A lithium ion battery electrochemical model parameter estimation method based on RA-TLBO

Abstract

This invention discloses a parameter estimation method for an electrochemical model of a lithium-ion battery based on RA-TLBO, including obtaining the reference terminal voltage V of the lithium-ion battery under a 5C high-rate charge-discharge condition. ref Curve; Based on the sensitivity of the pseudo-two-dimensional P2D electrochemical model parameters under 5C high-rate charge-discharge conditions, determine the estimable high-sensitivity parameters and their value ranges; Construct a P2D electrochemical model, solve the P2D electrochemical model using the finite element method, and obtain the simulated terminal voltage V of the lithium-ion battery. sim Curve; based on the reference terminal voltage V of the lithium-ion battery ref Curve, simulated terminal voltage V sim Curve data is used to construct an objective function; the optimal solution for P2D electrochemical model parameters is estimated through the reverse assimilation-teaching and learning optimization RA-TLBO algorithm. This invention features high accuracy, high computational efficiency, and strong robustness, and is suitable for estimating electrochemical model parameters for lithium-ion batteries.

Description

Technical Field

[0001] This invention relates to the field of lithium-ion battery chemical model parameter estimation, and in particular to a method for estimating lithium-ion battery electrochemical model parameters based on RA-TLBO. Background Technology

[0002] Electrochemical models are a key tool for evaluating lithium-ion battery performance and optimizing battery design. They describe the complex electrochemical processes within lithium-ion batteries, involving information such as battery state and behavior. However, electrochemical models have numerous parameters, requiring testing of each parameter using different equipment and methods, and sometimes even disassembling the lithium-ion battery. Therefore, proposing an efficient, accurate, and non-destructive method for estimating electrochemical model parameters is crucial for improving lithium-ion battery design and performance prediction. This invention identifies 13 highly sensitive and estimable electrochemical model parameters, constructs and solves a P2D electrochemical model using finite element method software, and estimates these 13 electrochemical model parameters based on the RA-TLBO (Reverse Assimilation-Teaching LearningBased Optimization) algorithm, enabling accurate and effective estimation of lithium-ion battery electrochemical model parameters.

[0003] The specific patent prior art documents and related literature mentioned above are as follows:

[0004] 1) The article "Offline Parameter Identification of Lithium-ion Batteries Based on RLS Method" by Cao Ming, Zhang Yue, Huang Juhua, et al. from Nanchang University, published in the June 2020 issue of "Batteries" (Volume 50, Issue 3), proposes an improved offline parameter identification method based on the recursive least squares (RLS) algorithm. It uses RLS as the initial value for offline parameter identification to overcome the limitations of initial value selection and ensure the accuracy of the identification results. However, the battery model used in this paper is a second-order RC equivalent circuit model, which cannot intuitively reflect the electrochemical characteristics of the battery. This invention aims to describe the parameters of a more comprehensive P2D electrochemical model, and uses the finite element method for solution. Finally, it proposes the RA-TLBO algorithm to optimize the estimated values ​​of the electrochemical model parameters. Summary of the Invention

[0005] To address the aforementioned technical problems, the purpose of this invention is to provide a method for estimating the parameters of an electrochemical model for lithium-ion batteries based on RA-TLBO, so as to meet the requirements for accurate and rapid estimation of the chemical model parameters of lithium-ion batteries.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A parameter estimation method for an electrochemical model of lithium-ion batteries based on RA-TLBO, comprising:

[0008] Step A: Obtain the reference terminal voltage V of the lithium-ion battery under 5C high-rate charge / discharge conditions. ref curve;

[0009] Step B determines the estimable high-sensitivity parameters and their value ranges based on the sensitivity of the pseudo-two-dimensional P2D electrochemical model parameters under 5C high-rate charge and discharge conditions.

[0010] Step C involves constructing a P2D electrochemical model and solving the P2D electrochemical model using the finite element method to obtain the simulated terminal voltage V of the lithium-ion battery. sim curve;

[0011] Step D is based on the reference terminal voltage V of the lithium-ion battery. ref Curve, simulated terminal voltage V sim Using curve data, construct the objective function;

[0012] Step E uses the reverse assimilation-teaching and learning optimization RA-TLBO algorithm to estimate the optimal solution of the P2D electrochemical model parameters.

[0013] Compared with the prior art, one or more embodiments of the present invention may have the following advantages:

[0014] This invention employs a P2D electrochemical model for modeling, which offers high accuracy and a more intuitive reflection of the electrochemical characteristics of lithium-ion batteries. Simultaneously, the finite element method is used to obtain the terminal voltage curve of the lithium-ion battery. By introducing a reverse assimilation mechanism, the traditional TLBO algorithm is improved, making it more consistent with real-world teaching and learning environments. This increases population diversity, making the algorithm more likely to escape local optima and improving the probability of obtaining the global optimum. Through this method, accurate and effective estimation of the parameters of the lithium-ion battery electrochemical model can be achieved. Attached Figure Description

[0015] Figure 1 This is a flowchart of a parameter estimation method for an electrochemical model of lithium-ion batteries based on RA-TLBO.

