Intelligent optimization method, device and storage medium for battery simulation design parameters

Abstract

The present application relates to an intelligent optimization method, device and storage medium for battery simulation design parameters, belonging to the field of battery design technology. The method includes: constructing a battery simulation model; obtaining battery performance evaluation indicators; determining the type of design parameters that can be optimized in the battery simulation model; establishing an objective function based on the performance evaluation indicators; searching for the optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches the expected value; it can solve the problems of high complexity and low efficiency of traditional simulation optimization methods; since the present application can accurately simulate the electrochemical process of the battery, it can ensure the accuracy and reliability of the evaluation results. At the same time, the whole process is automated, reducing human intervention and improving the efficiency and consistency of optimization.

Description

Technical Field

[0001] The present application relates to an intelligent optimization method, an operation method, a device and a storage medium for battery simulation design parameters, and belongs to the technical field of battery design. Background Art

[0002] With the rapid development of new energy technologies, battery technology has become an indispensable energy storage solution in modern electronic devices and electric vehicles. The performance of batteries directly affects the endurance, safety and service life of the equipment. In order to meet the growing performance requirements, battery design must be continuously optimized to improve energy density, power density, cycle stability and safety.

[0003] How to design efficient and long-life batteries has become a hot topic of research. Traditional battery design processes are mostly based on trial and error, relying on engineers' experience and a large number of experimental verifications. This is not only time-consuming and costly, but also incapable of coping with the complex and ever-changing battery design parameter space. Designers often need to rely on experience and intuition to select a few parameter combinations for experiments, which not only limits the exploration of a wider design space, but may also lead to missing the optimal or near-optimal design solution.

[0004] In recent years, the progress of computer-aided design and simulation technology has brought new possibilities to battery design. Simulation technology can predict the performance of batteries in a virtual environment, reducing the reliance on actual sample testing and accelerating the research and development process. Most existing simulation optimization methods rely on manually setting parameter ranges and optimization strategies.

[0005] However, manually setting parameter ranges and optimization strategies lacks intelligence and adaptability, and is difficult to cope with the high complexity and nonlinear characteristics of battery design. At this point, even with the use of simulation technology, how to efficiently and systematically screen and optimize numerous design parameters remains a major challenge.

[0006] Therefore, there is an urgent need for a more advanced and intelligent method that can automatically and efficiently explore and optimize battery design parameters to maximize battery performance while shortening product time to market and reducing R&D costs. Summary of the invention

[0007] This application provides a method, device and storage medium for intelligent optimization of battery simulation design parameters, which can realize automatic parameter adjustment and optimization, quickly find the optimal parameter combination of battery design, significantly improve the performance and life of the battery, and reduce R&D costs and time. This application provides the following technical solutions:

[0008] In a first aspect, a method for intelligently optimizing battery simulation design parameters is provided, the method comprising:

[0009] Constructing a battery simulation model, wherein the battery simulation model is a mathematical model for simulating the electrochemical behavior of a real battery under working conditions;

[0010] Obtaining a performance evaluation index of the battery; wherein the performance evaluation index is used to evaluate the basic performance and working performance of the battery, the basic performance is used to indicate the overall working capacity of the battery, and the working performance is used to indicate the charging and discharging capacity of the battery during operation;

[0011] Determining the type of design parameters that can be optimized in the battery simulation model, wherein the type of design parameters refers to the type of parameters that affect battery performance when designing a battery;

[0012] Establishing an objective function based on the performance evaluation index; wherein the objective function is used to indicate an optimization target for battery design;

[0013] Searching for optimal design parameters in a search space based on a preset parameter search algorithm so that a preset objective function reaches an expected value; wherein the search space includes all value solutions corresponding to each design parameter type;

[0014] A design solution for the battery is generated based on the optimal design parameters.

[0015] Optionally, searching for optimal design parameters in the search space based on a preset parameter search algorithm so that a preset objective function reaches an expected value includes:

[0016] quantifying the potential value of using different design parameters in the search space to evaluate the objective function through an acquisition function, to obtain an acquisition function value; the acquisition function is determined based on a mean estimate and a variance estimate of the objective function;

[0017] Determine the design parameter corresponding to the maximum acquisition function value as the candidate design parameter, input the objective function established based on the probability model, and obtain the objective function value; the objective function is obtained by training based on the training set, and the initialization training set includes initializing the design parameter;

[0018] Adding the candidate design parameters to the training set, training the objective function again, and obtaining an updated mean estimate and variance estimate of the objective function; triggering the step of obtaining a function value and subsequent steps by obtaining a function quantification to evaluate the objective function using different design parameters in the search space; and stopping until the objective function value remains substantially unchanged or the number of iterations reaches a preset number;

[0019] Determine the candidate design parameter corresponding to the maximum objective function value as the optimal design parameter.

[0020] Optionally, the probability model includes a Gaussian process; the acquisition function includes an upper confidence limit UCB function, and the UCB function UCB(x) is expressed by the following formula:

[0021] UCB(x)=μ(x)+βσ(x);

[0022] Among them, μ(x) is the mean estimate of the objective function at the design parameter x, σ(x) is the variance estimate of the objective function at the design parameter x, and β is a hyperparameter that controls the trade-off between exploration and exploitation.

[0023] Optionally, searching for optimal design parameters in the search space based on a preset parameter search algorithm so that a preset objective function reaches an expected value includes:

[0024] Obtaining current design parameters of the battery;

[0025] Determining the current gradient of the objective function based on the design parameters; the gradient indicates the direction and magnitude of the rate of change of the objective function and points to the direction in which the objective function grows fastest;

[0026] The design parameters are updated based on the gradient and the learning rate, where the learning rate is a positive scalar and is used to control the step size of each update of the design parameters; the step of determining the current gradient of the objective function based on the design parameters and subsequent steps are triggered to execute until an iteration stop condition is met;

[0027] The design parameters that meet the iteration stop condition are determined as the optimal design parameters.

