Energy storage battery soh estimation method based on electrochemical model and machine learning

Abstract

The application discloses a kind of based on electrochemical model and machine learning's energy storage battery SOH estimation method, comprising: S1: the electrochemical model of energy storage battery is constructed;S2: parameter identification is carried out to electrochemical model under initial aging cycle;S3: based on the model features related to battery aging of electrochemical model after parameter identification extraction;S4: based on experimental data extraction and battery aging related data features;S5: correlation analysis, feature dimension reduction and feature fusion are carried out to model features and data features, and obtain fusion features;S6: fusion features are input into trained support vector machine model, and the corresponding battery health state prediction value is output.The application extracts the features related to battery aging from electrochemical model and experimental data, obtains fusion features by principal component analysis method for dimension reduction processing of features, establishes battery aging model using support vector machine to realize SOH estimation, so as to effectively improve the accuracy and generalization of energy storage battery SOH estimation.

Description

Technical Field

[0001] This invention relates to the fields of Internet big data and battery management, specifically to a method for estimating the state of energy storage batteries (SOH) based on electrochemical models and machine learning. Background Technology

[0002] With increasing global pollution and the depletion of fossil fuels, energy conservation and environmental protection have become common goals for every country's development. A key aspect of achieving "dual-carbon" targets is making clean energy the primary source of power for the electricity system. New energy vehicles play a crucial role in the adjustment of national energy structures. Lithium-ion batteries, as a clean energy storage technology, have also become one of the most promising candidate batteries for electric vehicles.

[0003] While lithium-ion batteries offer advantages such as energy efficiency, environmental friendliness, lack of memory effect, long cycle life, and high energy density, their capacity and power performance gradually deteriorate over time. This severely impacts the driving range and lifespan of electric vehicles and can even lead to electrolyte leakage and micro-short circuits, causing battery failure and thermal runaway, potentially resulting in catastrophic accidents. Therefore, lithium battery safety has become a major bottleneck restricting the development of new energy vehicles. Effective battery management is essential to ensure the safe operation of electric vehicles. A Battery Management System (BMS) can monitor battery status, provide safety warnings, and ensure long-term safe and reliable battery operation. Battery State of Health (SOH) is one of the most important performance indicators of a BMS.

[0004] Currently, SOH estimation methods mainly fall into two categories: model-based methods and data-driven methods. Model-based methods primarily rely on establishing equivalent circuit models and electrochemical models to estimate SOH. Equivalent circuit models do not consider the internal chemical composition and corresponding reactions of the battery. Since internal resistance and capacity can effectively characterize battery aging, their changes indirectly reflect changes in battery health. However, this method is susceptible to the influence of operating conditions and the convergence of the estimation algorithm. Electrochemical models can effectively simulate the electrochemical reaction process of the battery, but they suffer from difficulties in model parameter identification, computational complexity due to numerous partial differential equations, and poor generalization ability. Data-driven model-based methods do not need to consider the internal electrochemical reaction behavior and failure mechanism characteristics of the battery; they only need to analyze large amounts of data to accurately estimate the battery's health. However, this method lacks clear physical meaning, depends on data quality, has a high computational load, and current research mainly focuses on constant current charge-discharge conditions, while user discharge behavior is random, leading to poor accuracy in lithium battery SOH estimation. Therefore, designing a method that can effectively improve the accuracy and generalization of SOH estimation for energy storage batteries is an urgent technical problem to be solved. Summary of the Invention

[0005] To address the shortcomings of the existing technologies, the technical problem to be solved by this invention is: how to provide a SOH estimation method for energy storage batteries based on electrochemical models and machine learning. First, features related to battery aging are extracted from electrochemical models and experimental data. Second, the features are dimensionality-reduced using principal component analysis to obtain fused features. Finally, a battery aging model is established using support vector machines to achieve SOH estimation, thereby effectively improving the accuracy and generalization of SOH estimation for energy storage batteries (lithium batteries).

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] A method for estimating the state of energy (SOH) of an energy storage battery based on electrochemical models and machine learning includes:

[0008] S1: Construct an electrochemical model for energy storage batteries;

[0009] S2: Parameter identification of the electrochemical model under initial aging cycles;

[0010] S3: Extracting model features related to battery aging based on the electrochemical model after parameter identification;

[0011] S4: Extract data features related to battery aging based on the acquired experimental data;

[0012] S5: Perform correlation analysis, feature dimensionality reduction, and feature fusion on model features and data features to obtain fused features;

[0013] S6: Input the fused features into the trained support vector machine model and output the corresponding battery health status prediction value.