[0016] Figure 2 This is a one-dimensional lithium-ion battery model and a schematic diagram of its mesh generation;

[0017] Figure 3 This is a comparison chart of the relative errors of the estimation results of various electrochemical parameters under different algorithms. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in further detail below with reference to the embodiments and accompanying drawings.

[0019] This embodiment uses the reverse assimilation-teaching and learning optimization RA-TLBO algorithm to achieve optimal solution estimation of P2D electrochemical model parameters. For example... Figure 1 The figure shows a parameter estimation method for an electrochemical model of a lithium-ion battery based on RA-TLBO, which includes the following steps:

[0020] Step 10: Obtain the reference terminal voltage V of the lithium-ion battery under 5C high-rate charge / discharge conditions through experiments or simulations. ref curve;

[0021] Step 20: Based on the sensitivity of the pseudo-two-dimensional P2D electrochemical model parameters under 5C high-rate charge and discharge conditions, determine the estimable high-sensitivity parameters and their value ranges.

[0022] Step 30: Construct a P2D electrochemical model using finite element software, and solve the P2D electrochemical model using the finite element method to obtain the simulated terminal voltage V of the lithium-ion battery. sim curve;

[0023] Step 40: Based on the reference terminal voltage V of the lithium-ion battery ref Curve, simulated terminal voltage V sim Using curve data, construct the objective function;

[0024] Step 50 uses the reverse assimilation-teaching and learning optimization RA-TLBO algorithm to estimate the optimal solution of the P2D electrochemical model parameters.

[0025] In step 10 above, the 5C high-rate charge and discharge condition is as follows: 5C rate discharge for 200s, rest for 100s, 5C rate charge for 200s, and finally rest for 100s.

[0026] In step 20 above, the estimable high-sensitivity parameters include the positive electrode length L. pos Negative electrode length L neg ε, solid phase volume fraction of positive electrode s_pos Negative electrode solid volume fraction ε s_neg Positive electrode particle radius R s_pos The radius of the negative electrode particle, R s_neg positive electrode Li + Diffusion coefficient D s_pos negative electrode Li + Diffusion coefficient D s_neg , positive electrode reaction rate constant k pos The negative electrode reaction rate constant k neg Negative electrode SEI film internal resistance R SEI_neg Maximum solid concentration at the positive electrode, c s,max_pos Maximum solid concentration c at the negative electrode s,max_neg There are 13 parameters. Table 1 shows the range of values ​​for the 13 highly sensitive electrochemical model parameters that can be estimated.

[0027] Table 1 shows the range of values ​​for the 13 highly sensitive electrochemical model parameters that can be estimated.

[0028]

[0029]

[0030] In step 30 above, the construction of the P2D model includes the liquid phase diffusion equation, solid phase diffusion equation, liquid phase potential equation, solid phase potential equation, and Butler-Volmer governing equation. The simulated terminal voltage V of the P2D model under 5C high-rate charge-discharge conditions is then solved using the finite element method. sim curve.

[0031] Figure 2 This is a schematic diagram of a one-dimensional lithium-ion battery model and its mesh generation. The three segments in the diagram directly represent the negative electrode, the separator, and the positive electrode, respectively. The ambient temperature is set to 25℃, and an extremely fine mesh generation is used. The diagram shows a magnified view of the mesh generation in some areas. The model contains a total of 300 units and 301 mesh vertices.

[0032] In step 40 above, let x be a vector composed of the electrochemical estimation parameters of the P2D model, and V ref V is the reference terminal voltage. sim Let I be the simulated terminal voltage, I be the load current, and m be the test data length. Then, the formula for calculating the objective function f(x) is:

[0033]

[0034] Step 50 above consists of a teaching phase and a learning phase that includes a reverse assimilation mechanism. The specific steps are as follows:

[0035] During the teaching phase, let teacher X... T Let X be the optimal solution obtained from the population, M be the average score of the entire class, and X be the score of the i-th student. i X i,new and X i,old These are the values ​​of r before and after the i-th student's learning, respectively. i Let r be the learning step size of the i-th student. i = rand(0,1), T F,i Let T be the teaching factor for the i-th student, and T F,i =round(1+rand(0,1)), N P Let X be the population size, then after the i-th student learns... i,new for:

[0036] X i,new =X i,old +r i (X T -T F,i M), where

[0037] During the learning phase, let the i-th student be X. i Randomly select a student X from the class. j (j≠i), adjust and improve the learning method, that is:

[0038]

[0039] In addition, during the learning phase, N is randomly selected. ra Let the number of students be the reverse assimilation set, and let the class reverse assimilation rate be p. ra Then there is Let the i-th student be X. i If the student is not in the reverse assimilation set, then a learning object X is randomly selected. j (j≠i), and update according to formula (3); if the student is in the reverse assimilation set, then:

[0040]

[0041] After the above teaching and learning stages, the students are updated, and the process is iterated continuously until the algorithm converges.

