Lithium-ion battery fast charging strategy optimization method, device, equipment and medium

Abstract

The present invention proposes a lithium-ion battery fast-charging strategy optimization method, device, equipment, and medium, including considering the voltage and temperature characteristics of the lithium-ion battery during charging, constructing an electrochemical-thermal coupling model of the lithium-ion battery; obtaining the electrochemical-thermal coupling model and the simulated and measured voltage and temperature of a real lithium-ion battery under different operating conditions; using a differential evolution algorithm to perform parameter identification on the electrochemical-thermal coupling model with the goal of minimizing the simulated and measured voltage and temperature errors to obtain a parameterized electrochemical-thermal coupling model; and using the parameterized electrochemical-thermal coupling model as an environmental model in a reinforcement learning model to guide the generation of a multi-stage constant-current charging strategy. Through the reinforcement learning model, the present invention can continuously explore and optimize the charging strategy to achieve the goal of safe and efficient fast charging.

Description

Technical Field

[0001] The present invention belongs to the technical field of lithium-ion battery charging, and in particular to a method, device, equipment and medium for optimizing a lithium-ion battery fast charging strategy. Background Art

[0002] As an efficient and reliable energy storage technology, lithium-ion batteries have been widely used in many fields such as electric vehicles, energy storage systems, and 3C products. However, in these application scenarios, the battery life of lithium-ion batteries has always been a key issue that needs to be solved urgently. Since it is difficult to achieve a major breakthrough in battery energy density in the short term, fast charging technology has become one of the effective ways to alleviate battery life anxiety. However, inappropriate fast charging methods can easily trigger lithium plating side reactions in lithium-ion batteries. This side reaction not only causes a sharp drop in battery capacity, but in severe cases may also cause internal short circuits in the battery, leading to serious safety accidents such as thermal runaway, fire, and even explosion. Therefore, developing a fast charging strategy that is both safe and efficient is of extremely important practical significance for the practical application of lithium-ion batteries.

[0003] Currently, common fast-charging strategies for lithium-ion batteries include high-rate constant-current charging, multi-stage constant-current charging, and pulse charging. High-rate constant-current charging significantly shortens charging time by significantly increasing the charging current, making it relatively easy to implement, but it has significant drawbacks. Improper charging rates can easily induce lithium plating, a side reaction. Furthermore, because high-current charging can cause severe polarization, high-rate constant-current charging strategies often have limited charging capacity, making it difficult to fully charge the battery. Multi-stage constant-current charging leverages the ability of lithium-ion batteries to accept different charging rates at different states of charge (SOC). By adjusting the charging current in stages, this strategy maintains charging speed while also ensuring safety. In particular, using a relatively low charge rate at high SOC levels can effectively increase the battery's charge capacity. However, this strategy is complex to design and places high demands on the accuracy and reliability of the battery management system. Pulse charging uses intermittent current charging to reduce polarization effects, effectively avoiding the lithium plating side reaction. However, this method requires precise control of pulse parameters, otherwise it may lead to other negative effects. In recent years, with the rapid development of artificial intelligence (AI) technology, AI-based intelligent charging strategies have gradually emerged. These strategies significantly improve charging safety and efficiency by monitoring the battery's operating status in real time and dynamically adjusting charging parameters. However, these strategies are complex and rely on real-time data from high-precision sensors. Overall, each fast-charging strategy has its own advantages and disadvantages, requiring comprehensive optimization and selection based on the specific application scenario and battery characteristics to achieve safe, efficient, and long-life fast charging. Summary of the Invention

[0004] In response to the technical problems existing in the prior art, the present invention provides a method, device, equipment and medium for optimizing the fast charging strategy of a lithium-ion battery.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] In one aspect, the present invention provides a method for optimizing a lithium-ion battery fast charging strategy, comprising:

[0007] Considering the voltage and temperature characteristics of lithium-ion batteries during charging, an electrochemical-thermal coupling model of lithium-ion batteries was constructed;

[0008] Obtain electrochemical-thermal coupling models and simulated and measured voltages and temperatures of real lithium-ion batteries under different operating conditions;

[0009] Aiming to minimize the voltage and temperature errors between simulation and measurement, a differential evolution algorithm was used to identify the parameters of the electrochemical-thermal coupling model, and a parameterized electrochemical-thermal coupling model was obtained.