[0014] Preferably, in step S1, the electrochemical model of the energy storage battery is constructed through the following steps:

[0015] S101: The relationship between the lithium-ion flux generated by the electrochemical reaction on the surface of the active particles of the negative and positive electrodes of the energy storage battery and the external current is calculated as follows:

[0016]

[0017] In the formula: j r,n This indicates the relationship between the lithium-ion flux generated by the electrochemical reaction on the surface of the negative electrode active particles of an energy storage battery and the external current; j r,p This represents the relationship between the lithium-ion flux generated by the electrochemical reaction on the surface of the positive electrode active particles of an energy storage battery and the external current; i represents the external current; L n Indicates the thickness of the negative electrode; A n This represents the area of ​​the negative electrode plate; an L represents the specific surface area of ​​the negative electrode active particles; F represents the Faraday constant; L represents the specific surface area of ​​the negative electrode active particles. p Indicates the thickness of the positive electrode; A p This represents the area of ​​the positive electrode plate; a p This represents the specific surface area of ​​the positive electrode active particles;

[0018] S102: The lithium-ion concentration on the particle surface of the positive and negative electrodes was obtained using the second-order Runge-Kutta method.

[0019]

[0020] In the formula: c surf,p and c surf,n D represents the lithium-ion concentration on the particle surface of the positive and negative electrodes, respectively; s,p and D s,n Indicates the solid-phase diffusion coefficient of the positive and negative electrodes;

[0021] S103: Calculate the electrode utilization rate of the positive and negative electrodes based on the lithium ion concentration on the particle surface, and then calculate the open circuit voltage;

[0022] The formula for calculating electrode utilization rate is as follows:

[0023]

[0024] In the formula: θ p θ n These represent the electrode utilization rates of the positive and negative electrodes, respectively; c smax,p and c smax,n These represent the maximum solid-phase lithium-ion concentrations of the positive and negative electrodes, respectively.

[0025] The formula for the positive open-circuit voltage is:

[0026] U p (θ p )=4.65-0.2076*tanh((θ p -0.4) / 0.06004)-0.06572*tanh((θ p -0.552) / 0.04231)

[0027] -0.1478*tanh((θ p -0.728) / 0.09524)+0.012814*tanh((θ p -0.4445) / 0.01732)

[0028] +0.006405*tanh((θ p -0.56) / 0.02347)-0.0728*tanh((θ p-0.90) / 0.06714)

[0029] -0.341*exp(223.8*(θ p -0.999))+0.004366*tanh((θ p -0.803) / 0.02835)

[0030] -0.005707*tanh((θ p -0.972) / 0.01034)+0.01604*exp(-((θ p -0.9927) / 0.003936) 2 )

[0031] +0.001*(-2.529*tanh((θ p -0.7) / 0.08168))-0.55

[0032] The formula for the open-circuit voltage of the negative terminal is as follows:

[0033] U n (θ n )=4.973-2.643*tanh((θ n -0.9832) / 0.01329)+5.621*tanh((θ n -0.6752) / 1.164)

[0034] -0.1455*tanh((θ n -0.5721) / 1.55)+4.688*tanh((θ n -0.472) / 0.2705)-0.5182

[0035] *tanh((θ n -0.4024) / 1.16)+2.935*exp(-28.9*θ n )+7.418*tanh((θ n -0.2756) / 0.204)

[0036] +14.46*tanh((θ n -0.1097) / 0.1035)-36.24*tanh((θ n -0.0624) / 0.4477)

[0037] +11.54*exp(-((θ n -0.03786) / 0.1157) 2 )

[0038] S104: The formula for defining liquid phase potential is:

[0039]

[0040] Where: φ e Represents the liquid phase potential; A represents the electrode surface area; k eff L represents the effective conductivity. p L n and L sep These represent the thicknesses of the positive electrode, negative electrode, and separator, respectively.

[0041] S105: The j in step S101... r,n and j r,p Substituting into the Butler-Volmer equation, we get:

[0042]

[0043] In the formula: k s,n and k s,p The rate constants representing the electrochemical reactions at the negative and positive electrodes; c e c represents the concentration of lithium ions in the liquid phase. e-s,n and c e-s,p Represents the lithium-ion concentration in the solid-liquid phase; α represents the transfer coefficient; R represents the gas constant; T represents the temperature constant; η p and η n These represent the overpotentials at the positive and negative terminals, respectively.

[0044] Wherein, auxiliary variable ξ p and ξ n for:

[0045]

[0046] The magnitude of the overpotential generated by the positive and negative electrode reactions is calculated as follows:

[0047]

[0048] S106: Establish an electrochemical model to estimate the battery's terminal voltage and state of charge (SOC);

[0049] The formula for calculating the terminal voltage in the electrochemical model is expressed as follows:

[0050] U t =φ e +(U p -U n )+η p -η n +IR f ;

[0051] In the formula: U tRepresents terminal voltage; R f U represents the internal resistance of the ohm; p U represents the open-circuit voltage of the positive terminal; n Indicates the open-circuit voltage of the negative terminal; η p Indicates the overpotential generated by the positive electrode reaction; η n φ represents the overpotential generated by the negative electrode reaction. e I represents the liquid phase potential; I represents the current.

[0052] The state of charge (SOC) in the electrochemical model is expressed using the electrode utilization rate of the positive electrode as follows:

[0053]

[0054] In the formula: SOC(t) represents the SOC value; θ p0％ θ p100％ θ represents the utilization rate of the spherical particle surface electrode when the battery is fully discharged and fully charged, respectively; p This indicates the electrode utilization rate of the positive electrode.