[0028] Optionally, searching for optimal design parameters in the search space based on a preset parameter search algorithm so that a preset objective function reaches an expected value includes:

[0029] quantifying the potential value of using different design parameters in the search space to evaluate the objective function through an acquisition function, to obtain an acquisition function value; the acquisition function is determined based on a mean estimate and a variance estimate of the objective function;

[0030] Determine the design parameter corresponding to the maximum acquisition function value as the candidate design parameter, input the objective function established based on the probability model, and obtain the objective function value; the objective function is obtained by training based on the training set, and the initialization training set includes initializing the design parameter;

[0031] The candidate design parameters are added to the training set, and the objective function is trained again to obtain the mean estimate and variance estimate of the updated objective function; triggering the step of obtaining the function value and the subsequent steps by obtaining the potential value of evaluating the objective function using different design parameters in the search space by obtaining the function quantification; until the first stopping condition is met, the design parameters corresponding to the first stopping condition are obtained;

[0032] Determining the current gradient of the objective function based on the design parameters corresponding to the first stop condition; the gradient indicates the direction and magnitude of the rate of change of the objective function and points to the direction in which the objective function grows fastest;

[0033] updating the design parameters based on the gradient and the learning rate, wherein the learning rate is a positive scalar and is used to control the step size of each update of the design parameters; triggering the step of determining the current gradient of the objective function based on the design parameters and subsequent steps until a second stop condition is met;

[0034] The design parameters when the second stop condition is satisfied are determined as the optimal design parameters.

[0035] Optionally, the battery simulation model is a pseudo two-dimensional P2D lithium-ion battery model, and constructing the battery simulation model includes:

[0036] Initializing a volume source having a uniform distribution, the volume source representing a rate of generation or consumption of lithium ions in an electrochemical reaction;

[0037] The solid-phase potential distribution is solved using zero gradient boundary conditions and fixed value boundary conditions. The actual distribution of the solid-phase potential is obtained by subtracting the initially set reference value from the solved potential value. The solid-phase potential refers to the potential distribution inside the positive and negative electrode materials of the battery.

[0038] In the current time step, the solution equation of the liquid phase potential is regarded as the Laplace equation with a fixed volume source term, and the liquid phase potential value is obtained; where the liquid phase potential refers to the potential distribution in the electrolyte;

[0039] Calculate the relative value of the solid phase potential and the liquid phase potential and the solid phase lithium ion concentration to obtain the relative value of the overpotential;

[0040] Using the charge conservation constraint combined with the Butler-Volmer equation, the potential field is associated with the concentration field so that the potential field affects the concentration field and updates the reference value of the solid-liquid phase potential;

[0041] The absolute value of the solid-liquid phase potential is calculated by combining the relative value and the updated reference value, and the reaction current density is updated; the volume source term is calculated by introducing the specific surface area;

[0042] The lithium ion concentration field in the solid and liquid phases is updated according to the updated reaction current density and concentration gradient boundary conditions.

[0043] Optionally, the performance evaluation indicators of the basic performance include: rated energy, continuous power, peak power, energy density and power density;

[0044] The rated energy is calculated by integrating the product of voltage and current over time after the battery reaches the lower cut-off voltage at a certain discharge current rate;

[0045] The continuous power is the power value corresponding to the battery being discharged at the first constant power for the first preset time and then continuously increasing the discharge power until the battery reaches the lower limit cut-off voltage at the specified time;

[0046] The peak power is the power value corresponding to the battery discharging at the second constant power for the second preset time and then continuously increasing the discharge power until the battery reaches the lower limit cut-off voltage at the specified time; wherein the second constant power is greater than the first constant power, and the second preset time is less than the first preset time;

[0047] The energy density includes mass energy density and volume energy density. The mass energy density is the energy stored in the battery per unit mass, which is calculated by dividing the total energy by the mass of the battery. The volume energy density is the energy stored in the battery per unit volume, which is calculated by dividing the total energy by the volume of the battery.

[0048] The power density includes mass power density and volume power density. The mass power density is the maximum power that a battery per unit mass can output or input per unit time, which is calculated by dividing the maximum power by the mass of the battery. The volume power density is the maximum power that a battery per unit volume can output or input per unit time, which is calculated by dividing the maximum power by the volume of the battery.

[0049] The performance evaluation indicators of the working performance include: energy efficiency and fast charging time;

[0050] Energy efficiency: refers to the ratio of output energy to input energy of a battery during energy conversion or use; it is calculated by dividing the battery discharge energy by the battery charge energy;

[0051] The fast charge time refers to the time required for the battery to charge from a low power state to a high power state.

[0052] Optionally, the types of design parameters that can be optimized include the following categories:

[0053] Appearance variable class, including: battery size information;

[0054] Mechanical parts variable class, including: battery end cap information and tab information;

[0055] Pole and diaphragm variable class, including: battery pole and diaphragm information;

[0056] Mass fraction variable class, including: mass fraction ratio of battery active materials, conductive agents and binders.

[0057] In a second aspect, a device for intelligently optimizing battery simulation design parameters is provided, the device comprising:

[0058] A model building module, used to build a battery simulation model, wherein the battery simulation model is a mathematical model used to simulate the electrochemical behavior of a real battery under working conditions;

[0059] An index acquisition module, used to acquire a performance evaluation index of a battery; wherein the performance evaluation index is used to evaluate the basic performance and working performance of the battery, the basic performance is used to indicate the overall working capacity of the battery, and the working performance is used to indicate the charging and discharging capacity of the battery during operation;

[0060] A parameter determination module, used to determine the type of design parameters that can be optimized in the battery simulation model, wherein the type of design parameters refers to the type of parameters that affect battery performance when designing a battery;

[0061] A function establishment module, used to establish an objective function based on the performance evaluation index; wherein the objective function is used to indicate an optimization target of a battery design;

[0062] A parameter optimization module, used to search for optimal design parameters in a search space based on a preset parameter search algorithm so that a preset objective function reaches an expected value; wherein the search space includes all value solutions corresponding to each design parameter type;

[0063] A solution generation module is used to generate a battery design solution based on the optimal design parameters.