[0010] The parameterized electrochemical-thermal coupling model is used as the environment model in the reinforcement learning model to guide the generation of a multi-stage constant current charging strategy.

[0011] On the other hand, a device for optimizing a fast charging strategy for a lithium-ion battery is provided, comprising:

[0012] The first module is used to consider the voltage and temperature characteristics of lithium-ion batteries during charging and construct an electrochemical-thermal coupling model of lithium-ion batteries;

[0013] The second module is used to obtain the electrochemical-thermal coupling model and the simulated and measured voltage and temperature of real lithium-ion batteries under different operating conditions;

[0014] The third module is used to identify the parameters of the electrochemical-thermal coupling model using a differential evolution algorithm with the goal of minimizing the voltage and temperature errors between the simulation and the measured values, thereby obtaining a parameterized electrochemical-thermal coupling model.

[0015] The fourth module is used to use the parameterized electrochemical-thermal coupling model as the environment model in the reinforcement learning model to guide the generation of a multi-stage constant current charging strategy.

[0016] On the other hand, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned lithium-ion battery fast charging strategy optimization method are implemented.

[0017] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned lithium-ion battery fast charging strategy optimization method.

[0018] On the other hand, the present invention provides a computer program product, which is stored on a computer-readable storage medium and includes computer instructions, which, when executed by a processor, enable a computer device to implement the steps of the above-mentioned lithium-ion battery fast charging strategy optimization method.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0020] In order to effectively solve the battery life anxiety of lithium-ion batteries and ensure the safety and efficiency of the charging process, the present invention proposes a multi-stage constant current fast charging strategy optimization method based on reinforcement learning. The present invention first constructs an electrochemical-thermal coupling model that can accurately describe the dynamic behavior of lithium-ion batteries under different working conditions; then, the electrochemical-thermal coupling model and the simulated and measured voltage and temperature of the real lithium-ion battery under different working conditions are obtained. The working condition test includes open circuit voltage test, intermittent constant current charging test, etc., through these test conditions. With the goal of minimizing the error between simulation and measured voltage and temperature, the differential evolution algorithm is used to accurately identify the unknown parameters in the electrochemical-thermal coupling model. Through this process, the parameterized electrochemical-thermal coupling model can accurately reflect the actual behavior of lithium-ion batteries under different working conditions. Subsequently, the parameterized electrochemical-thermal coupling model is used as the environment (Environment) in the reinforcement learning framework to guide the generation of a multi-stage constant current charging strategy. During reinforcement learning model training, the environment model interacts with the agent model to simulate and provide the state parameter characteristics (including SOC, voltage, temperature, and negative electrode potential) of the lithium-ion battery under different charging strategies and charging states. These key state parameter characteristics of the lithium-ion battery are used to calculate the score of the current charging strategy to guide the generation of subsequent charging strategies. In this way, the reinforcement learning model can continuously explore and optimize charging strategies to achieve the goal of safe, efficient and fast charging.

[0021] The present invention constructs a parameterized electrochemical-thermal coupling model as an environmental model in reinforcement learning, which can accurately simulate the evolution of the voltage, temperature, negative electrode potential and other states of lithium-ion batteries under different charging strategies. This simulation result provides guidance for the agent model to generate a safe, lossless and efficient multi-stage constant current charging strategy. The method of the present invention is suitable for the fast charging optimization of lithium-ion batteries under different temperature scenarios, and can adapt to lithium-ion batteries of different types and different aging states by adjusting the parameters of the electrochemical-thermal coupling model. In addition, the trained agent model can quickly generate the corresponding optimal multi-stage constant current charging curve according to the initial state of the battery, with low computational consumption, and can meet the computing power requirements of the existing battery management system. Therefore, the present method has strong practicality and can carry out safe and efficient charging strategy optimization work for lithium-ion batteries of different temperature scenarios, different types and different aging states in the existing battery management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0023] Figure 1 Flowchart of a method for optimizing a lithium-ion battery fast charging strategy in one embodiment;

[0024] Figure 2 A schematic diagram of the interaction relationship between the environment model and the agent model in a reinforcement learning model in one embodiment;

[0025] Figure 3 FIG. 1 is a flowchart of a multi-stage constant current charging optimization process for a battery based on reinforcement learning in one embodiment. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0027] In one embodiment, a method for optimizing a lithium-ion battery fast charging strategy is provided, comprising:

[0028] Considering the voltage and temperature characteristics of lithium-ion batteries during charging, an electrochemical-thermal coupling model of lithium-ion batteries was constructed;

[0029] Obtain electrochemical-thermal coupling models and simulated and measured voltages and temperatures of real lithium-ion batteries under different operating conditions;

[0030] Aiming to minimize the voltage and temperature errors between simulation and measurement, a differential evolution algorithm was used to identify the parameters of the electrochemical-thermal coupling model, and a parameterized electrochemical-thermal coupling model was obtained.