[0055] Preferably, in step S2, the electrochemical model is parameter identified using a particle swarm optimization algorithm under the initial aging cycle.

[0056] The processing steps of the particle swarm optimization algorithm include:

[0057] S201: Determine the model parameters that need to be identified;

[0058] S202: Define each particle as a set of model parameters to be identified;

[0059] S203: Initialize particle position and velocity:

[0060] X = [c smax,n ,c smax,p ,R n ,R p ,k s,n ,k s,p ] T ;

[0061] V = [V csmax,n V csmax,p V Rn V Rp V ks,n V ks,p ] T ;

[0062] In the formula: c smax,p and c smax,n R represents the maximum solid-phase lithium-ion concentration at the negative and positive electrodes, respectively. n and R p Let k represent the radii of the solid particles at the negative and positive electrodes, respectively.s,n , and k s,p represents the electrochemical reaction rate constants at the negative and positive electrodes, respectively, and X and V represent the particle position and velocity, respectively;

[0063] S204: Define the minimum root mean square error of velocity as the fitness function;

[0064] The formula is expressed as:

[0065]

[0066] In the formula: Fit represents the fitness function, V(t) and These are represented as the simulated terminal voltage and the measured terminal voltage, respectively.

[0067] S205: Calculate the fitness value of each particle; compare the fitness of the current particle position with its historical best position; if the current position is better, update the individual's best position.

[0068] S206: Find the particle with the best fitness among all particles and update the global optimal position;

[0069] S207: Update the velocity and position of each particle based on inertial weights, acceleration constants, and individual and global best positions;

[0070] The formula is expressed as:

[0071]

[0072] In the formula: X i (k) represents the position vector of the particle in the kth iteration; V i (k) represents the velocity vector of the particle in the kth iteration; P i (k) represents the historical best position of the particle in the kth iteration; G i (k) represents the historical best position of the group in the k-th iteration; w represents the inertia weight; c1 and c2 represent the individual learning factor and the group learning factor, respectively; r1 and r2 are random numbers in the interval [0,1].

[0073] S208: Repeat steps S205 to S207 until the preset number of iterations is reached or the convergence condition is met. Output the model parameters corresponding to the optimal particle as the parameter identification result of the electrochemical model.

[0074] Preferably, in step S201, by using the controlled variable method, while ensuring that other parameters of the electrochemical model remain unchanged, the value of a certain parameter is changed to analyze the battery voltage characteristics of the electrochemical model, and several parameters with the highest parameter sensitivity are selected as model parameters to be identified.

[0075] Preferably, in step S3, model features related to battery aging are extracted through the following steps:

[0076] S301: Input experimental data under different operating conditions during different aging cycles into the electrochemical model after parameter identification;

[0077] S302: Calculate the simulated SOC value of the energy storage battery using an electrochemical model; calculate the SOC error (SOC error) by comparing the simulated and actual SOC values. error Extracting SOC error (SOC) error The mean and integral are used as model features F1 and model features F2;

[0078] The formula is expressed as:

[0079] F1 = mean(SOC) error );

[0080]

[0081] S303: Calculate the simulated voltage value of the energy storage battery using an electrochemical model; calculate the voltage error U by comparing the simulated voltage value with the actual voltage value. error Extraction terminal voltage error U error The mean and integral are used as model features F3 and model features F4;

[0082] The formula is expressed as:

[0083] F3=mean(U error );

[0084]

[0085] Preferably, in step S4, data features related to battery aging are extracted through the following steps:

[0086] S401: Obtain experimental data;

[0087] S402: Extract discharge time as data feature F5 from experimental data;

[0088] The formula is expressed as:

[0089] F5 = t end -t0;

[0090] In the formula: t end t0 represents the discharge end time; t0 represents the discharge start time.

[0091] S403: Extract net discharge energy from experimental data as data feature F6;

[0092] The formula is expressed as:

[0093] F6=∫U d I d dt d -∫U c I c dt c ;

[0094] In the formula: U d and U c These represent the voltages corresponding to the discharge current and the charging current, respectively; I d and I c These represent the discharge current and the charging current, respectively; t d and t c These represent the discharge time and charging time, respectively.

[0095] S404: Extract the voltage change rate as a data feature F7 from experimental data;

[0096] The formula is expressed as:

[0097]

[0098] In the formula: U end U0 represents the voltage at the end of the discharge; U0 represents the voltage at the beginning of the discharge; Δt represents the time from the beginning to the end of the discharge.

[0099] Preferably, in step S5, the fusion feature is generated through the following steps:

[0100] S501: Put all model features and data features into one dataset;

[0101] S502: Normalize all features in the dataset;

[0102] The formula is expressed as:

[0103]

[0104] In the formula: x normalized x represents the normalized feature; x represents the original feature; x max and x min These represent the maximum and minimum values ​​of the features in the dataset, respectively.

