Battery offline and online hybrid modeling method and device, storage medium and product

Abstract

The invention discloses a battery off-line and on-line hybrid modeling method and device, a storage medium and a product. The method comprises the following steps: constructing a state dependent model of a battery; determining a state vector according to the state parameter of the battery; taking the state vector as input, and determining a function type coefficient of the state dependent model; converting the state-dependent model into a nonlinear least square model; according to the obtained battery measurement data, performing offline optimization on parameters of the nonlinear least square model by adopting an optimization algorithm to obtain all parameters, and further obtaining a battery offline model; and keeping the non-linear parameters of the battery offline model unchanged, and updating the linear parameters in the battery offline model online according to the real-time measurement data of the battery so as to obtain a real-time updated state dependent model. According to the method, high-precision modeling of the lithium ion battery is realized on the premise of ensuring real-time performance.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery management, and in particular relates to a battery offline-online hybrid modeling method, device, storage medium and product based on a state-dependent model. Background Art

[0002] With the rapid development of modern industry, mobile phones, computers, electric vehicles, etc. have been applied to all aspects of human life, greatly improving people's quality of life. Among them, lithium-ion batteries with advantages such as high energy density and long cycle life are the main energy storage units of these devices. Effective management of them is the prerequisite for the safe and economical application of related equipment. However, lithium-ion batteries are a type of complex nonlinear time-varying system, and their internal state is time-varying and unobservable. Therefore, constructing a mathematical model that can accurately describe the dynamic characteristics of lithium-ion batteries and meet the real-time requirements of actual engineering applications is crucial to realize online estimation of battery status and charge and discharge control.

[0003] At present, the modeling problems of lithium-ion batteries mainly include mechanism modeling and data-driven modeling. The equivalent circuit model that uses ideal circuit elements to simulate the dynamic process of the battery has been widely used in engineering due to its simple model structure, but this type of model is usually difficult to fully characterize the nonlinear dynamic time-varying characteristics of lithium-ion batteries under different actual operating conditions (such as battery temperature, aging degree, charge and discharge rate, etc.). The battery model based on the electrochemical mechanism description has a higher model accuracy in theory, but due to the existence of high-order partial differential equations, the parameter identification process is complicated and it is difficult to ensure real-time performance in actual engineering applications. The data-driven modeling method describes the dynamic behavior of the battery by directly constructing a nonlinear mapping relationship between the battery state and the battery measurable, but this type of model usually has a complex structure and is difficult to achieve online learning, while offline identification requires a large amount of historical data to support the accuracy of the model, and this type of modeling method is difficult to achieve a closed loop of state estimation and charge and discharge control. Summary of the invention

[0004] The purpose of the present invention is to provide a battery offline-online hybrid modeling method, device, storage medium and product to solve the problems that the traditional method has a complex parameter identification process, is difficult to achieve online learning, and is difficult to meet real-time requirements.

[0005] The present invention solves the above technical problems through the following technical solutions: a battery offline and online hybrid modeling method, comprising:

[0006] Construct a state-dependent model of the battery;

[0007] Determine a state vector of the battery according to a state parameter of the battery;

[0008] Determining functional coefficients of the state-dependent model using the state vector as input;

[0009] According to the functional coefficients, converting the state-dependent model into a nonlinear least squares model;

[0010] According to the acquired battery measurement data, an optimization algorithm is used to perform offline optimization on the parameters of the nonlinear least squares model to obtain all parameters, and then obtain a battery offline model;

[0011] The nonlinear parameters of the battery offline model are kept unchanged, and the linear parameters in the battery offline model are updated online according to the real-time measurement data of the battery, thereby obtaining a state-dependent model updated in real time.

[0012] Furthermore, the state dependency model is:

[0013] ;

[0014] in, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model, Both represent the state vector The functional coefficients of Indicates that the battery is The state vector at time, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is Current at the moment.

[0015] Furthermore, the state vector of the battery is:

[0016] ;

[0017] in, Indicates that the battery is The state vector at time, Indicates that the battery is The terminal voltage at the moment, represents the regression order of the terminal voltage, Indicates that the battery is The current at the moment, represents the regression order of the current, Indicates that the battery is The state of charge at the moment, Indicates that the battery is Health status at all times, Indicates that the battery is The temperature of the moment.

