponding to other battery health states is used to construct a second database including multiple second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is the total optimization target in the constrained multi-objective optimization problem corresponding to the charging protocol; an electrochemical-thermal-aging model is constructed, and an enhanced cutting Chebyshev function; based on the first database and the enhanced Chebyshev function, the constrained multi-objective optimization problem corresponding to each first charging protocol is scalarized to obtain a scalarized target data set; based on the second database and the enhanced Chebyshev function, the constrained multi-objective optimization problem corresponding to each second charging protocol is scalarized to obtain a scalarized source data set; based on the scalarized target data set, a Gaussian process model of the single-objective subproblem is constructed; based on the scalarized source data set, a migration acquisition function is designed using the Gaussian process model, the migration acquisition function is solved, and the solution results are evaluated using the electrochemical-thermal-aging model to determine a set of target charging protocols. In this way, the charging protocol is scalarized by adopting the enhanced Chebyshev function, and a Gaussian process model of a single-target subproblem is constructed based on the scalarized target data set. The Gaussian process model is trained with limited real simulation results to achieve rapid prediction and uncertainty assessment of the scalarized target of the charging protocol, which can greatly reduce expensive simulation calls, that is, reduce the model simulation time; then the Gaussian process model is used to design the migration acquisition function, which can effectively combine historical data and empirical data, dynamically adjust the optimization direction, avoid repeated exploration, and quickly generate a new charging protocol, which improves the accuracy of the charging protocol and improves the efficiency of charging protocol design.

[0083] The experience data corresponding to the other battery health states mentioned above may be experience data of other battery health states except the current battery health state, for example, charging design data of batteries at different health state levels, which represents corresponding experience.

[0084] The electrochemical-thermal-aging model may be a combined model including an electrochemical model, a thermal model, and an aging model.

[0085] In some embodiments, the electrochemical-thermal-aging model includes an electrochemical model, a thermal model, and an aging model. Constructing the electrochemical-thermal-aging model includes:

[0086] Build an electrochemical model:

[0087] ;

[0088] Build the thermal model:

[0089] ;

[0090] Building an aging model:

[0091] ;

[0092] in, express The terminal voltage at the moment, represents the solid phase potential, represents the spatial coordinates, Indicates the length of the negative electrode, represents the length of the diaphragm, Indicates the length of the positive electrode, Indicates the battery density, represents the specific heat capacity, Indicates temperature, represents thermal conductivity, represents the heat generation rate determined by the electrochemical model, Indicates lithium ion loss, represents the reaction current formed by the solid electrolyte interface layer, represents the reaction current associated with the continuous formation of the surface solid electrolyte interface, Represents the reaction current of the formation of a new solid electrolyte interface layer on the cracks of graphite particles.

[0093] In this embodiment, establishing a computationally efficient battery model that accurately describes the internal battery mechanisms effectively addresses the challenges of optimizing battery charging strategies. By combining electrochemical, thermal, and aging models to construct an electrochemical-thermal-aging model (i.e., a battery model), this model accurately simulates a real battery, thereby improving the accuracy of subsequent charging protocol performance assessments.

[0094] In some embodiments, constructing an enhanced Chebyshev function includes:

[0095] ;

[0096] in, represents the scalarized target generated by the augmented Chebyshev function, represents the number of single-objective subproblems in the constrained multi-objective optimization problem, Represents the weight vector No. A quantity, represents the protocol parameter vector, represents the first A normalized single-objective subproblem, Indicates a positive weight coefficient.

[0097] In this embodiment, by constructing an enhanced Chebyshev function, a good foundation is laid for later use of the enhanced Chebyshev function to realize charging protocol scalarization.

[0098] In some embodiments, constructing a Gaussian process model for a single-objective subproblem based on a scalarized target dataset includes:

[0099] The Gaussian kernel function is used to quantify the correlation between any two charging protocols in the scalar target dataset;

[0100] According to the correlation between any two charging protocols, a correlation matrix is ​​constructed;

[0101] Calculate the predicted value and uncertainty of the charging protocol based on the correlation matrix and the scalarized target dataset;

[0102] According to the predicted value and its uncertainty, a Gaussian process model of the single-objective subproblem is constructed.

[0103] In this example, a Gaussian kernel function is used to quantify the correlation between any two charging protocols in a scalarized target dataset. A correlation matrix is ​​constructed based on the correlation between any two charging protocols. The predicted value and uncertainty of the charging protocol are calculated based on the correlation matrix and the scalarized target dataset. Based on the predicted value and uncertainty, a Gaussian process model for the single-target subproblem is constructed. This approach constructs a Gaussian process model for the single-target subproblem based on the scalarized target dataset and trains the Gaussian process model using limited real-world simulation results. This allows for rapid prediction and uncertainty assessment of the scalarized target of the charging protocol, significantly reducing expensive simulation calls and, consequently, model simulation time.