[0064] In a third aspect, a device for intelligent optimization of battery simulation design parameters is provided, the device comprising a processor and a memory; a program is stored in the memory, and the program is loaded and executed by the processor to implement the intelligent optimization method for battery simulation design parameters described in the first aspect.

[0065] In a fourth aspect, a computer-readable storage medium is provided, wherein a program is stored in the storage medium, and the program is loaded and executed by the processor to implement the intelligent optimization method of battery simulation design parameters described in the first aspect.

[0066] The beneficial effects of the present application are: by constructing a battery simulation model to simulate a mathematical model of the electrochemical behavior of a real battery under working conditions; obtaining the performance evaluation index of the battery; determining the type of design parameters that can be optimized in the battery simulation model, the design parameter type refers to the parameter type that affects the battery performance when designing the battery; establishing an objective function based on the performance evaluation index; wherein the objective function is used to indicate the optimization target of the battery design; searching for the optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches the expected value; wherein the search space includes all value schemes corresponding to each design parameter type; generating a battery design scheme based on the optimal design parameters; can solve the problems of high complexity and low efficiency of traditional simulation optimization methods; since the present application can accurately simulate the electrochemical process of the battery, the accuracy and reliability of the evaluation results can be ensured. Parameter adjustment strategies include but are not limited to parameter fine-tuning, parameter range expansion and parameter reinitialization to cope with different optimization challenges. At the same time, the whole process is automated, reducing human intervention and improving the efficiency and consistency of optimization. Through this automated parameter adjustment and optimization, the present invention can quickly find the optimal parameter combination for battery design, significantly improve the performance and life of the battery, and reduce R&D costs and time. This intelligent optimization method provides an efficient and innovative solution in the field of battery design.

[0067] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present application in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a flow chart of an intelligent optimization method of battery simulation design parameters provided by an embodiment of the present application;

[0069] Figure 2 is a schematic diagram of an intelligent optimization process of battery simulation design parameters provided by an embodiment of the present application;

[0070] Figure 3 is a schematic diagram of an optimization process of battery peak power performance provided by an embodiment of the present application;

[0071] Figure 4 It is a schematic diagram of the change process of the volume percentage of the positive electrode porosity, a design variable provided by an embodiment of the present application;

[0072] Figure 5 is a block diagram of an intelligent optimization device for battery simulation design parameters provided by an embodiment of the present application;

[0073] Figure 6It is a block diagram of an intelligent optimization device for battery simulation design parameters provided by an embodiment of the present application. DETAILED DESCRIPTION

[0074] The specific implementation methods of the present application are further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0075] Figure 1 This is a flowchart of intelligent optimization of battery simulation design parameters provided by an embodiment of the present application. This embodiment is described by taking the method used in an electronic device with computing capabilities as an example. The electronic device can be a user terminal or a server. The user terminal includes but is not limited to: a computer, a tablet computer, etc. This embodiment does not limit the type of the electronic device. The method includes at least the following steps:

[0076] Step 101 : constructing a battery simulation model, where the battery simulation model is a mathematical model used to simulate the electrochemical behavior of a real battery under working conditions.

[0077] In this embodiment, the battery simulation model can describe the physical structure, chemical composition and electrical characteristics of the battery in detail. The battery simulation model includes but is not limited to: the microscopic pore structure of the electrode, the distribution of active materials, the composition of the electrolyte and the macroscopic geometry of the battery. The battery simulation model is designed to capture the complex electrochemical behavior of the battery under various working conditions, including but not limited to: the diffusion of lithium ions in the electrode material, the transmission of electrons in the electrode and the current collector, and the chemical reaction kinetics inside the battery. To achieve this, the battery simulation model adopts a multi-physics field coupling method, integrating physical phenomena in multiple fields such as electricity, thermodynamics, fluid mechanics and chemical kinetics. The construction of the battery simulation model adopts advanced numerical techniques, such as finite element analysis, finite difference method, finite volume method, boundary element method, multi-scale modeling, etc. Among them, the finite element method is to discretize the continuous space region into a finite number of small units, transform the partial differential equation into a local equation inside each unit, and then combine these local equations to form a system of equations for the entire region, and obtain an approximate solution to the partial differential equation by solving the system of equations. The basic idea of ​​the finite difference method is to divide the solution domain into differential grids, replace the continuous solution domain with finite grid nodes, and use methods such as Taylor series expansion to replace the differential quotient in the well-defined problem with the difference quotient, thereby discretizing the original problem into a differential format and obtaining a numerical solution. The finite volume method, also known as the finite volume method or the control volume method, divides the solution domain into finite discrete control volumes, integrates the equilibrium equations inside each control volume, and obtains a set of discrete equations, and then obtains an approximate solution by solving the discrete equations. The boundary element method is a numerical method developed on the basis of the classical integral equation method and the finite element method. Unlike the idea of ​​dividing units in the solution domain of the finite element method, the boundary element method only divides units on the boundary of the definition domain, discretizes the boundary integral equations into a set of linear algebraic equations, and obtains the physical quantities of the boundary nodes by solving these equations. Multiscale modeling technology combines models at different scales, from atomic level to macroscopic structure, to capture the behavior of materials or systems at different scales. Through this highly integrated and accurate simulation model, the present embodiment can provide in-depth insights into battery design and lay a solid foundation for subsequent intelligent optimization algorithms.