[0018] Furthermore, a radial basis function neural network is used to determine the functional coefficients of the state-dependent model, and the specific expression of the functional coefficients is:

[0019] ;

[0020] ;

[0021] ;

[0022] in, Both represent the state vector The functional coefficients of represents the number of neurons in the radial basis function neural network, , , , , , Both represent the linear weights of the radial basis function neural network, , Both represent the scale factor of the radial basis function neural network, , Both represent the center vector of the radial basis function neural network, represents the 2-norm.

[0023] Furthermore, the nonlinear least squares model is:

[0024] ;

[0025] ;

[0026] ;

[0027] in, Indicates that the battery is The terminal voltage at the moment, , Both represent intermediate quantities. represents the nonlinear parameters in the functional coefficients, Indicates that the battery is The state vector at time, Indicates the linear parameter in the function coefficient, the superscript represents transpose, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model.

[0028] Furthermore, the parameters of the nonlinear least squares model are optimized offline using a structured nonlinear parameter optimization algorithm, specifically including:

[0029] Step A1: Construct the optimization objective function, the specific expression is:

[0030] ;

[0031] ;

[0032] ;

[0033] in, represents the optimization target value, Represents the error vector between the predicted value and the true value of the battery terminal voltage, express The maximum value in represents the nonlinear parameters in the functional coefficients, represents the linear parameter in the functional coefficient, represents the 2-norm, , Both represent the order of the state-dependent model, represents the regression order of the terminal voltage, represents the regression order of the current, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is The terminal voltage at the moment, represents the terminal voltage of the battery at time M, represents the predicted value of the terminal voltage, represents the state vector, M represents the number of sampling points of the data used for parameter optimization;

[0034] Step A2: Initializing the nonlinear parameters and linear parameters in the functional coefficients;

[0035] Step A3: According to the initialized nonlinear parameters and the optimization objective function, the Levenberg-Marquardt method is used to calculate the nonlinear parameters. The specific calculation formula is:

[0036] ;

[0037] ;

[0038] ;

[0039] in, Indicates The nonlinear parameters of the iteration; Indicates The nonlinear parameters of the iteration; represents the step length; represents the Jacobian matrix; Indicates the intermediate quantity, For control The direction and size of Approaching 0, Along the Gauss-Newton direction, when As it approaches infinity, It will be in the direction of steepest descent; represents the identity matrix, Indicates The linear parameters of the iterations, Indicates the symbol for partial derivative; Represents the error vector between the predicted value and the true value of the battery terminal voltage; represents transpose;

[0040] Step A4: According to the nonlinear parameters and the nonlinear least squares model, the linear parameters are calculated using the least squares method. The specific calculation formula is:

[0041] ;

[0042] ;

[0043] in, Indicates The linear parameter of the iteration, superscript represents the inverse matrix, Indicates that the battery is The terminal voltage at the moment, represents the amount of intermediate calculation, express The state vector at the moment;

[0044] Step A5: Determine whether the optimization termination condition is met. If so, output the optimized nonlinear parameters and linear parameters to obtain the battery offline model; if not, repeat steps A3 and A4 until the optimization termination condition is met.

[0045] Furthermore, the recursive least square method with a forgetting factor is used to update the linear parameters in the battery offline model online. The specific update formula is:

[0046] ;

[0047] ;

[0048] ;

[0049] ;

[0050] ;

[0051] in, express The gain matrix at time instant, express The covariance matrix at time , , , Both represent intermediate quantities. represents the nonlinear parameters in the functional coefficients, Indicates that the battery is The state vector at time, represents the forgetting factor, express The updated value of the linear parameter at time , express The updated value of the linear parameter at time , Indicates that the battery is The terminal voltage at the moment, express The covariance matrix at time , represents the identity matrix, the superscript represents transpose, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model, represents the number of neurons in the radial basis function neural network, represents the scale factor of the radial basis function neural network, represents the center vector of the radial basis function neural network, represents the 2-norm.