[0104] In some embodiments, calculating a predicted value and an uncertainty of the predicted value of a charging protocol based on a correlation matrix and a scalarized target dataset includes:

[0105] ;

[0106] ;

[0107] ;

[0108] ;

[0109] in, represents the predicted value of the charging protocol, Represents N-dimensional full vector, represents transpose, represents the correlation matrix, represents the inverse transformation of the correlation matrix, represents the target value set in the scalar target dataset, represents an N-dimensional vector, represents the uncertainty of the predicted value, represents N dimensions, Represents the protocol parameter vector.

[0110] In some embodiments, a Gaussian process model is used to design a migration acquisition function based on a scalarized source data set, the migration acquisition function is solved, and the solution is evaluated using an electrochemical-thermal-aging model to determine a set of target charging protocols, including:

[0111] Based on the scalarized source data set, the migration acquisition function is designed using the predicted value and the uncertainty of the predicted value in the Gaussian process model;

[0112] An evolutionary algorithm is used to solve the migration acquisition function and obtain a new charging protocol;

[0113] Adding the new charging protocol to the first database to obtain an updated database;

[0114] The new charging protocol was evaluated using an electrochemical-thermal-aging model, and the evaluation results were obtained.

[0115] Based on the evaluation results, a set of target charging protocols is selected from the updated database.

[0116] In this embodiment, a migration acquisition function is designed based on a scalarized source data set, using the predicted values ​​and their uncertainties in a Gaussian process model. An evolutionary algorithm is used to solve the migration acquisition function to obtain a new charging protocol. The new charging protocol is added to the first database to obtain an updated database. The new charging protocol is evaluated using an electrochemical-thermal-aging model to obtain evaluation results. Based on the evaluation results, a set of target charging protocols is selected from the updated database. This designed migration acquisition function effectively combines historical and empirical data, dynamically adjusts the optimization direction, avoids repeated exploration, and rapidly generates new charging protocols, thereby improving both the accuracy of the charging protocol and the efficiency of its design.

[0117] In some embodiments, based on the scalarized source data set, a migration acquisition function is designed using the predicted values ​​and the uncertainty of the predicted values ​​in the Gaussian process model, including:

[0118] ;

[0119] ;

[0120] in, represents the migration acquisition function, represents the predicted value of the charging protocol, represents the balance coefficient, represents the uncertainty of the predicted value, represents the weight term, represents a Gaussian distribution derived from a scalarized source dataset, represents the maximum number of iterations, Indicates the current iteration number, Indicates a preset constant, represents the KL divergence, represents the Gaussian distribution derived from the scalarized target dataset, Represents the protocol parameter vector.

[0121] In this embodiment, based on the scalarized source data set, the predicted value and the uncertainty of the predicted value in the Gaussian process model are used to design a migration acquisition function. The historical optimization data in the high health state can be modeled as a Gaussian distribution, and its similarity with the current health state (SOH) task is quantified by KL divergence, and the experience weight is dynamically adjusted. , avoiding interference from irrelevant experience (negative transfer). It enables cross-SOH experience reuse and reduces repeated exploration.

[0122] To facilitate understanding by those skilled in the art, a set of best embodiments is provided below:

[0123] As the most popular energy storage device at present, lithium-ion batteries have been widely used in many fields such as electric vehicles, portable electronic devices and clean energy storage due to their high energy density, high power density, long cycle life, good safety and relative environmental friendliness. However, the long charging time caused by slow charging speed and the battery performance degradation caused by fast charging are still key issues hindering the further promotion and development of lithium-ion batteries. Therefore, the charging problem of lithium-ion batteries is an important topic. Under the condition of a given battery capacity, optimizing the battery charging process and designing an efficient charging strategy can effectively reduce the charging time while ensuring the health of the battery, thereby bringing convenience to users of lithium-ion battery products. The design process of the lithium-ion battery charging protocol is essentially a constrained multi-objective optimization problem (CMOP), because the ideal optimal charging strategy must not only achieve a significant reduction in charging time, but also ensure the safety of the charging process, while minimizing the degree of battery aging caused by charging. It can be seen that the optimization of the charging strategy requires balancing the charging time ( ), peak temperature ( ) and lithium ion loss ( ) and other goals while meeting dynamic constraints such as current, voltage, temperature and SOC.