[0105] S503: Perform correlation analysis on the normalized features of the dataset using the Pearson correlation coefficient;

[0106] The formula for correlation analysis is:

[0107]

[0108] In the formula: ρ X,YThe Pearson correlation coefficient is represented by cov(X,Y); cov(X,Y) represents the covariance matrix of two continuous variables X and Y, and σ X and σ Y Let X and Y represent the standard deviations respectively; E represents the expected value; μ X and μ Y These are the means of X and Y, respectively;

[0109] S504: Perform principal component analysis to reduce the dimensionality of several target features with the highest correlation coefficients, and select the feature with the largest contribution rate as the fusion feature after dimensionality reduction;

[0110] The steps involved in principal component analysis dimensionality reduction include:

[0111] S5041: Centralize the target features;

[0112] The formula is expressed as:

[0113]

[0114] In the formula: Represents the input features X1, X2, X3, ..., X n The mean of y1, y2, ... y2; n This represents the target characteristics after centralized processing.

[0115] S5042: Calculate the covariance matrix based on the target features after centralization;

[0116] The formula is expressed as:

[0117]

[0118] In the formula: cov(y1,y2) represents the calculation of covariance;

[0119] S5043: Solving for the eigenvalues ​​and eigenvectors of the covariance matrix through eigenvalue decomposition;

[0120] S5044: Arrange the eigenvectors into a matrix from top to bottom according to the magnitude of their corresponding eigenvalues, and take the first k columns to form the principal components V. k ;

[0121] S5045: Multiply the original dataset by the principal component V k The features in the dataset are projected into a low-dimensional space to obtain the dimensionality-reduced data, which is the contribution rate of the features in the principal component space.

[0122] The formula is expressed as:

[0123] W k =X·V k ;

[0124] In the formula: W k X represents the data after dimensionality reduction; X represents the original dataset.

[0125] S5046: Select the feature with the highest contribution rate as the fusion feature after dimensionality reduction.

[0126] Preferably, in step S6, the support vector machine model is a support vector regression model.

[0127] Preferably, in step S6, the support vector regression model is trained using the k-fold cross-validation algorithm. The training steps include:

[0128] S601: Obtain a dataset containing input features and target variables; divide the dataset into k non-overlapping subsets;

[0129] S602: For each of the k subsets, perform the following steps:

[0130] S6021: Use the current subset as the test set and the other k-1 subsets as the training set;

[0131] S6022: Train a support vector regression model on the training set;

[0132] S6023: Evaluate the performance of the support vector regression model on the test set and record the performance metrics;

[0133] S6024: Repeat S6021 to S6023, average the performance metrics obtained from k iterations, and obtain an overall evaluation of the performance of the support vector regression model.

[0134] S603: Based on the results of k-fold cross-validation, compare the average performance indicators under different parameter combinations; select the parameter combination with the best performance indicators as the optimal parameters for the support vector regression model;

[0135] S604: The optimal parameters of the support vector regression model were determined by training the entire dataset until the model converged.

[0136] Preferably, in step S6023, the performance of the support vector regression model is evaluated by the root mean square error and the mean absolute error.

[0137] The formula is expressed as:

[0138]

[0139] In the formula: N represents the number of samples; This represents the predicted battery health status output by the support vector regression model; y m This indicates the true value of the battery's health status.

[0140] The SOH estimation method for energy storage batteries based on electrochemical models and machine learning in this invention has the following advantages compared with existing technologies:

[0141] First, this invention constructs an electrochemical model of the energy storage battery and identifies parameters under initial aging cycles to accurately describe the battery's internal electrochemical behavior. Dynamic operating condition data under different aging cycles are input into the electrochemical model to obtain voltage error and SOC error under different aging cycles, enabling the acquisition of model features closely related to battery aging based on model errors. Simultaneously, data features are extracted from experimental data, and the extracted model features and data features are subjected to correlation analysis, feature dimensionality reduction, and feature fusion to further screen and optimize the feature set used for SOH estimation, thereby improving the accuracy of SOH estimation for energy storage batteries (lithium batteries), making the prediction results closer to the actual health state of the battery, and better meeting the needs of battery management systems for accurate SOH estimation. Then, this invention reduces the number of features through feature dimensionality reduction, lowering computational complexity and helping to remove redundant information, improving the model's generalization ability. Furthermore, this invention utilizes a support vector machine (support vector regression, SVR) model for SOH estimation, leveraging SVR's advantages in handling nonlinear problems and avoiding overfitting to enhance the model's generalization performance, enabling the model to maintain stable predictive performance under different operating conditions and battery types, thereby improving the practicality and reliability of energy storage battery SOH estimation. Secondly, by applying feature fusion and the SVR model, this invention achieves fast and real-time SOH estimation while maintaining accuracy. The SVR model has high computational efficiency when processing large datasets, while feature fusion reduces the number of features and lowers computational complexity. Finally, addressing the issues of long parameter identification time and complex calculation process in electrochemical models, this invention only requires parameter identification for the electrochemical model under the initial cycle, without updating the electrochemical parameters for each cycle. By extracting model features using an electrochemical model with practical physical meaning, the accuracy and computational efficiency of SOH estimation for energy storage batteries can be improved. Attached Figure Description

[0142] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0143] Figure 1 and Figure 2 The diagrams are the logic block diagram and schematic diagram of the SOH estimation method for energy storage batteries.