[0078] This implementation can be implemented using a variety of lithium-ion battery models, such as single-particle models, multi-particle models, pseudo-two-dimensional models, pseudo-three-dimensional models, multi-scale models, etc. Taking the establishment of a pseudo-two-dimensional lithium-ion battery model as an example, within each time step, the model will update the potential field, lithium ion concentration field and other related physical quantities to reflect the state of the battery at the end of the time step. Specifically, building a battery simulation model includes the following steps:

[0079] Step 1, initializing a volume source with a uniform distribution, the volume source representing the generation or consumption rate of lithium ions in an electrochemical reaction;

[0080] Step 2, using zero gradient boundary conditions and fixed value boundary conditions to solve the solid phase potential distribution, subtracting the initially set reference value from the solved potential value to obtain the actual distribution of the solid phase potential; wherein the solid phase potential refers to the potential distribution inside the positive and negative electrode materials of the battery;

[0081] Step 3, in the current time step, the solution equation of the liquid phase potential is regarded as the Laplace equation with a fixed volume source term, and the liquid phase potential value is obtained; wherein the liquid phase potential refers to the potential distribution in the electrolyte;

[0082] Step 4, calculating the relative value of the solid phase potential and the liquid phase potential and the solid phase lithium ion concentration to obtain the relative value of the overpotential;

[0083] Step 5, using the charge conservation constraint combined with the Butler-Volmer equation, the potential field is associated with the concentration field so that the potential field affects the concentration field and the reference value of the solid-liquid phase potential is updated;

[0084] Step 6, combining the relative value and the updated reference value to calculate the absolute value of the solid-liquid phase potential, and updating the reaction current density; and calculating the volume source term by introducing the specific surface area;

[0085] Step 7, updating the lithium ion concentration field of the solid and liquid phases according to the updated reaction current density and concentration gradient boundary conditions.

[0086] Step 102, obtaining a performance evaluation index of the battery; wherein the performance evaluation index is used to evaluate the basic performance and working performance of the battery, the basic performance is used to indicate the overall working capacity of the battery, and the working performance is used to indicate the charging and discharging capacity of the battery during operation.

[0087] Performance evaluation indicators are key elements in the battery simulation design process and determine the goal and direction of the optimization algorithm. In order to comprehensively evaluate the performance of the battery design, this embodiment defines a series of comprehensive performance indicators, which include not only the basic performance of the battery, such as rated energy, continuous power, peak power, energy density, power density, etc., but also the performance of the battery in practical applications, such as energy efficiency, fast charging time, etc. Through these evaluation indicators, a comprehensive evaluation of the battery design can be carried out to ensure that the optimization results are not only optimal in theory, but also reliable and effective in practical applications. These indicators will be used to evaluate the simulation results and guide the optimization process.

[0088] Schematically, the performance evaluation indicators of basic performance include: rated energy, continuous power, peak power, energy density and power density;

[0089] Rated energy: After the battery reaches the lower cut-off voltage at a certain discharge current rate, the energy released by the battery is calculated by multiplying the voltage and current and integrating them over time;

[0090] Continuous power: refers to the power value corresponding to the battery being discharged at the first constant power for the first preset time, and then the discharge power is continuously increased until the battery reaches the lower limit cut-off voltage at the specified time;

[0091] Peak power: refers to the power value corresponding to the battery being discharged at the second constant power for the second preset time, and then continuously increasing the discharge power until the battery reaches the lower limit cut-off voltage at the specified time; wherein the second constant power is greater than the first constant power, and the second preset time is less than the first preset time;

[0092] Energy density: includes mass energy density and volume energy density. Mass energy density is the energy stored in the battery per unit mass, calculated by dividing the total energy by the mass of the battery; volume energy density is the energy stored in the battery per unit volume, calculated by dividing the total energy by the volume of the battery;

[0093] Power density: including mass power density and volume power density. Mass power density is the maximum power that a battery per unit mass can output or input per unit time, calculated by dividing the maximum power by the mass of the battery; volume power density is the maximum power that a battery per unit volume can output or input per unit time, calculated by dividing the maximum power by the volume of the battery;

[0094] The performance evaluation indicators of working performance include: energy efficiency and fast charging time;

[0095] Energy efficiency: refers to the ratio of the output energy to the input energy of the battery during energy conversion or use; calculated by dividing the battery discharge energy by the battery charge energy;

[0096] Fast charge time: refers to the time required for the battery to charge from a low power state to a high power state. The calculation method of fast charge time includes: given a state of charge (SOC) sequence, each two adjacent SOCs are a SOC interval, calculate the instantaneous charging current corresponding to the lower limit SOC in each SOC interval, use this current as the charging current in the current SOC interval, use the difference between the upper and lower limits of the SOC interval to get the SOC percentage, and then multiply it by the total capacity of the battery to get the capacity corresponding to the SOC interval, calculate the charging time in the SOC interval by dividing the capacity by the charging current, and then add up the charging time of all SOC intervals to get the total fast charge time.

[0097] Step 103, determining the type of design parameters that can be optimized in the battery simulation model, where the design parameter type refers to the type of parameters that affect battery performance when designing the battery.

[0098] Before performing target optimization, it is necessary to select designable battery parameters, that is, design parameter types. These parameters include, but are not limited to: the type and mass fraction of electrode materials, the geometric dimensions of the battery, the thickness and compaction density of the electrode, and the size information of the pole. When selecting these parameters, this embodiment comprehensively considers the manufacturing process and expected performance of the battery, screens out the key parameters that have a significant impact on the battery performance, and provides a practical parameter space for the intelligent optimization algorithm. The selection of these parameters is not only based on the current technical level, but also takes into account future technological development trends to ensure the foresight and adaptability of the design. In addition, this embodiment also takes into account the interactions and constraints between the parameters to ensure that the selected parameters can meet the technical requirements of the actual manufacturing process while ensuring the battery performance. By selecting designable parameters, this embodiment provides a flexible, changeable and scientifically reasonable optimization starting point for battery design.