[0124] In order to accurately evaluate the impact of different charging protocols on battery performance, the academic community generally relies on coupled electrochemical-thermal-aging models. This model is based on pseudo-two-dimensional (P2D) electrochemical equations, heat conduction equations, and SEI nucleation and growth kinetics equations. It accurately reproduces the internal ion diffusion, electrode reactions, and heat source distribution of the battery by solving a set of partial differential equations. However, the simulation of the full-order model is extremely time-consuming, which is not conducive to online optimization and large-scale sample screening. For this reason, some people have proposed equivalent circuit models (ECMs) and reduced-order models (ROMs) to replace the highly complex P2D model:

[0125] ECM uses a series of RC networks to approximate the battery polarization and transmission process, and can estimate the terminal voltage and internal impedance in real time, but the characterization of the temperature field and aging mechanism is relatively rough; ROM (such as the single-particle model SPM and its coupled thermal-aging extension) simplifies and assumes the key modes of the P2D model, reducing the amount of calculation while retaining some physical interpretability, but accuracy deviations may still occur under high-rate, low-temperature or non-uniform boundary conditions.

[0126] The existing related technologies have the following shortcomings:

[0127] 1. High-fidelity battery model simulation is expensive: Full-order electrochemical-thermal-aging models require decoupling of partial differential equations. Single simulations are time-consuming and difficult to meet online or real-time optimization requirements.

[0128] 2. Insufficient accuracy of equivalent circuits and reduced-order models: Although ECM and ROM (such as single-particle models) can accelerate calculations, they make it difficult to accurately describe the temperature field and aging mechanism under high-rate fast charging due to simplified assumptions, affecting the reliability of the optimization results.

[0129] 3. Data starvation in multi-objective Bayesian optimization: Existing multi-objective BO methods (such as the framework based on Chebyshev decomposition) require a large number of simulation samples to converge when jointly optimizing the three objectives of charging time, peak temperature, and capacity attenuation, resulting in low optimization efficiency.

[0130] 4. Lack of experience transfer across health states: As batteries age with use, the optimal charging protocol changes with the state of health (SOH). Existing methods require optimization from scratch for each SOH, resulting in serious waste of computing power and response delays.

[0131] In order to solve the problems existing in the existing technology, this embodiment proposes an experience-driven multi-objective Bayesian optimization algorithm (EIMO), which introduces experience migration under the decomposition multi-objective BO framework and uses the optimal protocol data under the historical SOH level to accelerate the optimization under the current SOH; by constructing a Gaussian process model and training the GP model with limited real simulation results, it can achieve rapid prediction and uncertainty assessment of the scalarized target of the charging protocol (enhanced Tchebycheff function), greatly reducing expensive simulation calls; by designing a transfer acquisition function, on the basis of the common lower confidence bound (LCB) acquisition function, an empirical distribution term is introduced , and through adaptive weights Adjust the balance between experience and exploration / exploitation to accelerate the exploration of the optimal protocol. This embodiment designs an adaptive experience weight strategy based on the experience distribution. With current agent distribution KL divergence, dynamic adjustment , ensuring that experience is fully utilized in similar tasks, and that later optimizations gradually rely on real-time data to avoid negative transfer. Ultimately, this embodiment achieves efficient, online, and cross-SOH adaptive charging design, providing new methodological support for intelligent charging management and promoting the development of related industries.

[0132] The technical solution of this embodiment specifically includes the following contents:

[0133] like Figure 2 As shown in Figure 1, the charging design problem determines the charging protocol through optimization techniques that consider the battery state. It is usually formulated as a constrained multi-objective optimization problem (CMOP). Generally speaking, higher charging currents can shorten charging times, but this leads to higher voltage and temperature peaks and a shorter service life. It must simultaneously meet the conflicting requirements of speed, safety, and sustainability. Considering these different requirements, the charging protocol design problem can be formulated as a multi-objective optimization problem. The optimization objectives of the multi-objective optimization problem are defined as follows:

[0134] (1);

[0135] in, is the charging time, representing the goal of rapidity, Is the peak temperature during the charging process, representing the safety target, and These two target values ​​can be directly obtained through the state quantity calculated by battery simulation. It is the battery capacity attenuation caused by the charging process, calculated by the aging model:

[0136] (2);

[0137] in, The amount of power increased by one charge, is the total available cycle charge and discharge capacity under the current strategy, calculated by formula (3).

[0138] (3);

[0139] in, Represents a nonlinear aging function, which is used to describe the influence of relevant battery state quantities on battery capacity attenuation. represents the exponential factor, Indicates the average SOC of a lithium-ion battery during a charging cycle. Indicates the average input current. Indicates the average battery temperature.

[0140] To ensure the safety of the charging process, the following constraints also need to be considered:

[0141] (4);

[0142] in, Indicates the battery temperature, Indicates the terminal voltage, Indicates the battery state of charge (SOC), Indicates the input current. Indicates the start time, represents the initial temperature, Indicates the upper limit of the battery temperature. and Respectively represent the upper and lower limits of the battery state of charge, and Represent the upper and lower limits of the terminal voltage respectively, and The above constraints include the limits on battery voltage and temperature, the limits on input current, and the definition of the SOC range during charging.