[0052] Based on the same concept, the present invention also provides an electronic device, including a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the battery offline-online hybrid modeling method as described above.

[0053] Based on the same concept, the present invention also provides a computer-readable storage medium having a computer program / instruction stored thereon, which implements the above-mentioned battery offline-online hybrid modeling method when executed by a processor.

[0054] Based on the same concept, the present invention also provides a computer program product, including a computer program / instruction, which implements the above-mentioned battery offline-online hybrid modeling method when executed by a processor. Beneficial Effects

[0055] Compared with the prior art, the advantages of the present invention are:

[0056] The present invention constructs a battery model based on a state-dependent model, without the need to understand the internal mechanism of the battery, and effectively avoids a complex multi-step parameter identification process; by embedding the state vector of the battery, it effectively matches the multi-factor coupled nonlinear characteristics of the battery; through an offline online modeling strategy, on the basis of effectively learning the dynamic characteristics of the battery, it can follow the parameter changes of the battery caused by factors such as temperature and operating conditions in real time, thereby achieving efficient and high-precision modeling of the battery, facilitating state estimation, charge and discharge control and other management based on the constructed battery model, and having high theoretical value and application prospects. The present invention achieves high-precision modeling of lithium-ion batteries while ensuring real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 This is a flow chart of a battery offline and online hybrid modeling method according to an embodiment of the present invention;

[0059] Figure 2 is a comparison curve between the terminal voltage prediction value and the actual value in the embodiment of the present invention;

[0060] Figure 3 is a terminal voltage error curve in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The following is a clear and complete description of the technical solutions in the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] The technical solution of the present application is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0063] Example 1

[0064] Figure 1 The flow chart of the battery offline and online hybrid modeling method provided by the present invention is shown, and this embodiment takes lithium-ion batteries as an example. Figure 1 As shown, the modeling method includes two steps: offline modeling and online updating, wherein the offline modeling part includes:

[0065] Step 1: Construct a state-dependent model of the battery.

[0066] In this embodiment, the state dependency model is:

[0067] (1)

[0068] in, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model, Both represent the state vector The functional coefficients of Indicates that the battery is The state vector at time, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is Current at the moment.

[0069] Step 2: Determine the state vector of the battery according to the state parameters of the battery.

[0070] Considering that lithium-ion batteries are affected by state parameters such as temperature, state of charge, state of health, current and terminal voltage, the state vector of the battery is:

[0071] (2)

[0072] in, Indicates that the battery is The terminal voltage at the moment, represents the regression order of the terminal voltage, represents the regression order of the current, Indicates that the battery is The state of charge at the moment, Indicates that the battery is Health status at all times, Indicates that the battery is The temperature of the moment.

[0073] According to different modeling requirements, different regression orders can be selected to consider the time-dependent characteristics of lithium-ion batteries. , When , the state vector of the battery is:

[0074] (3)

[0075] Step 3: Take the battery state vector As input, the functional coefficients of the state-dependent model are determined.

[0076] In a specific embodiment of the present invention, a radial basis function neural network is used to determine the functional coefficients of the state-dependent model. The specific expression of the functional coefficients is:

[0077] (4)

[0078] (5)

[0079] (6)

[0080] in, Both represent the state vector The functional coefficients of represents the number of neurons in the radial basis function neural network, , , , , , Both represent the linear weights of the radial basis function neural network, , Both represent the scale factor of the radial basis function neural network, , Both represent the center vector of the radial basis function neural network, represents the 2-norm. Linear weights, scaling factors, and center vectors are all parameters that need to be optimized in the state-dependent model.

[0081] Step 4: According to the functional coefficients, the state-dependent model is converted into a nonlinear least squares model.

[0082] The parameters in the functional coefficients are divided into linear parameters and nonlinear parameters, so the state-dependent model can be converted into a nonlinear least squares model. The specific expression is:

[0083] (7)

[0084] ; (8)

[0085] (9)

[0086] (10)

[0087] in, , Both represent intermediate quantities. represents the nonlinear parameters in the functional coefficients, Indicates the linear parameter in the function coefficient, the superscript represents transpose, , Both represent the order of the state-dependent model.