[0144] Figure 3 This is a schematic diagram of the electrochemical model.

[0145] Figure 4 This is a flowchart of parameter identification using the particle swarm optimization algorithm.

[0146] Figure 5 and Figure 6 These are schematic diagrams showing voltage error and SOC error, respectively.

[0147] Figure 7 A flowchart for generating fusion features.

[0148] Figure 8 A schematic diagram illustrating the SOH estimation for a support vector regression model. Detailed Implementation

[0149] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but only to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0150] The following detailed explanation illustrates the specific implementation methods:

[0151] Example:

[0152] This embodiment discloses a method for estimating the SOH of an energy storage battery based on an electrochemical model and machine learning.

[0153] like Figure 1 and Figure 2 As shown, the SOH estimation method for energy storage batteries based on electrochemical models and machine learning includes:

[0154] S1: Construct an electrochemical model for energy storage batteries (lithium batteries);

[0155] S2: Parameter identification of the electrochemical model under initial aging cycles;

[0156] S3: Extracting model features related to battery aging based on the electrochemical model after parameter identification;

[0157] S4: Extract data features related to battery aging based on the acquired experimental data;

[0158] S5: Perform correlation analysis, feature dimensionality reduction, and feature fusion on model features and data features to obtain fused features;

[0159] S6: Input the fused features into the trained support vector machine model and output the corresponding battery health status prediction value.

[0160] First, this invention constructs an electrochemical model of the energy storage battery and identifies parameters under initial aging cycles to accurately describe the battery's internal electrochemical behavior. Dynamic operating condition data under different aging cycles are input into the electrochemical model to obtain voltage error and SOC error under different aging cycles, enabling the acquisition of model features closely related to battery aging based on model errors. Simultaneously, data features are extracted from experimental data, and the extracted model features and data features are subjected to correlation analysis, feature dimensionality reduction, and feature fusion to further screen and optimize the feature set used for SOH estimation, thereby improving the accuracy of SOH estimation for energy storage batteries (lithium batteries), making the prediction results closer to the actual health state of the battery, and better meeting the needs of battery management systems for accurate SOH estimation. Then, this invention reduces the number of features through feature dimensionality reduction, lowering computational complexity and helping to remove redundant information, improving the model's generalization ability. Furthermore, this invention utilizes a support vector machine (support vector regression, SVR) model for SOH estimation, leveraging SVR's advantages in handling nonlinear problems and avoiding overfitting to enhance the model's generalization performance, enabling the model to maintain stable predictive performance under different operating conditions and battery types, thereby improving the practicality and reliability of energy storage battery SOH estimation. Secondly, by applying feature fusion and the SVR model, this invention achieves fast and real-time SOH estimation while maintaining accuracy. The SVR model has high computational efficiency when processing large datasets, while feature fusion reduces the number of features and lowers computational complexity. Finally, addressing the issues of long parameter identification time and complex calculation process in electrochemical models, this invention only requires parameter identification for the electrochemical model under the initial cycle, without updating the electrochemical parameters for each cycle. By extracting model features using an electrochemical model with practical physical meaning, the accuracy and computational efficiency of SOH estimation for energy storage batteries can be improved.

[0161] To better illustrate the technical solution of the present invention, this embodiment is described in the following parts.

[0162] I. Electrochemical Model

[0163] In the specific implementation process, combined with Figure 3 As shown, the electrochemical model of the energy storage battery is constructed through the following steps:

[0164] S101: Ignoring the diffusion of solid-phase lithium ion concentration along the electrode, it is assumed that the solid spherical particles distributed along the x-direction in each electrode are indistinguishable, and the solid-phase diffusion of lithium ions occurs in a representative spherical particle. For a certain position with coordinate x' in the negative electrode region, we have:

[0165]

[0166] In the formula: i e (x') represents the liquid current density at x'; a n This represents the specific surface area of ​​the negative electrode active particles; j r,n The flux of lithium ions generated by the electrochemical reaction on the surface of the negative electrode active particles; F represents the Faraday constant;

[0167] From the boundary conditions of the negative electrode liquid phase current, we can know that:

[0168]

[0169] Therefore, the relationship between the lithium-ion flux generated by the electrochemical reaction on the surface of the negative electrode active particles of the energy storage battery and the external current is calculated as follows:

[0170]

[0171] In the formula: i represents the external current; L n Indicates the thickness of the negative electrode; A n This represents the area of ​​the negative electrode plate; a n The specific surface area of ​​the negative electrode active particles is represented by F; F represents the Faraday constant.