[0143] Considering that the multi-stage constant current (MCC) protocol does not have a slow constant voltage stage and a static stage, it has higher charging efficiency. In this embodiment, the MCC protocol is optimized. Figure 3 As shown, the MCC protocol divides the charging process into Each stage uses an independent current to charge the battery. The goal of the MCC protocol is to reduce the battery SOC from the minimum Increase to maximum In the The charging current is , the charge capacity is , charging time is To simplify the representation, a protocol parameter vector can be used To describe the MCC protocol. Generally speaking, increasing the number of stages in the protocol can improve its theoretical optimal performance, but it also introduces a larger parameter space, increases the complexity of the optimization task, and makes the search for the optimal protocol more difficult.

[0144] Establishing a battery model that can accurately describe the internal mechanism of the battery and is computationally efficient can effectively solve the difficulties in optimizing the charging strategy mentioned above. According to different modeling process objects, battery models can be divided into electrochemical models, thermal models, and aging models, such as Figure 4As shown in Figure 2, an electrochemical model is used to describe the electrochemical dynamics within lithium-ion batteries. Based on solution theory, porous electrode theory, and kinetic equations, this pseudo-two-dimensional (P2D) model is known for its ability to accurately simulate real batteries. Specifically, the model consists of a set of partial differential equations and algebraic equations that describe the internal dynamics of the battery. Based on these equations, the terminal voltage is calculated as follows for a given input current:

[0145] (5);

[0146] in, represents the solid phase potential, represents the spatial coordinates, 、 and They represent the length of the negative electrode, the length of the separator and the length of the positive electrode respectively.

[0147] The thermal model is used to represent the temperature distribution inside the lithium-ion battery. This embodiment uses the thermal model widely used in cylindrical batteries:

[0148] (6);

[0149] The model needs to meet the boundary conditions:

[0150] (7);

[0151] (8);

[0152] And the initial condition constraints:

[0153] (9);

[0154] in, are spatial coordinates, represents temperature, Indicates the battery density, is the specific heat capacity, represents thermal conductivity, is the heat generation rate determined by the electrochemical model, is the specific thermal conductivity, represents a non-uniform Robin boundary, is the heat transfer coefficient, is the ambient temperature, is the initial temperature.

[0155] The aging model is used to describe the capacity decay of lithium-ion batteries. This embodiment uses a physical-based aging model that reflects battery degradation by simulating the formation of a solid electrolyte interface (SEI) layer. Specifically, the lithium ion loss (expressed as ) is described as follows:

[0156] (10);

[0157] in, represents the reaction current of SEI formation, is the reaction current associated with the continuous formation of surface SEI, Represents the reaction current of forming new SEI layer on the cracks of graphite particles.

[0158] The above models are combined to form the electrochemical-thermal-aging model. Because it involves solving partial differential equations, each simulation of this model is time-consuming. Therefore, when using this model for charging design, it is crucial to develop suitable optimization techniques to accelerate the design process.

[0159] In constrained multi-objective optimization problems that represent charging design problems, Bayesian optimization has been applied to the design of charging protocols, namely multi-stage constant current (MCC) protocols. Although it can speed up the design process by building an information-rich proxy model with a minimum number of evaluation samples, Bayesian optimization often has the problem of large data requirements when solving CMOP problems. In fact, the experience of other similar scenarios can be used to reduce the dependence on the current task data. For charging design, the experience gained from lithium-ion batteries at different states of health (SOH) can be used to optimize the charging protocol of the battery at the current SOH. For example Figure 5 As shown, the EIMO proposed in this embodiment uses Bayesian optimization for charging design and combines it with transfer learning to combine historical data of the current SOH level with experience of higher SOH levels, further improving efficiency.

[0160] The framework diagram of EIMO is as follows Figure 6As shown in Figure 2. First, a database is constructed by sampling a set of charging protocols and evaluating the corresponding battery states. The charging design problem described by Equations (1) and (4) is decomposed into several single-objective subproblems. In each iteration, a subproblem is randomly selected and a new charging protocol is generated for the subproblem using Bayesian optimization. Specifically, Bayesian optimization consists of two key components: a Gaussian process (GP) model and an acquisition function. The GP model is used to approximate the time-consuming electrochemical-thermal-aging model mentioned above, while the acquisition function is used to sample new charging protocols. The GP model is constructed using historical data in the database, and a transfer learning technique is proposed to integrate historical data with empirical data in the formulation of the acquisition function. A new charging protocol is obtained by solving the acquisition function using an evolutionary algorithm. This charging protocol is then evaluated using the electrochemical-thermal-aging model and added to the database. This process is repeated until the computing resources are exhausted, after which the optimal charging protocol set (i.e., the optimal charging protocol set that meets the user's individual needs) is selected and output from the database.