[0042] In this embodiment, Figure 3 This is a comparison chart of the relative errors of the estimation results of various electrochemical parameters under different algorithms. Compared with the TLBO algorithm, the RA-TLBO algorithm improves the estimation accuracy of all 13 electrochemical model parameters, including the positive electrode particle radius R. s_pos Negative electrode particle radius R s_neg positive electrode Li + Diffusion coefficient D s_pos negative electrode Li + Diffusion coefficient D s_neg and the internal resistance R of the negative electrode SEI film SEI_neg The accuracy of parameter estimation is significantly improved; compared with the MultiGA algorithm and BSA algorithm, except for the positive electrode reaction rate constant k... pos and the maximum solid concentration c of the positive electrode s,max_pos Aside from slightly higher values ​​for parameters like the anode reaction rate constant k, the RA-TLBO algorithm has a significant advantage in estimating other electrochemical parameters. Furthermore, among all estimated parameters, the RA-TLBO algorithm has the best estimation performance for k. neg It has the largest relative error value, about 15.2%, but for this parameter estimation, the MultiGA algorithm is as high as about 34.3% and the BSA algorithm is as high as 24.7%, and its estimation performance is significantly better.

[0043] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and variations in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection for this invention shall still be determined by the scope defined in the appended claims.

Claims

1. A parameter estimation method for an electrochemical model of lithium-ion batteries based on RA-TLBO, characterized in that, The method includes the following steps: Step A: Obtain the reference terminal voltage of the lithium-ion battery under 5C high-rate charge / discharge conditions. V ref curve; Step B determines the estimable high-sensitivity parameters and their value ranges based on the sensitivity of the pseudo-two-dimensional P2D electrochemical model parameters under 5C high-rate charge and discharge conditions. Step C involves constructing a P2D electrochemical model and solving the P2D electrochemical model using the finite element method to obtain the simulated terminal voltage of the lithium-ion battery. V sim curve; Step D is based on the reference terminal voltage of the lithium-ion battery. V ref Curves, simulated terminal voltage V sim Using curve data, construct the objective function; Step E uses the reverse assimilation-teaching and learning optimization RA-TLBO algorithm to estimate the optimal solution of the P2D electrochemical model parameters; Step E consists of a teaching phase and a learning phase that includes a reverse mechanism. The specific steps are as follows: During the teaching phase, teachers are appointed. X T This is the optimal solution obtained from the population. M The average score of the entire class, the first i One student X i , X i,new and X i,old The first i The values ​​of each student before and after learning. r i For the first i The learning pace of each student r i =rand(0,1), T F,i For the first i The teaching factors for each student, and T F,i =round(1+rand(0,1)), N P Let be the population number, then the th i After each student learned X i,new for: （2）； During the learning phase, let the first... i students X i Randomly select a student from the class. X j ,in j ≠ i Adjust and improve learning methods, namely: （3）； In addition, during the learning phase, random selection N ra Let the number of students be the reverse assimilation set, and let the class reverse assimilation rate be... p ra Then there is Let the first i One student X i If the student is not in the reverse assimilation set, then a learning object is randomly selected. X j ,in j ≠ i And update according to formula (3); if the student is in the reverse assimilation set, then: （4）； After the above teaching and learning stages, the students are updated, and the process is iterated continuously until the algorithm converges.

2. The method for estimating parameters of an electrochemical model for lithium-ion batteries based on RA-TLBO according to claim 1, characterized in that, In step A, the 5C high-rate charge and discharge condition is as follows: 5C rate discharge for 200s, rest for 100s, 5C rate charge for 200s, and finally rest for 100s.

3. The method for estimating parameters of an electrochemical model of a lithium-ion battery based on RA-TLBO according to claim 1, characterized in that, In step B, the estimable high-sensitivity parameters include: positive electrode length. L pos Negative electrode length L neg Volume fraction of positive electrode solid phase ε s_pos Negative electrode solid phase volume fraction ε s_neg Positive electrode particle radius R s_pos Negative electrode particle radius R s_neg positive electrode Li + diffusion coefficient D s_pos negative electrode Li + diffusion coefficient D s_neg , positive electrode reaction rate constant k pos negative electrode reaction rate constant k neg Negative electrode SEI film internal resistance R SEI_neg Maximum solid concentration at the positive electrode c s,max_pos and the maximum solid concentration of the negative electrode c s,max_neg 13 parameters.

4. The method for estimating parameters of an electrochemical model of a lithium-ion battery based on RA-TLBO according to claim 1, characterized in that, In step C, a P2D electrochemical model is constructed using finite element software. The construction of the P2D electrochemical model includes liquid-phase diffusion equations, solid-phase diffusion equations, liquid-phase potential equations, solid-phase potential equations, and the Butler-Volmer governing equations. The simulated terminal voltage of the P2D model under 5C high-rate charge-discharge conditions is then solved using the finite element method. V sim curve.

5. The method for estimating parameters of an electrochemical model of a lithium-ion battery based on RA-TLBO according to claim 1, characterized in that, In step D, let x be a vector composed of the electrochemical estimation parameters of the P2D model. V ref Reference terminal voltage, V sim To simulate terminal voltage, I For load current, m Given the length of the test data, the objective function is... f The formula for calculating (x) is: （1）。

Images