[0031] The parameterized electrochemical-thermal coupling model is used as the environment model in the reinforcement learning model to guide the generation of a multi-stage constant current charging strategy.

[0032] The electrochemical-thermal coupling model of the lithium-ion battery includes an electrochemical model and a thermal model of the lithium-ion battery, wherein the electrochemical model includes multiple sets of mutually coupled mathematical models for describing the lithium ion solid-phase diffusion, liquid-phase diffusion, lithium ion solid-phase potential distribution, lithium ion liquid-phase potential distribution, and solid-liquid interface reaction processes inside the lithium-ion battery.

[0033] The lithium ion solid phase diffusion model and lithium ion liquid phase diffusion model used to describe the lithium ion solid phase diffusion and liquid phase diffusion inside the lithium ion battery are as follows:

[0034] ;

[0035] Where, is the solid phase lithium ion concentration, Ds is the solid phase lithium ion diffusion coefficient, is the radial coordinate inside the solid particle, ranging from (center of particle) to (particle surface radius);

[0036] ;

[0037] In the formula, the first term on the right side of the equation is the lithium ion diffusion process, which is driven by the liquid phase concentration gradient; the second term is the lithium ion migration process, which is driven by the liquid phase potential difference. is the liquid phase lithium ion concentration, is the liquid volume fraction (i.e. porosity), is the equivalent diffusion coefficient of lithium ions in the liquid phase, is the lithium ion migration number, is the surface density of liquid phase current, is the Faraday constant, which is approximately j is the reaction current density at the solid-liquid interface, and its unit is usually , the lithium ion adsorption and desorption process on the surface of the electrode particles is coupled to the two-phase diffusion equation through j. It is the spatial coordinate of the electrode on the macro scale, that is, the distance coordinate from the current collector side to the separator side (or from the negative terminal to the positive terminal) along the thickness direction of the electrode.

[0038] In the lithium ion solid phase diffusion model, Used to characterize the concentration gradient inside a single particle.

[0039] In the lithium ion liquid phase diffusion model, Used to characterize the concentration gradient and potential change of the electrolyte in the electrode; j is the coupled interface reaction flux between the two phases, which brings in the Faraday constant To ensure that the material flux is consistent with the current.

[0040] The lithium ion solid phase potential distribution model and the lithium ion liquid phase potential distribution model are as follows:

[0041] ;

[0042] Where, is the solid phase volume fraction, is the solid phase potential, is the solid phase current surface density, is the equivalent lithium ion conductivity of the solid phase.

[0043] Liquid phase potential Satisfies the following formula:

[0044] ;

[0045] Where R is the ideal gas constant, is the battery thermodynamic temperature, is the equivalent ionic conductivity of the electrolyte, is the lithium ion migration number, is the surface density of liquid phase current, is the Faraday constant, which is approximately .

[0046] The solid-liquid interface reaction equation is as follows:

[0047] ;

[0048] Where, It is the electrochemical reaction current per unit area at the solid-liquid interface. and It is the transfer coefficient of the electrochemical reaction. For lithium ion insertion or deinsertion electrochemical reactions, it is usually taken as 0.5. It is the exchange current density of lithium ion insertion or deinsertion reaction, which is related to the lithium ion concentration in the solid and liquid phases. is the total area of ​​the solid-liquid interface per unit electrode volume, usually in units of (or simply ), coupled with the current density multiplied by the area, is used to convert the local reaction current density Expanded to overall current. The activation overpotential is the difference between the actual electrode potential and the equilibrium electrode potential.

[0049] The thermal model uses a lumped thermal model, treating the lithium-ion battery as a single mass point, ignoring the temperature conduction process inside it, and only considering the heat generation of the lithium-ion battery and the heat exchange with the environment. The thermal model includes a battery heat generation model, a battery heat exchange model, and a battery temperature change model.