[0161] The main steps of EIMO include initialization, problem decomposition, GP modeling and migration acquisition function, as follows:

[0162] 1. Initialization: Set the initial number of iterations Set to 0. Next, randomly sample Based on these protocols, a database (i.e., the first database) is constructed and expressed as ,in and Respectively represent The charging protocols of the samples and their corresponding total target values ​​(the total target value is the overall optimization goal, that is, Here, data from other scenarios, such as battery charging design data at different SOH levels, represent the corresponding experience. These data are integrated into an experience database (i.e., the second database) Among them and represent the first charging protocols and their corresponding total target values, represents the total charging protocol for SOH level sampling, represents the set of charging protocols sampled from other state of health (SOH) levels, Represents The target value set corresponding to the charging protocol, Indicates the various stages of the charging process.

[0163] 2. Problem decomposition: EIMO uses a decomposition method to solve the target CMOP problem. In the example, a predefined set of Riesz s-energy methods is used to generate Randomly select a weight vector from , wherein, the Riesz s-energy in this embodiment is a prior art and will not be described in detail in this embodiment. Please refer to the document “Rieszs-energy-based Reference Sets for Multi-Objective optimization”. Using the selected weight vector, based on and the enhanced Chebyshev function, in the target dataset Generate scalarized targets and obtain scalarized target datasets , Represents the total charging protocol sampled in the target dataset as follows:

[0164] (11);

[0165] in, represents the scalarized target generated by the augmented Chebyshev function, represents the protocol parameter vector, is the number of single-objective subproblems in the target CMOP problem, The target CMOP is A normalized single-objective subproblem, yes No. A quantity, is a small positive weight. Note that The normalized formula for a single-objective subproblem is as follows:

[0166] (12);

[0167] in, It is The original objective function corresponding to the single-objective subproblem is: and It is The single-objective subproblem in the dataset (i.e. ). In the same way, based on and enhanced Chebyshev function, which can also generate scalarized source data sets .

[0168] 3. GP modeling: For the selected single-objective sub-problem, a GP model is constructed, denoted as , for dataset-based Approximate Chebyshev scalarization targets to speed up the design process:

[0169] (13);

[0170] (14);

[0171] (15);

[0172] (16);

[0173] in, Represents the scalarized target dataset The target value set after enhanced Chebyshev scalarization processing, represents the predicted value, Indicates the uncertainty of the predicted value, 1 is a dimensional (i.e. N=3M) all-1 vector, is an (N×N) matrix (i.e., correlation matrix), whose elements ), Represents the target dataset Middle A sampled charging protocol, , is a N -dimensional vector whose elements ).here, Represents the correlation (i.e., correlation) between two vectors, quantified by the Gaussian kernel function:

[0174] (17);

[0175] in, is a weight parameter, and They are and No. Dimensional components.

[0176] 4. Migration acquisition function: In traditional Bayesian optimization, the GP model is used to design a migration acquisition function to sample a new decision vector that can effectively balance utilization and exploration. In order to effectively combine data with experience, EIMO introduces a migration acquisition function (expressed as ), which combines the :

[0177] (18);

[0178] As shown in formula (18), It consists of three key parts: Lower Confidence Bound (LCB) , items representing experience and weight terms .

[0179] LCB is a commonly used sampling criterion in traditional Bayesian optimization and is known for its ability to effectively balance exploitation and exploration. Specifically, minimizing It helps to exploit in promising areas with good target values ​​while minimizing ( ) encourages exploration in decision spaces with high uncertainty. , a balance between exploitation and exploration can be achieved. It is from First, from Select The optimal decision vector is then used to calculate the mean using the maximum likelihood estimation method. and covariance matrix Intuitively speaking, Decision vectors with larger values ​​may perform well in scenarios where the battery charging protocol under different SOH levels is optimized. Since these scenarios are similar to the current task, minimizing ( ) can use the experience of these scenarios to accelerate the design process.

[0180] Weight The influence of control experience on the optimization process. Adjust There are two reasons: First, when and When the two models are very similar, ensure that experience has a greater influence; second, as the optimization proceeds and new data continues to increase, the influence of experience should gradually weaken to reduce the risk of negative transfer.

[0181] Based on the above considerations, The following adaptive adjustment method is used:

[0182] (19);

[0183] in, is the Kullback-Leibler (KL) divergence, which is used to quantify and The similarities between and The derivation method is the same as that from , It is from The Gaussian distribution derived from is the maximum number of iterations, is a preset constant. As shown in formula (19), Affected by KL divergence and number of iterations The addition of KL divergence is intended to and When high similarity is shown, EIMO is encouraged to transfer more experience. The smaller the KL divergence, and The higher the similarity between them. In addition, This is to guide EIMO to rely primarily on experience in the early stages of optimization, and gradually reduce the influence of experience as the optimization progresses. This adjustment strategy is consistent with the above analysis.