[0172] Similarly, for a position x' in the positive polar region, we have:

[0173]

[0174] In the formula: a p This represents the specific surface area of ​​the positive electrode active particles; j r,p The lithium-ion flux generated by the electrochemical reaction on the surface of the positive electrode active particles;

[0175] From the boundary conditions of the positive electrode liquid phase current, we know that:

[0176]

[0177] Therefore, the relationship between the lithium-ion flux generated by the electrochemical reaction on the surface of the positive electrode active particles of the energy storage battery and the external current is calculated as follows:

[0178]

[0179] In the formula: i represents the external current; L p Indicates the thickness of the positive electrode; A p This represents the area of ​​the positive electrode plate; a p The positive electrode active particle has a specific surface area; F represents the Faraday constant.

[0180] S102: The pseudo-two-dimensional model is the most classic electrochemical model. It contains four partial differential equations, one algebraic equation, and one equation for calculating the terminal voltage. The equations are heavily coupled, and the partial differential equations are difficult to calculate. Therefore, the pseudo-two-dimensional model needs to be simplified. The solid-phase diffusion equation is represented by a complex partial differential equation. Therefore, the finite difference method is used to divide the radius of the active particles into m segments, converting the partial differential equation into an ordinary differential equation for calculation, thus obtaining the state-space expression of the system.

[0181] The formula is expressed as:

[0182]

[0183] In the formula: R represents the radius of the solid particle; m is 10; D s The solid-phase diffusion coefficient; The state vector representing the lithium-ion concentration in solid particles;

[0184] Taking the negative electrode solid particles as an example, the system's state equation is expressed as follows:

[0185]

[0186] In short:

[0187]

[0188] Among them, C n Let A be the state vector representing the lithium-ion concentration of the active particles. n and B n These are the coefficient matrices corresponding to the state-space expressions.

[0189] The higher-order system was solved using the second-order Runge-Kutta method to obtain the lithium-ion concentrations on the particle surfaces of the positive and negative electrodes:

[0190]

[0191] In the formula: c surf,p and c surf,n D represents the lithium-ion concentration on the particle surface of the positive and negative electrodes, respectively; s,p and D s,n Indicates the solid-phase diffusion coefficient of the positive and negative electrodes;

[0192] S103: Calculate the electrode utilization rate of the positive and negative electrodes based on the lithium ion concentration on the particle surface, and then calculate the open circuit voltage;

[0193] The formula for calculating electrode utilization rate is as follows:

[0194]

[0195] In the formula: θp θ n These represent the electrode utilization rates of the positive and negative electrodes, respectively; c smax,p and c smax,n These represent the maximum solid-phase lithium-ion concentrations of the positive and negative electrodes, respectively.

[0196] The formula for the positive open-circuit voltage is:

[0197] U p (θ p )=4.65-0.2076*tanh((θ p -0.4) / 0.06004)-0.06572*tanh((θ p -0.552) / 0.04231)

[0198] -0.1478*tanh((θ p -0.728) / 0.09524)+0.012814*tanh((θ p -0.4445) / 0.01732)

[0199] +0.006405*tanh((θ p -0.56) / 0.02347)-0.0728*tanh((θ p -0.90) / 0.06714)

[0200] -0.341*exp(223.8*(θ p -0.999))+0.004366*tanh((θ p -0.803) / 0.02835)

[0201] -0.005707*tanh((θ p -0.972) / 0.01034)+0.01604*exp(-((θ p -0.9927) / 0.003936) 2 )

[0202] +0.001*(-2.529*tanh((θ p -0.7) / 0.08168))-0.55

[0203] The formula for the open-circuit voltage of the negative terminal is as follows:

[0204] U n (θ n )=4.973-2.643*tanh((θ n -0.9832) / 0.01329)+5.621*tanh((θn -0.6752) / 1.164)

[0205] -0.1455*tanh((θ n -0.5721) / 1.55)+4.688*tanh((θ n -0.472) / 0.2705)-0.5182

[0206] *tanh((θ n -0.4024) / 1.16)+2.935*exp(-28.9*θ n )+7.418*tanh((θ n -0.2756) / 0.204)

[0207] +14.46*tanh((θ n -0.1097) / 0.1035)-36.24*tanh((θ n -0.0624) / 0.4477)

[0208] +11.54*exp(-((θ n -0.03786) / 0.1157) 2 )

[0209] S104: The formula for defining liquid phase potential is:

[0210]

[0211] Where: φ e Represents the liquid phase potential; A represents the electrode surface area; k eff L represents the effective conductivity. p L n and L sep These represent the thicknesses of the positive electrode, negative electrode, and separator, respectively.

[0212] S105: The j in step S101... r,n and j r,p Substituting into the Butler-Volmer equation, we get:

[0213]

[0214] In the formula: k s,n and k s,p The rate constants representing the electrochemical reactions at the negative and positive electrodes; c e c represents the concentration of lithium ions in the liquid phase. e-s,n and c e-s,p Represents the lithium-ion concentration in the solid-liquid phase; α represents the transfer coefficient; R represents the gas constant; T represents the temperature constant; ηp and η n These represent the overpotentials at the positive and negative terminals, respectively.