[0050] The battery temperature change model is as follows:

[0051] ;

[0052] Where T is the battery temperature, C is the specific heat capacity of the battery, m is the mass of the battery, q is the heat generation power of the battery, and It represents the heat exchange power between the battery and the environment.

[0053] The battery heat generation model is as follows:

[0054] ;

[0055] Where I and V are the battery current and terminal voltage, respectively, where current I is negative during charging and positive during discharging. is the open circuit voltage (OCV) of the battery, is the entropy thermal coefficient of the battery.

[0056] The heat exchange between the battery and the environment is mainly manifested as a typical thermal convection process, which can be described by Newton's law of cooling. The battery heat exchange model is as follows:

[0057] ;

[0058] in is the convection heat transfer coefficient, A is the contact area between the battery and the environment, is the ambient temperature.

[0059] Table 1 shows the parameters to be identified in this embodiment:

[0060] Table 1

[0061]

[0062] The electrochemical-thermal coupling model of the lithium-ion battery describes the key physical and chemical processes such as solid-phase diffusion, liquid-phase diffusion and electrochemical reaction inside the lithium-ion battery through multiple sets of mutually coupled mathematical equations, and can accurately characterize the battery's state parameters such as voltage and negative electrode potential. The thermal model uses a lumped thermal model, which regards the battery as a single particle, ignores the temperature conduction process inside it, and only considers the heat generation of the battery and the heat exchange with the environment. While meeting the battery temperature monitoring accuracy requirements in fast charging optimization, this lumped model significantly improves the efficiency of simulation calculations due to its simple structure. In addition, in the electrochemical model, processes such as ion solid-phase diffusion, liquid-phase diffusion and electrochemical reaction are all affected by temperature. To this end, based on the Arrhenius formula, the relationship between the above-mentioned physical and chemical processes and temperature is characterized, thereby achieving effective coupling of the electrochemical model and the thermal model.

[0063] Based on the electrochemical-thermal coupling model of the lithium-ion battery, three working condition test simulations including open circuit voltage test, constant current charging test and intermittent constant current charging test are carried out to obtain the simulated voltage of the lithium-ion battery in different working condition test simulations. V sim and simulation temperature T sim Based on real lithium-ion batteries, three working condition tests are carried out: open circuit voltage test, constant current charge test and intermittent constant current charge test, to obtain the real test voltage of real lithium-ion batteries in different working condition tests. V test and test temperature T test The open circuit voltage test uses a low charge rate of 0.05C to charge the lithium-ion battery from 0% SOC to the upper cutoff voltage. The constant current charge test uses a charge rate of 1C to charge the lithium-ion battery from 0% SOC to the upper cutoff voltage. The intermittent constant current charge test charges the lithium-ion battery at a rate of 1C to 20% SOC, rests for 30 minutes, and repeats this cycle until the upper cutoff voltage is reached. By adding a rest period, this operating condition can enhance the influence of ion solid-phase diffusion and liquid-phase diffusion on voltage characteristics, thereby improving the identifiability of related parameters.

[0064] With the goal of minimizing the voltage and temperature errors between simulation and measurement, a differential evolution algorithm is used to identify the parameters of the electrochemical-thermal coupling model, including:

[0065] Based on the actual test voltage of real lithium-ion batteries in different working conditions V test and test temperature T test The maximum and minimum values ​​of the test voltage, test temperature, and simulation voltage and temperature are normalized. This step is to eliminate the impact of the dimensional differences in voltage and temperature characteristics on subsequent error calculations.

[0066] Test voltage V test Taking normalization as an example, the normalized test voltage ,in 、 are the maximum and minimum values ​​of the test voltage respectively. Similarly, the normalized test temperature can be obtained , normalized simulation voltage , normalized simulation temperature .

[0067] An objective function is constructed with the goal of minimizing the error between the test voltage and the simulation voltage, as well as the error between the test temperature and the simulation temperature. The differential evolution algorithm is used to identify the unknown parameters in the electrochemical-thermal coupling model, and the parameterized electrochemical-thermal coupling model is obtained when the objective function meets the set requirements.

[0068] The objective function is:

[0069] ;

[0070] in: n is the total amount of test data samples, that is, the total amount of simulation data samples. Each test data sample corresponds to a test temperature and a test voltage under a test condition, and each simulation data sample corresponds to a simulation temperature and a simulation voltage under a simulation test condition.