[0184] Using evolutionary algorithms to solve the migration acquisition function, a new protocol (expressed as Then, the electrochemical-thermal-aging model was used to Evaluate and append it to Repeat the above process until the computing resources are exhausted (i.e. ).

[0185] The evolutionary algorithms described above form a "family of algorithms." Despite their numerous variations, including different genetic expression methods, crossover and mutation operators, the use of specialized operators, and diverse regeneration and selection methods, they all draw inspiration from biological evolution in nature. Compared to traditional optimization algorithms such as calculus-based methods and exhaustive methods, evolutionary computation is a mature, highly robust, and widely applicable global optimization method. Its self-organizing, self-adaptive, and self-learning properties allow it to effectively address complex problems that are difficult for traditional optimization algorithms to solve, regardless of the nature of the problem.

[0186] Compared with the existing technology, the technical solution of this embodiment has the following advantages:

[0187] 1. This embodiment uses the transfer learning technology driven by distribution similarity to model the historical optimization data under high health status as a Gaussian distribution , quantify its similarity with the current SOH task through KL divergence, and dynamically adjust the experience weight ( ), avoiding interference from irrelevant experience (negative transfer). This embodiment reuses experience across SOHs, significantly reducing the number of simulations. This embodiment integrates high SOH historical data (such as optimization experience with SOH = 1.0) through transfer learning, dynamically adjusting the optimization direction and avoiding repeated exploration.

[0188] 2. This embodiment adopts the design of migration acquisition function that integrates experience. On the basis of traditional Bayesian optimization LCB, the experience term ( ), forming a migration acquisition function. And through LCB balanced exploration and development, the empirical term guides the search towards the historical optimal solution area, accelerating convergence to the high-quality Pareto frontier, and improving the convergence and distribution uniformity of the solution set.

[0189] 3. This embodiment employs a decomposition-based multi-objective optimization and evolutionary solution framework, using the Riesz s-energy method to generate a uniformly distributed set of weight vectors to cover diverse user requirements (such as fast charging, low temperature, and low attenuation), avoiding the sparsity issues of traditional uniform sampling. Furthermore, this embodiment incorporates a differential evolution algorithm to solve high-dimensional, non-convex transfer acquisition functions, supporting multi-stage constant current charging (MCC) protocol optimization and reducing computation time by 23%-41%.

[0190] 4. This embodiment dynamically adapts to battery aging and supports full life cycle management. The similarity between experience and current SOH is quantified by KL divergence, and the migration weight is dynamically adjusted ( ), avoiding negative migration without having to optimize from scratch for each SOH, and quickly generating new protocols using historical experience.

[0191] Reference Figure 7 The embodiment of the present application further provides a multi-objective optimization lithium-ion battery charging design system, which includes a first construction unit 100, a second construction unit 200, a third construction unit 300, a first scalarization unit 400, a second scalarization unit 500, a model construction unit 600, and a charging protocol determination unit 700, wherein:

[0192] A first construction unit 100 is configured to construct a constrained multi-objective optimization problem comprising a plurality of single-objective sub-problems and constraints, wherein each single-objective sub-problem satisfies the constraints;

[0193] A second construction unit 200 is configured to construct a first database comprising a plurality of first charging protocols and total target values ​​corresponding to the first charging protocols using historical data corresponding to the current battery health state, and to construct a second database comprising a plurality of second charging protocols and total target values ​​corresponding to the second charging protocols using empirical data corresponding to other battery health states, wherein the total target value is the total optimization objective in the constrained multi-objective optimization problem corresponding to the charging protocol;

[0194] The third construction unit 300 is used to construct an electrochemical-thermal-aging model and an enhanced Chebyshev function;

[0195] A first scalarization unit 400 is configured to scalarize the constrained multi-objective optimization problem corresponding to each first charging protocol based on the first database and the enhanced Chebyshev function to obtain a scalarized target data set;

[0196] A second scalarization unit 500 is configured to scalarize the constrained multi-objective optimization problem corresponding to each second charging protocol based on the second database and the enhanced Chebyshev function to obtain a scalarized source data set;

[0197] A model building unit 600 is used to build a Gaussian process model of a single-objective subproblem based on a scalarized target data set;

[0198] The charging protocol determination unit 700 is used to design a migration acquisition function based on a scalarized source data set using a Gaussian process model, solve the migration acquisition function, and evaluate the solution using an electrochemical-thermal-aging model to determine a set of target charging protocols.

[0199] It should be noted that, since the multi-objective optimization lithium-ion battery charging design system in this embodiment and the above-mentioned multi-objective optimization lithium-ion battery charging design method are based on the same inventive concept, the corresponding contents in the method embodiment are also applicable to the present system embodiment and will not be described in detail here.