[0099] Specifically, the types of design parameters that can be optimized include the following:

[0100] Appearance variable class, including: battery size information; such as: battery width, height, thickness and other information;

[0101] Mechanical parts variable class, including: battery end cap information and terminal lug information; such as: end cap thickness, terminal column margin and other information;

[0102] The electrode and diaphragm variable class includes: the battery electrode information and diaphragm information; such as: electrode compaction density and coating thickness, diaphragm thickness and porosity and other information.

[0103] Mass fraction variable class, including: mass fraction ratio of battery active materials, conductive agents and binders.

[0104] Step 104, establishing an objective function based on the performance evaluation index; wherein the objective function is used to indicate an optimization target for battery design.

[0105] The optimization target of the objective function may be to maximize the peak power, or may be to maximize other parameters. This embodiment does not limit the optimization target of the objective function.

[0106] Step 105 , searching for optimal design parameters in a search space based on a preset parameter search algorithm, so that a preset objective function reaches an expected value; wherein the search space includes all value solutions corresponding to each design parameter type.

[0107] The parameter search algorithm is the core of the present invention and is used to automatically adjust the battery design parameters. This embodiment provides a variety of advanced optimization algorithms that can handle complex multivariable, nonlinear optimization problems, ensure efficient and accurate searches in the battery design parameter space, find the optimal or near-optimal solution, and effectively avoid local optimal solutions, find the global optimal design parameters, thereby significantly improving the performance and efficiency of the battery design. In addition, this embodiment also takes into account the scalability and adaptability of the algorithm to ensure that the algorithm can continue to evolve and improve with the development of battery technology and the emergence of new design requirements. These optimization algorithms include but are not limited to the following:

[0108] The first type: Bayesian optimization, which uses a probability model to predict the objective function and selects the next sampling point by maximizing the collected information. Specifically, based on a preset parameter search algorithm, the optimal design parameters are searched in the search space so that the preset objective function reaches the expected value, including:

[0109] The acquisition function is used to quantify the potential value of using different design parameters in the search space to evaluate the objective function, and the acquisition function value is obtained; the acquisition function is determined based on the mean estimate and variance estimate of the objective function;

[0110] The design parameter corresponding to the maximum acquisition function value is determined as a candidate design parameter, and the objective function established based on the probability model is input to obtain the objective function value; the objective function is obtained by training based on the training set, and the initialization training set includes initializing the design parameter;

[0111] The candidate design parameters are added to the training set, and the objective function is trained again to obtain the mean estimate and variance estimate of the updated objective function; triggering the execution of the step of obtaining the function value and the subsequent steps by obtaining the potential value of the objective function when evaluating the objective function using different design parameters in the search space; and stopping when the objective function value remains substantially unchanged or the number of iterations reaches a preset number;

[0112] The candidate design parameters corresponding to the maximum objective function value are determined to be the optimal design parameters.

[0113] Herein, remaining substantially unchanged means that the objective function value remains unchanged or the change in the objective function value is less than a preset allowable range.

[0114] The probability model includes a Gaussian process; the acquisition function includes an upper confidence limit UCB function, and the UCB function UCB(x) is expressed by the following formula:

[0115] UCB(x)=μ(x)+βσ(x);

[0116] Among them, μ(x) is the mean estimate of the objective function at the design parameter x, σ(x) is the variance estimate of the objective function at the design parameter x, and β is a hyperparameter that controls the trade-off between exploration and exploitation.

[0117] The second type: gradient descent algorithm, which iteratively calculates the gradient of the objective function, updates the solution to reduce the error, converges quickly but may fall into a local minimum.

[0118] Based on the preset parameter search algorithm, the optimal design parameters are searched in the search space so that the preset objective function reaches the expected value, including:

[0119] Obtain the current design parameters of the battery;

[0120] Determine the current gradient of the objective function based on the design parameters; the gradient indicates the direction and magnitude of the rate of change of the objective function and points to the direction in which the objective function is growing fastest;

[0121] Update the design parameters based on the gradient and the learning rate. The learning rate is a positive scalar used to control the step size of each update of the design parameters. Trigger the execution of the step of determining the current gradient of the objective function based on the design parameters and the subsequent steps until the iteration stop condition is met.

[0122] The design parameters that meet the iteration stopping conditions are determined as the optimal design parameters.

[0123] Optionally, the iteration condition may be that the gradient is less than a certain threshold, or a predetermined number of iterations is reached, or the parameter update is small enough. This embodiment does not limit the implementation method of the iteration condition.

[0124] The third method: combining Bayesian optimization and gradient descent algorithm to optimize parameters. That is, first fully explore the parameter space through Bayesian optimization, then use the gradient descent algorithm to further optimize the top-ranked results obtained by Bayesian optimization, and finally compare all optimization results to obtain the global optimal solution. The algorithm used in the present invention can balance the relationship between exploration (discovering new design areas) and utilization (improving the current best solution) to ensure the globality and effectiveness of the search process. Specifically, based on the preset parameter search algorithm, the optimal design parameters are searched in the search space so that the preset objective function reaches the expected value, including:

[0125] The acquisition function is used to quantify the potential value of using different design parameters in the search space to evaluate the objective function, and the acquisition function value is obtained; the acquisition function is determined based on the mean estimate and variance estimate of the objective function;

[0126] The design parameter corresponding to the maximum acquisition function value is determined as a candidate design parameter, and the objective function established based on the probability model is input to obtain the objective function value; the objective function is obtained by training based on the training set, and the initialization training set includes initializing the design parameter;

[0127] The candidate design parameters are added to the training set, and the objective function is trained again to obtain the mean estimate and variance estimate of the updated objective function; triggering the execution of the step of obtaining the function value and the subsequent steps by obtaining the potential value of the objective function when evaluating the objective function using different design parameters in the search space; until the first stopping condition is met, the design parameters corresponding to the first stopping condition are obtained;

[0128] Determine the current gradient of the objective function based on the design parameters corresponding to the first stop condition; the gradient indicates the direction and magnitude of the rate of change of the objective function and points to the direction in which the objective function grows fastest;

[0129] The design parameters are updated based on the gradient and the learning rate, where the learning rate is a positive scalar and is used to control the step size of each update of the design parameters; the step of determining the current gradient of the objective function based on the design parameters and subsequent steps are triggered to execute until the second stop condition is met;

[0130] The design parameters when the second stop condition is satisfied are determined as the optimal design parameters.