[0088] Step 5: Based on the acquired battery measurement data, an optimization algorithm is used to perform offline optimization on the parameters of the nonlinear least squares model to obtain all parameters of the nonlinear least squares model, and then obtain the battery offline model.

[0089] The battery measurement data is obtained through dynamic working condition testing, including the battery terminal voltage, current, state of charge, temperature, etc., and all parameters of the nonlinear least squares model are optimized offline using a structured nonlinear parameter optimization algorithm, so that it can learn the dynamic nonlinear characteristics of lithium-ion batteries. In a specific embodiment of the present invention, the parameters of the nonlinear least squares model are optimized offline using a structured nonlinear parameter optimization algorithm, specifically including:

[0090] Step 5.1: Construct the optimization objective function. The specific expression is:

[0091] (11)

[0092] (12)

[0093] (13)

[0094] in, represents the optimization target value, Represents the error vector between the predicted value and the true value of the battery terminal voltage, express The maximum value in represents the nonlinear parameters in the functional coefficients, represents the linear parameter in the functional coefficient, represents the 2-norm, , Both represent the order of the state-dependent model, represents the regression order of the terminal voltage, represents the regression order of the current, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is The terminal voltage at the moment, represents the terminal voltage of the battery at time M, Represents the predicted value of terminal voltage, represents the state vector, Represents the number of sampling points of the data used for parameter optimization.

[0095] Step 5.2: Nonlinear parameters in functional coefficients and linear parameters Initialize.

[0096] In this embodiment, random generation is used to initialize the nonlinear parameters. , and then initialize the nonlinear parameters Substituting into the nonlinear least squares model, we get the linear parameters The initial value of the linear parameter initialization.

[0097] Step 5.3: According to the initialized nonlinear parameters and the optimization objective function, the Levenberg-Marquitt method is used to calculate the nonlinear parameters. The specific calculation formula is:

[0098] (14)

[0099] (15)

[0100] (16)

[0101] in, Indicates The nonlinear parameters of the iteration; Indicates The nonlinear parameters of the iteration; represents the step length; represents the Jacobian matrix; represents the intermediate quantity, calculated according to formula (15), For control The direction and size of Approaching 0, Along the Gauss-Newton direction, when As it approaches infinity, It will be in the direction of steepest descent; represents the identity matrix, Indicates The linear parameters of the iterations, Indicates the symbol for partial derivative; Represents the error vector between the predicted value and the true value of the battery terminal voltage; Indicates transpose.

[0102] Step 5.4: According to the nonlinear parameters and the nonlinear least squares model, the least squares method is used to calculate the linear parameters. The specific calculation formula is:

[0103] (17)

[0104] (18)

[0105] in, Indicates The linear parameters of the iterations, express The maximum value in represents the inverse matrix, represents the number of sampling points of the data used for parameter optimization, Indicates that the battery is The terminal voltage at the moment, represents the amount of intermediate calculation, Indicates The nonlinear parameters of the iterations, express The state vector at time.

[0106] Step 5.5: Determine whether the optimization termination condition is met. If so, output the optimized nonlinear parameters and linear parameters, that is, obtain the battery offline model; if not, repeat steps 5.3 and 5.4 until the optimization termination condition is met.

[0107] In this embodiment, the optimization termination condition is that the number of iterations reaches a set number, or the optimization target value is less than a set error.

[0108] The construction of the battery offline model is completed through steps 1 to 5, that is, the state-dependent model of the battery after offline optimization is obtained. On this basis, online updates are performed, which include:

[0109] Step 6: Keep the nonlinear parameters of the battery offline model unchanged, and update the linear parameters in the battery offline model online according to the real-time measurement data of the battery.

[0110] The real-time measurement data includes the terminal voltage, current, state of charge, temperature, etc. of the battery during real-time operation. In a specific embodiment of the present invention, a recursive least square method with a forgetting factor is used to update the linear parameters in the battery offline model online to obtain the battery online model, that is, to obtain a state-dependent model updated in real time. Linear parameters The specific update formula is:

[0111] (19)

[0112] (20)

[0113] (twenty one)

[0114] in, express The gain matrix at time instant, express The covariance matrix at time , represents the forgetting factor, express The updated value of the linear parameter at time , express The updated value of the linear parameter at time , express The covariance matrix at time , Represents the identity matrix.