[0215] Wherein, auxiliary variable ξ p and ξ n for:

[0216]

[0217] The magnitude of the overpotential generated by the positive and negative electrode reactions is calculated as follows:

[0218]

[0219] S106: Finally, an electrochemical model is established to estimate the battery's terminal voltage and state of charge (SOC).

[0220] The formula for calculating the terminal voltage in the electrochemical model is expressed as follows:

[0221] U t =φ e +(U p -U n )+η p -η n +IR f ;

[0222] In the formula: U t Represents terminal voltage; R f U represents the internal resistance of the ohm; p U represents the open-circuit voltage of the positive terminal; n Indicates the open-circuit voltage of the negative terminal; η p Indicates the overpotential generated by the positive electrode reaction; η n φ represents the overpotential generated by the negative electrode reaction. e I represents the liquid phase potential; I represents the current.

[0223] The state of charge (SOC) in the electrochemical model is expressed using the electrode utilization rate of the positive electrode as follows:

[0224]

[0225] In the formula: SOC(t) represents the SOC value; θ p0％ θ p100％ θ represents the utilization rate of the spherical particle surface electrode when the battery is fully discharged and fully charged, respectively; p This indicates the electrode utilization rate of the positive electrode.

[0226] This invention extracts model features based on an electrochemical model and simplifies the model using the finite difference method. Compared with equivalent circuit models and data-driven methods, the electrochemical model can describe the internal electrochemical reaction process of the battery, more accurately describe the battery's aging state, and improve the accuracy of SOH estimation. However, the traditional pseudo-two-dimensional model contains four partial differential equations, one algebraic equation, and one terminal voltage calculation equation. These equations are heavily coupled, the partial differential equations are difficult to calculate, and there are more than 30 electrochemical parameters that are difficult to obtain and are also affected by temperature and aging. Therefore, this invention simplifies the electrochemical model.

[0227] II. Parameter Identification

[0228] In the specific implementation process, the particle swarm optimization algorithm is used to identify the parameters of the electrochemical model under the initial aging cycle.

[0229] Combination Figure 4 As shown, the processing steps of the particle swarm optimization algorithm include:

[0230] S201: Determine the model parameters that need to be identified;

[0231] In this embodiment, by using the controlled variable method, while ensuring that other parameters of the electrochemical model remain unchanged, the value of a certain parameter is changed to analyze the battery voltage characteristics of the electrochemical model, and several parameters with the highest parameter sensitivity are selected as model parameters to be identified.

[0232] S202: Define each particle as a set of model parameters to be identified;

[0233] S203: Initialize particle position and velocity:

[0234] X = [c smax,n ,c smax,p ,R n ,R p ,rk n ,rk p ] T ;

[0235] V = [V csmax,n V csmax,p V Rn V Rp V rkn V rkp ] T ;

[0236] In the formula: c smax,p and c smax,n R represents the maximum solid-phase lithium-ion concentration at the negative and positive electrodes, respectively. n and R pLet k represent the radii of the solid particles at the negative and positive electrodes, respectively. s,n , and k s,p represents the electrochemical reaction rate constants at the negative and positive electrodes, respectively, and X and V represent the particle position and velocity, respectively;

[0237] S204: Define the minimum root mean square error of velocity as the fitness function;

[0238] The formula is expressed as:

[0239]

[0240] In the formula: Fit represents the fitness function, V(t) and These are represented as the simulated terminal voltage and the measured terminal voltage, respectively.

[0241] In this embodiment, necessary parameter settings can also be made for the particle swarm optimization algorithm as needed, including population size, number of iterations, inertia weight, acceleration constant, and boundary conditions. These parameter settings are accomplished using existing methods.

[0242] S205: Calculate the fitness value of each particle; compare the fitness of the current particle position with its historical best position; if the current position is better, update the individual's best position.

[0243] S206: Find the particle with the best fitness among all particles and update the global optimal position;

[0244] S207: Update the velocity and position of each particle based on inertial weights, acceleration constants, and individual and global best positions;

[0245] The formula is expressed as:

[0246]

[0247] In the formula: X i (k) represents the position vector of the particle in the kth iteration; V i (k) represents the velocity vector of the particle in the kth iteration; P i (k) represents the historical best position of the particle in the kth iteration; G i (k) represents the historical best position of the group in the kth iteration; w represents the inertia weight; c1 and c2 represent the individual learning factor and the group learning factor, respectively; r1 and r2 are random numbers in the interval [0,1] to increase the randomness of the search;

[0248] S208: Repeat steps S205 to S207 until the preset number of iterations is reached or the convergence condition is met. Output the model parameters corresponding to the optimal particle as the parameter identification result of the electrochemical model.

[0249] In this embodiment, the final parameter identification results are shown in Table 1.

[0250] Table 1. Identification results of highly sensitive parameters at the initial aging point.

[0251]

[0252] This invention only identifies the model parameters under the initial aging cycle, without needing to update the model parameters under each cycle, thus reducing the amount of computation and improving the computational efficiency of SOH estimation for energy storage batteries.