[0071] Reference Figure 2 Figure 2 illustrates the interaction between the environment model and the agent model in a reinforcement learning model. During reinforcement learning model training, the environment model interacts with the agent model to simulate and provide key parameter characteristics (including SOC, voltage, temperature, and cathode potential) of lithium-ion batteries under different charging strategies and charging states. These key parameter characteristics are used to calculate the current charging strategy score, which guides the generation of subsequent charging strategies.

[0072] The parameterized electrochemical-thermal coupling model is used as the environment model in the reinforcement learning model to guide the generation of multi-stage constant current charging strategy. Figure 3 , which is a flow chart for optimizing multi-stage constant current charging of batteries based on reinforcement learning. The entire process includes:

[0073] Step (1) Initialization parameters, including initialization of agent model parameters, initial state parameters of lithium-ion batteries and strategy scores. The state parameters of lithium-ion batteries include state of charge, temperature, voltage and negative electrode potential of lithium-ion batteries; initial temperature T The value range can be set according to the actual charging scenario requirements of the battery, for example, within the range of [-20℃, 40℃];

[0074] Step (2) The agent model generates a stage charging strategy based on the current state parameters of the lithium-ion battery and the strategy score I n and V n ,in I n is the constant current charging current in the current stage, V n It is the cut-off voltage of constant current charging in the current stage;

[0075] Step (3) The stage charging strategy I n and V n Input into the environmental model for simulation and update the current state parameters of the lithium-ion battery (including SOC, temperature T ,Voltage V , negative electrode potential , charging time t c ), and calculate the corresponding strategy score based on the current state parameters of the updated lithium-ion battery ;

[0076] ;

[0077] The first term to the right of the equal sign in the equation represents the battery's charging rate. The second term constrains the battery temperature during charging, ensuring it does not exceed 50°C to prevent accelerated aging. The third term constrains lithium deposition during charging, ensuring the negative electrode potential remains above 0V to prevent lithium deposition. The higher the strategy score, the better.

[0078] Step (4) determines whether the set charging cutoff condition is met based on the battery state of charge in the updated current state parameters of the lithium-ion battery. If so, the current multi-stage constant current charging strategy is output, and the corresponding strategy score is calculated based on the updated current state parameters of the lithium-ion battery to update the parameters of the agent model. If not, the process returns to step (2). The charging cutoff condition can be determined based on actual conditions or empirical values, such as if the battery state of charge in the current state parameters is higher than a preset charging cutoff power value (e.g., 90%).

[0079] Calculate the corresponding strategy score based on the current state parameters of the updated lithium-ion battery, and use the REINFORCE algorithm to adjust the parameters of the agent model. When updating the parameters, the output of the agent model is first considered as a set of parameterized strategies. , then perform gradient estimation and define the objective function of the strategy as , whose gradient can be given by the following formula:

[0080] ;

[0081] in , is the number of stages, That is, the corresponding strategy score is calculated based on the current state parameters of the updated lithium-ion battery , in a single episode, use sample estimation for gradient ascent:

[0082] ;

[0083] in is the learning rate, is a baseline value, which can be taken as the average return of historical episodes to reduce the variance of the gradient estimate.

[0084] The present invention can train the proxy model according to the above steps by setting the number of training iterations of the reinforcement learning model. After the training is completed, the proxy model can generate the optimal multi-stage constant current charging curve under the current state by inputting the initial state of the battery. Since the constraints on temperature rise and lithium precipitation reaction during battery charging are added to the strategy scoring model, the charging strategy generated by the proxy model can effectively ensure the safety of the battery during the charging process and slow down the aging rate of the battery. In addition, since the initial battery temperature accepted by the proxy model during the training process is randomly generated from the set value range, the proxy model can generate the optimal, safe and fast multi-stage constant current charging curve for any temperature point within the value range, realizing the strategy generation requirements of the model under a wide temperature range.

[0085] In summary, the present invention has the following three major innovations in model design and optimization path: First, the existing technology is based on traditional single-particle or simplified P2D models. Starting from the microscopic electrode structure, the present invention completely solves the spatiotemporal distribution of the solid phase and electrolyte in the thickness direction, and realizes deep electro-thermal coupling through temperature-dependent diffusion coefficient, conductivity and reaction kinetic rate, thereby capturing local concentration, temperature and overpotential gradients; secondly, the differential evolution algorithm is used to perform global parameter identification of the voltage-current-temperature curve under a wide temperature range and intermittent / constant current conditions to ensure that the model maintains the lowest error under any operating condition; finally, the high-fidelity model is embedded in the reinforcement learning framework to realize adaptive optimization of the charging strategy, and dynamically adjust the current and cut-off voltage according to voltage and temperature feedback, so as to achieve "learning and adjusting", thereby improving the model's precision, robustness and optimization capability.