[0115] In this embodiment, the covariance matrix The initial value is ,in, The dimension is The specific value of the forgetting factor in this embodiment is 0.995.

[0116] Based on the offline battery model, the linear parameters of the model are updated in combination with the real-time measurement data of the battery, which can more accurately describe the time-varying dynamic characteristics of the lithium-ion battery and reduce the modeling error of the terminal voltage.

[0117] The nonlinear parameters obtained by offline optimization and the linear parameters obtained by online updating are substituted into the state-dependent model of the battery to obtain the battery terminal voltage model.

[0118] Figure 2 The comparison curve between the terminal voltage prediction value and the actual value is shown, wherein the terminal voltage prediction value is predicted by the battery terminal voltage model constructed by the present invention, and the actual terminal voltage value is obtained by actual measurement. Figure 2 It can be seen that the predicted value of the terminal voltage is close to the actual value of the terminal voltage, indicating that the present invention has good modeling accuracy.

[0119] Figure 3 The terminal voltage error (i.e., the difference between the predicted terminal voltage and the actual terminal voltage) curve is shown in Figure 3 It can be seen that the maximum absolute value of the error does not exceed 150 mV, and the absolute value of the error of most sampling points is less than 50 mV, indicating that the model constructed by the present invention can effectively characterize the nonlinear time-varying dynamic characteristics of lithium-ion batteries.

[0120] Example 2

[0121] An embodiment of the present invention also provides an electronic device, which includes: a memory, a processor, and a computer program / instructions stored in the memory, and the processor executes the computer program / instructions to implement the battery offline-online hybrid modeling method in the embodiment of the present application.

[0122] Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes according to the programs and / or data stored in the read-only memory (ROM) or the programs and / or data loaded from the storage part into the random access memory (RAM). The processor can be a multi-core processor or can include multiple processors. In some embodiments, the processor can include a general-purpose main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. In RAM, various programs and data required for device operation are also stored. The processor, ROM, and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.

[0123] The processor and memory are used together to execute the program / instructions stored in the memory. When the program / instructions are executed by the computer, the methods, steps or functions described in the above embodiments can be implemented.

[0124] Although not shown, an embodiment of the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the battery offline-online hybrid modeling method in the embodiment of the present application.

[0125] Readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0126] Although not shown, an embodiment of the present invention further provides a computer program product, including: a computer program / instruction, which, when executed by a processor, implements the battery offline-online hybrid modeling method in the embodiment of the present application.

[0127] What is disclosed above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, which should be covered within the protection scope of the present invention.

Claims

1. A battery offline and online hybrid modeling method, characterized in that: The modeling method comprises: Construct a state-dependent model of the battery; Determine a state vector of the battery according to a state parameter of the battery; Determining functional coefficients of the state-dependent model using the state vector as input; According to the functional coefficients, converting the state-dependent model into a nonlinear least squares model; According to the acquired battery measurement data, an optimization algorithm is used to perform offline optimization on the parameters of the nonlinear least squares model to obtain all parameters, and then obtain a battery offline model; The nonlinear parameters of the battery offline model are kept unchanged, and the linear parameters in the battery offline model are updated online according to the real-time measurement data of the battery, thereby obtaining a state-dependent model updated in real time.

2. The battery offline and online hybrid modeling method according to claim 1, characterized in that: The state dependency model is: ； in, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model, Both represent the state vector The functional coefficients of Indicates that the battery is The state vector at time, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is Current at the moment.

3. The battery offline and online hybrid modeling method according to claim 1, characterized in that: The state vector of the battery is: ； in, Indicates that the battery is The state vector at time, Indicates that the battery is The terminal voltage at the moment, represents the regression order of the terminal voltage, Indicates that the battery is The current at the moment, represents the regression order of the current, Indicates that the battery is The state of charge at the moment, Indicates that the battery is Health status at all times, Indicates that the battery is The temperature of the moment.