[0200] Reference Figure 8 , an embodiment of the present application further provides an electronic device, the electronic device comprising:

[0201] at least one memory;

[0202] at least one processor;

[0203] at least one program;

[0204] The programs are stored in the memory, and the processor executes at least one program to implement the multi-objective optimization lithium-ion battery charging design method disclosed in the present invention.

[0205] The electronic device may be any intelligent terminal including a mobile phone, a tablet computer, a personal digital assistant (PDA), a car computer, etc.

[0206] The electronic device according to the embodiment of the present application is described in detail below.

[0207] The processor 1600 may be implemented as a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is configured to execute relevant programs to implement the technical solutions provided by the embodiments of the present disclosure.

[0208] The memory 1700 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1700 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1700 and is called by the processor 1600 to execute the multi-objective optimization lithium-ion battery charging design method of the embodiment of the present disclosure.

[0209] Input / output interface 1800, used for information input and output;

[0210] Communication interface 1900, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WiFi, Bluetooth, etc.);

[0211] Bus 2000 , which transmits information between various components of the device (e.g., processor 1600 , memory 1700 , input / output interface 1800 , and communication interface 1900 );

[0212] The processor 1600 , the memory 1700 , the input / output interface 1800 , and the communication interface 1900 are connected to each other in communication within the device via the bus 2000 .

[0213] An embodiment of the present disclosure also provides a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions, which are used to enable a computer to execute the above-mentioned multi-objective optimization lithium-ion battery charging design method.

[0214] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0215] The embodiments described in the embodiments of the present disclosure are intended to more clearly illustrate the technical solutions of the embodiments of the present disclosure and do not constitute a limitation on the technical solutions provided by the embodiments of the present disclosure. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present disclosure are also applicable to similar technical problems.

[0216] Those skilled in the art will understand that the technical solutions shown in the drawings do not constitute a limitation on the embodiments of the present disclosure, and may include more or fewer steps than shown in the drawings, or a combination of certain steps, or different steps.

[0217] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0218] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0219] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0220] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0221] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0222] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0223] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0224] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions for enabling an electronic device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk. The embodiments of the present application are described in detail above in conjunction with the accompanying drawings, but the present application is not limited to the above embodiments. Various changes can be made within the scope of knowledge possessed by ordinary technicians in the relevant technical field without departing from the purpose of the present application.

[0225] The embodiments of the present application are described in detail above in conjunction with the accompanying drawings, but the present application is not limited to the above embodiments. Various changes can be made within the scope of knowledge possessed by ordinary technicians in the relevant technical field without departing from the purpose of the present application.

Claims

1. A multi-objective optimization lithium-ion battery charging design method, characterized in that: The method comprises: Constructing a constrained multi-objective optimization problem comprising a plurality of single-objective sub-problems and constraints, wherein each single-objective sub-problem satisfies the constraints; Using historical data corresponding to the current battery health state, a first database is constructed containing multiple first charging protocols and total target values ​​corresponding to the first charging protocols. Furthermore, using empirical data corresponding to other battery health states, a second database is constructed containing multiple second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is the total optimization objective in the constrained multi-objective optimization problem corresponding to the charging protocols. Constructing electrochemical-thermal-aging models and enhanced Chebyshev functions; scalarizing the constrained multi-objective optimization problem corresponding to each of the first charging protocols based on the first database and the enhanced Chebyshev function to obtain a scalarized target data set; scalarizing the constrained multi-objective optimization problem corresponding to each second charging protocol based on the second database and the enhanced Chebyshev function to obtain a scalarized source data set; Based on the scalarized target data set, a Gaussian process model of a single target subproblem is constructed; Based on the scalarized source data set, a migration acquisition function is designed using the Gaussian process model, the migration acquisition function is solved, and the solution is evaluated using the electrochemical-thermal-aging model to determine a set of target charging protocols, specifically: Based on the scalarized source data set, using the predicted value and the uncertainty of the predicted value in the Gaussian process model, a migration acquisition function is designed, including: ； ； in, represents the migration acquisition function, represents the predicted value of the charging protocol, represents the balance coefficient, represents the uncertainty of the predicted value, represents the weight term, represents a Gaussian distribution derived from a scalarized source dataset, represents the maximum number of iterations, Indicates the current iteration number, Indicates a preset constant, represents the KL divergence, represents the Gaussian distribution derived from the scalarized target dataset, represents the protocol parameter vector; An evolutionary algorithm is used to solve the migration acquisition function to obtain a new charging protocol; Adding the new charging protocol to the first database to obtain an updated database; Evaluating the new charging protocol using the electrochemical-thermal-aging model to obtain an evaluation result; A set of target charging protocols is selected from the updated database according to the evaluation result.