[0086] In one embodiment, a device for optimizing a fast-charging strategy for a lithium-ion battery is provided, comprising:

[0087] The first module is used to consider the voltage and temperature characteristics of lithium-ion batteries during charging and construct an electrochemical-thermal coupling model of lithium-ion batteries;

[0088] The second module is used to obtain the electrochemical-thermal coupling model and the simulated and measured voltage and temperature of real lithium-ion batteries under different operating conditions;

[0089] The third module is used to identify the parameters of the electrochemical-thermal coupling model using a differential evolution algorithm with the goal of minimizing the voltage and temperature errors between the simulation and the measured values, thereby obtaining a parameterized electrochemical-thermal coupling model.

[0090] The fourth module is used to use the parameterized electrochemical-thermal coupling model as the environment model in the reinforcement learning model to guide the generation of a multi-stage constant current charging strategy.

[0091] On the other hand, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the lithium-ion battery fast charging strategy optimization method provided in any of the above embodiments are implemented. The computer device may be a server. The computer device comprises a processor, a memory, a network interface, and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store sample data. The network interface of the computer device is used to communicate with an external terminal via a network connection.

[0092] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the lithium-ion battery fast charging strategy optimization method provided in any of the above embodiments.

[0093] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0094] Matters not covered by the present invention are known technologies.

[0095] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0096] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the spirit of the present application, and such modifications and improvements are all within the scope of protection of the present application.

[0097] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A lithium-ion battery fast charging strategy optimization method, characterized in that: include: Considering the voltage and temperature characteristics of the lithium-ion battery during charging, an electrochemical-thermal coupling model of the lithium-ion battery is constructed. The electrochemical-thermal coupling model of the lithium-ion battery includes an electrochemical model and a thermal model of the lithium-ion battery. The electrochemical model includes multiple sets of mutually coupled mathematical models for describing the lithium-ion solid-phase diffusion, liquid-phase diffusion, lithium-ion solid-phase potential distribution, lithium-ion liquid-phase potential distribution, and solid-liquid interface reaction processes within the lithium-ion battery. The thermal model uses a lumped thermal model, treating the lithium-ion battery as a single particle, ignoring the temperature conduction process within it, and only considering the heat generation of the lithium-ion battery and the heat exchange with the environment. The thermal model includes a battery heat generation model, a battery heat exchange model, and a battery temperature change model. Obtain electrochemical-thermal coupling models and simulated and measured voltages and temperatures of real lithium-ion batteries under different operating conditions; Aiming to minimize the voltage and temperature errors between simulation and measurement, a differential evolution algorithm was used to identify the parameters of the electrochemical-thermal coupling model, and a parameterized electrochemical-thermal coupling model was obtained. The parameterized electrochemical-thermal coupling model is used as the environment model in the reinforcement learning model to guide the generation of a multi-stage constant current charging strategy, including: Step (1) initializing parameters, including initializing agent model parameters, initial state parameters of the lithium-ion battery, and strategy scores. The state parameters of the lithium-ion battery include the state of charge, temperature, voltage, and negative electrode potential of the lithium-ion battery; Step (2) The agent model generates a stage charging strategy based on the current state parameters of the lithium-ion battery and the strategy score I n and V n ,in I n is the constant current charging current in the current stage, V n It is the cut-off voltage of constant current charging in the current stage; Step (3) The stage charging strategy I n and V n Input the data into the environment model for simulation, update the current state parameters of the lithium-ion battery, and calculate the corresponding strategy score based on the updated current state parameters of the lithium-ion battery; Step (4) determines whether the set charging cut-off condition is met based on the battery state of charge in the updated current state parameters of the lithium-ion battery. If so, the current multi-stage constant current charging strategy is output, and the corresponding strategy score is calculated based on the updated current state parameters of the lithium-ion battery to update the parameters of the agent model. If not, return to step (2).