4. The battery offline and online hybrid modeling method according to claim 1, characterized in that: The radial basis function neural network is used to determine the functional coefficients of the state-dependent model. The specific expression of the functional coefficients is: ； ； ； in, , , Both represent the state vector The functional coefficient, m represents the number of neurons in the radial basis function neural network, , , , , , Both represent the linear weights of the radial basis function neural network, , Both represent the scale factor of the radial basis function neural network, , Both represent the center vector of the radial basis function neural network, represents the 2-norm.

5. The battery offline and online hybrid modeling method according to claim 4, characterized in that: The nonlinear least squares model is: ； ； ； in, Indicates that the battery is The terminal voltage at the moment, , Both represent intermediate quantities. represents the nonlinear parameters in the functional coefficients, Indicates that the battery is The state vector at time, Indicates the linear parameter in the function coefficient, the superscript represents transpose, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model.

6. The battery offline and online hybrid modeling method according to claim 1, characterized in that: The parameters of the nonlinear least squares model are optimized offline using a structured nonlinear parameter optimization algorithm, specifically including: Step A1: Construct the optimization objective function, the specific expression is: ； ； ； in, represents the optimization target value, Represents the error vector between the predicted value and the true value of the battery terminal voltage, express The maximum value in represents the nonlinear parameters in the functional coefficients, represents the linear parameter in the functional coefficient, represents the 2-norm, , Both represent the order of the state-dependent model, represents the regression order of the terminal voltage, represents the regression order of the current, Indicates that the battery is The terminal voltage at the moment, Indicates that the battery is The terminal voltage at the moment, represents the terminal voltage of the battery at time M, represents the predicted value of the terminal voltage, represents the state vector, M represents the number of sampling points of the data used for parameter optimization; Step A2: Initializing the nonlinear parameters and linear parameters in the functional coefficients; Step A3: According to the initialized nonlinear parameters and the optimization objective function, the Levenberg-Marquardt method is used to calculate the nonlinear parameters. The specific calculation formula is: ； ； ； in, Indicates The nonlinear parameters of the iteration; Indicates The nonlinear parameters of the iteration; represents the step length; represents the Jacobian matrix; Indicates the intermediate quantity, For control The direction and size of Approaching 0, Along the Gauss-Newton direction, when As it approaches infinity, It will be in the direction of steepest descent; represents the identity matrix, Indicates The linear parameters of the iterations, Indicates the symbol for partial derivative; Represents the error vector between the predicted value and the true value of the battery terminal voltage; represents transpose; Step A4: According to the nonlinear parameters and the nonlinear least squares model, the linear parameters are calculated using the least squares method. The specific calculation formula is: ； ； in, Indicates The linear parameter of the iteration, superscript represents the inverse matrix, Indicates that the battery is The terminal voltage at the moment, represents the amount of intermediate calculation, express The state vector at the moment; Step A5: Determine whether the optimization termination condition is met. If so, output the optimized nonlinear parameters and linear parameters to obtain the battery offline model; if not, repeat steps A3 and A4 until the optimization termination condition is met.

7. The battery offline-online hybrid modeling method according to any one of claims 1 to 6, characterized in that: The linear parameters in the battery offline model are updated online using the recursive least square method with a forgetting factor. The specific update formula is: ； ； ； ； ； in, express The gain matrix at time instant, express The covariance matrix at time , , , Both represent intermediate quantities. represents the nonlinear parameters in the functional coefficients, Indicates that the battery is The state vector at time, represents the forgetting factor, express The updated value of the linear parameter at time , express The updated value of the linear parameter at time , Indicates that the battery is The terminal voltage at the moment, express The covariance matrix at time , represents the identity matrix, the superscript represents transpose, Indicates that the battery is The current at the moment, , Both represent the order of the state-dependent model, represents the number of neurons in the radial basis function neural network, represents the scale factor of the radial basis function neural network, represents the center vector of the radial basis function neural network, represents the 2-norm.

8. An electronic device comprising a memory, a processor, and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instructions to implement the battery offline-online hybrid modeling method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the battery offline-online hybrid modeling method as described in any one of claims 1 to 7 is implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the battery offline-online hybrid modeling method as described in any one of claims 1 to 7 is implemented.

Images