[0131] Among them, the first stopping condition may be that the number of iterations reaches a certain number, or the objective function value reaches a certain value. The first stopping condition may be the same as or different from the stopping condition in the first method. This embodiment does not limit the implementation method of the first stopping condition.

[0132] The second stop condition may be the same as or different from the iteration stop condition in the second mode. The second stop condition may be that the number of iterations reaches a preset number, etc. This embodiment does not limit the implementation method of the second stop condition.

[0133] Step 106: Generate a battery design solution based on the optimal design parameters.

[0134] In one example, the electronic device may directly output the optimal design parameters; or, may generate a battery parameter design report output based on the optimal design parameters.

[0135] In summary, the intelligent optimization method of battery simulation design parameters provided in this embodiment constructs a battery simulation model to simulate a mathematical model of the electrochemical behavior of a real battery under working conditions; obtains the performance evaluation index of the battery; determines the type of design parameters that can be optimized in the battery simulation model, and the design parameter type refers to the parameter type that affects the battery performance when designing the battery; establishes an objective function based on the performance evaluation index; wherein the objective function is used to indicate the optimization target of the battery design; searches for the optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches the expected value; wherein the search space includes all value schemes corresponding to each design parameter type; generates a battery design scheme based on the optimal design parameters; can solve the problems of high complexity and low efficiency of traditional simulation optimization methods; since the present application can accurately simulate the electrochemical process of the battery, the accuracy and reliability of the evaluation results can be ensured. Parameter adjustment strategies include but are not limited to parameter fine-tuning, parameter range expansion and parameter reinitialization to meet different optimization challenges. At the same time, the whole process is automated, reducing human intervention and improving the efficiency and consistency of optimization. Through this automated parameter adjustment and optimization, the present invention can quickly find the optimal parameter combination for battery design, significantly improving the performance and life of the battery while reducing R&D costs and time. This intelligent optimization method provides an efficient and innovative solution for the field of battery design.

[0136] Below, taking the peak power class as an optimization target as an example, the specific execution process of the intelligent optimization method for battery simulation design parameters provided by this application is introduced, and finally a battery parameter design report is formed. The method includes the following steps:

[0137] Step 1: First, establish a battery simulation model. For example, this embodiment uses the P2D model of a non-iterative volume source implemented by the open source software OpenFOAM.

[0138] Step 2: Input the initial battery design parameters, including battery configuration, electrode materials, geometric dimensions, thermodynamics, etc. These initial parameters are derived from battery parameters that have been simulated for long-term or short-term performance, or battery parameters that have been tested for proofing. This ensures the rationality of the initial parameters, makes the starting point of parameter optimization calculable, and avoids the problem that the optimization algorithm may not be able to find a reasonable parameter combination in a given parameter space.

[0139] Step 3, define the performance evaluation index, select the peak discharge power of the battery as the optimization target to establish the objective function, the relevant optimization target parameters are the initial temperature of 25°C, the discharge starting point of 20% SOC, and the target discharge time of 10s. The final optimization maximizes the peak power index of the battery.

[0140] Step 4, select battery design parameters, select the D50 particle size of the positive and negative electrode materials, the volume percentage of the positive and negative electrode porosity, the thickness of the positive electrode single-sided coating, the NP ratio, the diaphragm thickness and the diaphragm porosity as the design parameters.

[0141] Step 5, select the parameter search algorithm. Since the selected performance indicator is a single objective and there are many optimization parameters, Bayesian optimization is used, which is a global optimization method based on a probability model. By constructing a probability prior of the objective function (peak power), the model is continuously updated using observation data to guide the search for new solutions. The Bayesian algorithm can effectively balance exploration and utilization and quickly converge to an approximate optimal solution. If there are fewer optimization parameters and the parameter space is likely to be continuous, the gradient descent algorithm can be used. If you want to get a more accurate global optimal solution and do not have high requirements for algorithm efficiency, you can use a hybrid optimization algorithm. First, use Bayesian optimization to get some better parameter combinations, then use the gradient descent method for each group of parameter combinations, and finally compare all solutions to get the global optimal solution.

[0142] Step 6, set the number of optimization rounds to 30. In each round of optimization, the Bayesian optimizer will first generate a set of design variable values. These design variables will be used to replace the corresponding parameter values ​​in the battery design parameters. The updated design parameters will then be used to balance the battery to obtain more basic information about the battery. All parameters will then be input into the simulation model to simulate and calculate the electrochemical performance of the battery. The target performance, i.e., the peak power value of the objective function value, is obtained through simulation calculation, and the simulation results are then fed back to the Bayesian optimizer to start a new round of iterations. As the iterations proceed, the design variable parameters are dynamically adjusted, and the performance indicators gradually approach the optimal solution. The specific optimization process flow chart is shown below. Figure 2 shown.

[0143] Step 7: After the optimization algorithm converges, a report will be generated for the final battery design. The report includes design parameters, performance evaluation results, detailed records of the optimization process, and design recommendations, providing important information for further research and development of battery design.

[0144] The optimization process of battery peak power performance is as follows Figure 3 As shown in the figure, the change process of the volume percentage of the design variable positive electrode porosity is as follows Figure 4 shown.