2. The lithium-ion battery fast charging strategy optimization method according to claim 1, characterized in that: Based on the electrochemical-thermal coupling model of the lithium-ion battery, three operating condition test simulations are carried out, namely, open circuit voltage test, constant current charge test and intermittent constant current charge test, to obtain the simulated voltage and simulated temperature of the lithium-ion battery in the different operating condition test simulations; based on the real lithium-ion battery, three operating condition tests are carried out, namely, open circuit voltage test, constant current charge test and intermittent constant current charge test, to obtain the real test voltage and test temperature of the real lithium-ion battery in the different operating condition tests.

3. The lithium-ion battery fast charging strategy optimization method according to claim 2, characterized in that: The open circuit voltage test uses a small rate of 0.05C to charge the lithium-ion battery from 0% SOC to the upper cut-off voltage; the constant current charge test uses a rate of 1C to charge the lithium-ion battery from 0% SOC to the upper cut-off voltage; The intermittent constant current charging test was to charge the battery to 20% SOC at a 1C rate, let it rest for 30 minutes, and repeat this cycle until the battery reached the upper cutoff voltage.

4. The lithium-ion battery fast charging strategy optimization method according to claim 2 or 3, characterized in that: With the goal of minimizing the voltage and temperature errors between simulation and measurement, a differential evolution algorithm is used to identify the parameters of the electrochemical-thermal coupling model, including: Based on the actual test voltage of real lithium-ion batteries in different working conditions V test and test temperature T test The maximum and minimum values ​​of the test voltage, test temperature, simulation voltage and simulation temperature are normalized; An objective function is constructed with the goal of minimizing the error between the test voltage and the simulation voltage, as well as the error between the test temperature and the simulation temperature. The differential evolution algorithm is used to identify the unknown parameters in the electrochemical-thermal coupling model, and the parameterized electrochemical-thermal coupling model is obtained when the objective function meets the set requirements.

5. A lithium-ion battery fast charging strategy optimization device, characterized in that: include: The first module is used to consider the voltage and temperature characteristics of the lithium-ion battery during charging and construct an electrochemical-thermal coupling model of the lithium-ion battery. The electrochemical-thermal coupling model of the lithium-ion battery includes an electrochemical model and a thermal model of the lithium-ion battery. The electrochemical model includes multiple sets of mutually coupled mathematical models for describing the lithium-ion solid-phase diffusion, liquid-phase diffusion, lithium-ion solid-phase potential distribution, lithium-ion liquid-phase potential distribution, and solid-liquid interface reaction processes within the lithium-ion battery. The thermal model uses a lumped thermal model, treating the lithium-ion battery as a single particle, ignoring the temperature conduction process within it, and only considering the heat generation of the lithium-ion battery and the heat exchange with the environment. The thermal model includes a battery heat generation model, a battery heat exchange model, and a battery temperature change model. The second module is used to obtain the electrochemical-thermal coupling model and the simulated and measured voltage and temperature of real lithium-ion batteries under different operating conditions; The third module is used to identify the parameters of the electrochemical-thermal coupling model using a differential evolution algorithm with the goal of minimizing the voltage and temperature errors between the simulation and the measured values, thereby obtaining a parameterized electrochemical-thermal coupling model. The fourth module is used to use the parameterized electrochemical-thermal coupling model as the environment model in the reinforcement learning model to guide the generation of a multi-stage constant current charging strategy, including: Step (1) initializing parameters, including initializing agent model parameters, initial state parameters of the lithium-ion battery, and strategy scores. The state parameters of the lithium-ion battery include the state of charge, temperature, voltage, and negative electrode potential of the lithium-ion battery; Step (2) The agent model generates a stage charging strategy based on the current state parameters of the lithium-ion battery and the strategy score I n and V n ,in I n is the constant current charging current in the current stage, V n It is the cut-off voltage of constant current charging in the current stage; Step (3) The stage charging strategy I n and V n Input the data into the environment model for simulation, update the current state parameters of the lithium-ion battery, and calculate the corresponding strategy score based on the updated current state parameters of the lithium-ion battery; Step (4) determines whether the set charging cut-off condition is met based on the battery state of charge in the updated current state parameters of the lithium-ion battery. If so, the current multi-stage constant current charging strategy is output, and the corresponding strategy score is calculated based on the updated current state parameters of the lithium-ion battery to update the parameters of the agent model. If not, return to step (2).

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the lithium-ion battery fast charging strategy optimization method as described in claim 1 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the lithium-ion battery fast charging strategy optimization method as claimed in claim 1 are implemented.

Images