2. The multi-objective optimization lithium-ion battery charging design method according to claim 1, characterized in that: The electrochemical-thermal-aging model includes an electrochemical model, a thermal model, and an aging model. The construction of the electrochemical-thermal-aging model includes: Build an electrochemical model: ； Build the thermal model: ； Building an aging model: ； in, express The terminal voltage at the moment, represents the solid phase potential, represents the spatial coordinates, Indicates the length of the negative electrode, represents the length of the diaphragm, Indicates the length of the positive electrode, Indicates the battery density, represents the specific heat capacity, Indicates temperature, represents thermal conductivity, represents the heat generation rate determined by the electrochemical model, Indicates lithium ion loss, represents the reaction current formed by the solid electrolyte interface layer, represents the reaction current associated with the continuous formation of the surface solid electrolyte interface, Represents the reaction current of the formation of a new solid electrolyte interface layer on the cracks of graphite particles.

3. The multi-objective optimization lithium-ion battery charging design method according to claim 1, characterized in that: The construction of the enhanced Chebyshev function includes: ； in, represents the scalarized target generated by the augmented Chebyshev function, represents the number of single-objective subproblems in the constrained multi-objective optimization problem, Represents the weight vector No. A quantity, represents the protocol parameter vector, represents the first A normalized single-objective subproblem, Indicates a positive weight coefficient.

4. The multi-objective optimization lithium-ion battery charging design method according to claim 1, characterized in that: The step of constructing a Gaussian process model for a single-target subproblem based on the scalarized target data set includes: A Gaussian kernel function is used to quantify the correlation between any two charging protocols in the scalarized target data set; Constructing a correlation matrix based on the correlation between any two charging protocols; Calculating a predicted value and an uncertainty of the predicted value of a charging protocol based on the correlation matrix and the scalarized target data set; A Gaussian process model of the single-objective subproblem is constructed according to the predicted value and the uncertainty of the predicted value.

5. The multi-objective optimization lithium-ion battery charging design method according to claim 4, characterized in that: The calculating, based on the correlation matrix and the scalarized target data set, a predicted value of the charging protocol and an uncertainty of the predicted value, includes: ； ； ； ； in, represents the predicted value of the charging protocol, Represents N-dimensional full vector, represents transpose, represents the correlation matrix, represents the inverse transformation of the correlation matrix, represents the target value set in the scalar target dataset, represents an N-dimensional vector, represents the uncertainty of the predicted value, represents N dimensions, Represents the protocol parameter vector.

6. A multi-objective optimization lithium-ion battery charging design system, characterized in that: The system comprises: A first construction unit is configured to construct a constrained multi-objective optimization problem comprising a plurality of single-objective sub-problems and constraint conditions, wherein each single-objective sub-problem satisfies the constraint conditions; a second construction unit, configured to construct, using historical data corresponding to the current battery health state, a first database comprising a plurality of first charging protocols and total target values ​​corresponding to the first charging protocols, and to construct, using empirical data corresponding to other battery health states, a second database comprising a plurality of second charging protocols and total target values ​​corresponding to the second charging protocols, wherein the total target value is a total optimization objective in a constrained multi-objective optimization problem corresponding to the charging protocols; A third building unit is used to build an electrochemical-thermal-aging model and an enhanced Chebyshev function; a first scalarization unit, configured to scalarize the constrained multi-objective optimization problem corresponding to each of the first charging protocols based on the first database and the enhanced Chebyshev function to obtain a scalarized target data set; a second scalarization unit, configured to scalarize the constrained multi-objective optimization problem corresponding to each second charging protocol based on the second database and the enhanced Chebyshev function to obtain a scalarized source data set; A model building unit, configured to build a Gaussian process model of a single-objective subproblem based on the scalarized target data set; A charging protocol determination unit is configured to design a migration acquisition function based on the scalarized source data set using the Gaussian process model, solve the migration acquisition function, and evaluate the solution using the electrochemical-thermal-aging model to determine a set of target charging protocols, specifically: Based on the scalarized source data set, using the predicted value and the uncertainty of the predicted value in the Gaussian process model, a migration acquisition function is designed, including: ； ； in, represents the migration acquisition function, represents the predicted value of the charging protocol, represents the balance coefficient, represents the uncertainty of the predicted value, represents the weight term, represents a Gaussian distribution derived from a scalarized source dataset, represents the maximum number of iterations, Indicates the current iteration number, Indicates a preset constant, represents the KL divergence, represents the Gaussian distribution derived from the scalarized target dataset, represents the protocol parameter vector; An evolutionary algorithm is used to solve the migration acquisition function to obtain a new charging protocol; Adding the new charging protocol to the first database to obtain an updated database; Evaluating the new charging protocol using the electrochemical-thermal-aging model to obtain an evaluation result; A set of target charging protocols is selected from the updated database according to the evaluation result.

7. An electronic device, characterized in that: The invention comprises at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor to enable the at least one control processor to perform the multi-objective optimization lithium-ion battery charging design method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the multi-objective optimization lithium-ion battery charging design method according to any one of claims 1 to 5.

Images