[0145] In summary, this application can not only improve the efficiency and performance of battery design, but also provide a flexible and customizable optimization tool for battery design to meet the needs of different application scenarios.

[0146] Figure 5This is a block diagram of an intelligent optimization device for battery simulation design parameters provided by an embodiment of the present application. The device includes at least the following modules: a model building module 510, an indicator acquisition module 520, a parameter determination module 530, a function establishment module 540, a parameter optimization module 550 and a solution generation module 560.

[0147] A model building module 510 is used to build a battery simulation model, where the battery simulation model is a mathematical model used to simulate the electrochemical behavior of a real battery under working conditions;

[0148] The index acquisition module 520 is used to acquire the performance evaluation index of the battery; wherein the performance evaluation index is used to evaluate the basic performance and working performance of the battery, the basic performance is used to indicate the overall working capacity of the battery, and the working performance is used to indicate the charging and discharging capacity of the battery during operation;

[0149] A parameter determination module 530, used to determine the type of design parameters that can be optimized in the battery simulation model, wherein the type of design parameters refers to the type of parameters that affect battery performance when designing a battery;

[0150] A function establishment module 540 is used to establish an objective function based on the performance evaluation index; wherein the objective function is used to indicate an optimization target of a battery design;

[0151] The parameter optimization module 550 is used to search for the optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches the expected value; wherein the search space includes all value solutions corresponding to each design parameter type;

[0152] The solution generation module 560 is used to generate a battery design solution based on the optimal design parameters.

[0153] For relevant details, refer to the above method embodiment.

[0154] It should be noted that: the intelligent optimization device for battery simulation design parameters provided in the above embodiments only uses the division of the above functional modules as an example when performing intelligent optimization of battery simulation design parameters. In actual applications, the above functional allocation can be completed by different functional modules as needed, that is, the internal structure of the intelligent optimization device for battery simulation design parameters is divided into different functional modules to complete all or part of the functions described above. In addition, the intelligent optimization device for battery simulation design parameters provided in the above embodiments and the intelligent optimization method embodiment for battery simulation design parameters belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0155] Figure 66 is a block diagram of an intelligent optimization device for battery simulation design parameters provided by an embodiment of the present application. The device at least includes a processor 601 and a memory 602.

[0156] The processor 601 may include one or more processing cores, such as a 4-core processor, a 6-core processor, etc. The processor 601 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 601 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 601 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 601 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0157] The memory 602 may include one or more computer-readable storage media, which may be non-transitory. The memory 602 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 602 is used to store at least one instruction, which is used to be executed by the processor 601 to implement the intelligent optimization method of battery simulation design parameters provided in the method embodiment of the present application.

[0158] In some embodiments, the intelligent optimization device for battery simulation design parameters may also optionally include: a peripheral device interface and at least one peripheral device. The processor 601, the memory 602 and the peripheral device interface may be connected via a bus or a signal line. Each peripheral device may be connected to the peripheral device interface via a bus, a signal line or a circuit board. Schematically, the peripheral devices include but are not limited to: a radio frequency circuit, a touch display screen, an audio circuit, and a power supply.

[0159] Of course, the intelligent optimization device for battery simulation design parameters may also include fewer or more components, which is not limited in this embodiment.

[0160] Optionally, the present application also provides a computer-readable storage medium, in which a program is stored, and the program is loaded and executed by a processor to implement the intelligent optimization method of battery simulation design parameters of the above method embodiment.

[0161] Optionally, the present application also provides a computer product, which includes a computer-readable storage medium, in which a program is stored, and the program is loaded and executed by a processor to implement the intelligent optimization method of battery simulation design parameters of the above method embodiment.

[0162] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0163] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. An intelligent optimization method for battery simulation design parameters, characterized in that: The method comprises: Constructing a battery simulation model, wherein the battery simulation model is a mathematical model for simulating the electrochemical behavior of a real battery under working conditions; Wherein, the battery simulation model is a pseudo two-dimensional P2D lithium-ion battery model, and the construction of the battery simulation model includes: Initializing a volume source having a uniform distribution, the volume source representing a rate of generation or consumption of lithium ions in an electrochemical reaction; The solid-phase potential distribution is solved using zero gradient boundary conditions and fixed value boundary conditions. The actual distribution of the solid-phase potential is obtained by subtracting the initially set reference value from the solved potential value. The solid-phase potential refers to the potential distribution inside the positive and negative electrode materials of the battery. In the current time step, the solution equation of the liquid phase potential is regarded as the Laplace equation with a fixed volume source term, and the liquid phase potential value is obtained; where the liquid phase potential refers to the potential distribution in the electrolyte; Calculate the relative value of the solid phase potential and the liquid phase potential and the solid phase lithium ion concentration to obtain the relative value of the overpotential; Using the charge conservation constraint combined with the Butler-Volmer equation, the potential field is associated with the concentration field so that the potential field affects the concentration field and updates the reference value of the solid-liquid phase potential; The absolute value of the solid-liquid phase potential is calculated by combining the relative value and the updated reference value, and the reaction current density is updated; the volume source term is calculated by introducing the specific surface area; Update the lithium ion concentration field of the solid and liquid phases according to the updated reaction current density and concentration gradient boundary conditions; Obtaining a performance evaluation index of the battery; wherein the performance evaluation index is used to evaluate the basic performance and working performance of the battery, the basic performance is used to indicate the overall working capacity of the battery, and the working performance is used to indicate the charging and discharging capacity of the battery during operation; Determining the type of design parameters that can be optimized in the battery simulation model, wherein the type of design parameters refers to the type of parameters that affect battery performance when designing a battery; Establishing an objective function based on the performance evaluation index; wherein the objective function is used to indicate an optimization target for battery design; Based on a preset parameter search algorithm, the optimal design parameters are searched in the search space so that the preset objective function reaches the expected value; wherein the search space includes all value solutions corresponding to each design parameter type; The method of searching for optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches the expected value includes: quantifying the potential value of using different design parameters in the search space to evaluate the objective function through an acquisition function, to obtain an acquisition function value; the acquisition function is determined based on a mean estimate and a variance estimate of the objective function; Determine the design parameter corresponding to the maximum acquisition function value as the candidate design parameter, input the objective function established based on the probability model, and obtain the objective function value; the objective function is obtained by training based on the training set, and the initialization training set includes initializing the design parameter; The candidate design parameters are added to the training set, and the objective function is trained again to obtain the mean estimate and variance estimate of the updated objective function; triggering the step of obtaining the function value and the subsequent steps by obtaining the potential value of evaluating the objective function using different design parameters in the search space by obtaining the function quantification; until the first stopping condition is met, the design parameters corresponding to the first stopping condition are obtained; Determining the current gradient of the objective function based on the design parameters corresponding to the first stop condition; the gradient indicates the direction and magnitude of the rate of change of the objective function and points to the direction in which the objective function grows fastest; updating the design parameters based on the gradient and the learning rate, wherein the learning rate is a positive scalar and is used to control the step size of each update of the design parameters; triggering the step of determining the current gradient of the objective function based on the design parameters and subsequent steps until a second stop condition is met; Determining the design parameters when the second stop condition is satisfied as the optimal design parameters; A design solution for the battery is generated based on the optimal design parameters.

2. The method according to claim 1, characterized in that The method of searching for optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches an expected value includes: quantifying the potential value of using different design parameters in the search space to evaluate the objective function through an acquisition function, to obtain an acquisition function value; the acquisition function is determined based on a mean estimate and a variance estimate of the objective function; Determine the design parameter corresponding to the maximum acquisition function value as the candidate design parameter, input the objective function established based on the probability model, and obtain the objective function value; the objective function is obtained by training based on the training set, and the initialization training set includes initializing the design parameter; Adding the candidate design parameters to the training set, training the objective function again, and obtaining an updated mean estimate and variance estimate of the objective function; triggering the step of obtaining a function value and subsequent steps by obtaining a function quantification to evaluate the objective function using different design parameters in the search space; and stopping until the objective function value remains substantially unchanged or the number of iterations reaches a preset number; Determine the candidate design parameter corresponding to the maximum objective function value as the optimal design parameter.

3. The method according to claim 2, characterized in that The probability model includes a Gaussian process; the acquisition function includes an upper confidence limit UCB function, and the UCB function UCB(x) is expressed by the following formula: UCB(x)=μ(x)+βσ(x); Among them, μ(x) is the mean estimate of the objective function at the design parameter x, σ(x) is the variance estimate of the objective function at the design parameter x, and β is a hyperparameter that controls the trade-off between exploration and exploitation.

4. The method according to claim 1, characterized in that The method of searching for optimal design parameters in the search space based on a preset parameter search algorithm so that the preset objective function reaches an expected value includes: Obtaining current design parameters of the battery; Determining the current gradient of the objective function based on the design parameters; the gradient indicates the direction and magnitude of the rate of change of the objective function and points to the direction in which the objective function grows fastest; The design parameters are updated based on the gradient and the learning rate, where the learning rate is a positive scalar and is used to control the step size of each update of the design parameters; the step of determining the current gradient of the objective function based on the design parameters and subsequent steps are triggered to execute until an iteration stop condition is met; The design parameters that meet the iteration stop condition are determined as the optimal design parameters.

5. The method according to any one of claims 1 to 4, characterized in that: The performance evaluation indicators of the basic performance include: rated energy, continuous power, peak power, energy density and power density; The rated energy is calculated by integrating the product of voltage and current over time after the battery reaches the lower cut-off voltage at a certain discharge current rate; The continuous power is the power value corresponding to the battery being discharged at the first constant power for the first preset time and then continuously increasing the discharge power until the battery reaches the lower limit cut-off voltage at the specified time; The peak power is the power value corresponding to the battery discharging at the second constant power for the second preset time and then continuously increasing the discharge power until the battery reaches the lower limit cut-off voltage at the specified time; wherein the second constant power is greater than the first constant power, and the second preset time is less than the first preset time; The energy density includes mass energy density and volume energy density. The mass energy density is the energy stored in the battery per unit mass, which is calculated by dividing the total energy by the mass of the battery. The volume energy density is the energy stored in the battery per unit volume, which is calculated by dividing the total energy by the volume of the battery. The power density includes mass power density and volume power density. The mass power density is the maximum power that a battery per unit mass can output or input per unit time, which is calculated by dividing the maximum power by the mass of the battery. The volume power density is the maximum power that a battery per unit volume can output or input per unit time, which is calculated by dividing the maximum power by the volume of the battery. The performance evaluation indicators of the working performance include: energy efficiency and fast charging time; Energy efficiency: refers to the ratio of output energy to input energy of a battery during energy conversion or use; it is calculated by dividing the battery discharge energy by the battery charge energy; The fast charge time refers to the time required for the battery to charge from a low power state to a high power state.

6. The method according to any one of claims 1 to 4, characterized in that: The types of design parameters that can be optimized include the following categories: Appearance variable class, including: battery size information; Mechanical parts variable class, including: battery end cap information and tab information; Pole and diaphragm variable class, including: battery pole and diaphragm information; Mass fraction variable class, including: mass fraction ratio of battery active materials, conductive agents and binders.

7. A computer device, characterized in that: The computer device includes a processor and a memory; a program is stored in the memory, and the program is loaded and executed by the processor to implement the intelligent optimization method of battery simulation design parameters as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that: The storage medium stores a program, which, when executed by a processor, is used to implement the intelligent optimization method for battery simulation design parameters as described in any one of claims 1 to 6.

Images