[0253] III. Model Feature Extraction

[0254] In practice, experimental data under different aging cycles and operating conditions are input into the electrochemical model. Since the model parameters are not updated, the SOC and terminal voltage output by the model will have errors, and these errors will gradually increase as the battery ages. For example, under FUDS conditions... Figure 5 and Figure 6 As shown.

[0255] Specifically, the model features related to battery aging are extracted through the following steps:

[0256] S301: Input experimental data under different operating conditions during different aging cycles into the electrochemical model after parameter identification;

[0257] S302: Calculate the simulated SOC value of the energy storage battery using an electrochemical model; calculate the SOC error (SOC error) by comparing the simulated and actual SOC values. error Extracting SOC error (SOC) error The mean and integral are used as model features F1 and model features F2;

[0258] The formula is expressed as:

[0259] F1 = mean(SOC) error );

[0260]

[0261] S303: Calculate the simulated voltage value of the energy storage battery using an electrochemical model; calculate the voltage error U by comparing the simulated voltage value with the actual voltage value. error Extraction terminal voltage error U error The mean and integral are used as model features F3 and model features F4;

[0262] The formula is expressed as:

[0263] F3=mean(U error );

[0264]

[0265] IV. Data Feature Extraction

[0266] In the specific implementation process, the following steps are used to extract data features related to battery aging:

[0267] S401: Obtain experimental data;

[0268] In this embodiment, the process for obtaining experimental data includes: conducting battery aging tests on a 1.1Ah lithium iron phosphate battery at 25℃, including capacity testing, OCV testing, and tests under different dynamic operating conditions, including five dynamic operating conditions: DST, FUDS, UDDS, BJDST, and US06. An accelerated aging test is then conducted at 15℃, and the test ends when the failure threshold is reached.

[0269] S402: Extract discharge time as data feature F5 from experimental data;

[0270] The formula is expressed as:

[0271] F5 = t end -t0;

[0272] In the formula: t end t0 represents the discharge end time; t0 represents the discharge start time.

[0273] S403: Extract net discharge energy from experimental data as data feature F6;

[0274] The formula is expressed as:

[0275] F6=∫U d I d dt d -∫U c I c dt c

[0276] In the formula: U d and U c These represent the voltages corresponding to the discharge current and the charging current, respectively; I d and I c These represent the discharge current and the charging current, respectively; t d and t c These represent the discharge time and charging time, respectively.

[0277] S404: Extract the voltage change rate as a data feature F7 from experimental data;

[0278] The formula is expressed as:

[0279]

[0280] In the formula: U end U0 represents the voltage at the end of the discharge; U0 represents the voltage at the beginning of the discharge; Δt represents the time from the beginning to the end of the discharge.

[0281] In addition to extracting model features reflecting the internal structure of the battery, this invention also extracts data features reflecting the external structure of the battery based on experimental data. Compared with traditional data-driven methods, the method of this invention, by simultaneously extracting features reflecting both the internal and external aspects of the battery, enables the machine learning model to acquire more comprehensive and complete information, improves the model's sensitivity to changes in battery performance, and enhances prediction accuracy and robustness.

[0282] V. Feature Fusion Generation

[0283] In practice, features extracted based on models and data show a strong correlation with battery aging, but the differences in features under different operating conditions are significant. Therefore, combining... Figure 7 As shown, the fused features are generated through the following steps:

[0284] S501: Put all model features and data features into one dataset;

[0285] S502: The number of features varies under different operating conditions, so all features in the dataset need to be normalized.

[0286] The formula is expressed as:

[0287]

[0288] In the formula: x normalized x represents the normalized feature; x represents the original feature; x max and x min These represent the maximum and minimum values ​​of the features in the dataset, respectively.

[0289] S503: Perform correlation analysis on the normalized features of the dataset using the Pearson correlation coefficient;

[0290] The formula for correlation analysis is:

[0291]

[0292] In the formula: ρ X,Y The Pearson correlation coefficient is represented by cov(X,Y); cov(X,Y) represents the covariance matrix of two continuous variables X and Y, and σ X and σ Y Let X and Y represent the standard deviations respectively; E represents the expected value; μ X and μ Y These are the means of X and Y, respectively;

[0293] S504: Perform principal component analysis (PCA) dimensionality reduction on the five target features with the highest correlation coefficients, and select the first principal component with the largest contribution rate as the fusion feature after dimensionality reduction;

[0294] The steps involved in principal component analysis dimensionality reduction include:

[0295] S5041: Centralize the target features;

[0296] The formula is expressed as:

[0297]

[0298] In the formula: Represents the input features X1, X2, X3, ..., X n The mean of y1, y2, ... y2; n This represents the target characteristics after centralized processing.

[0299] S5042: Calculate the covariance matrix based on the target features after centralization;

[0300] The formula is expressed as:

[0301]

[0302] In the formula: cov(y1,y2) represents the calculation of covariance;

[0303] S5043: Solving for the eigenvalues ​​and eigenvectors of the covariance matrix through eigenvalue decomposition;

[0304] S5044: Arrange the eigenvectors into a matrix from top to bottom according to the magnitude of their corresponding eigenvalues, and take the first k columns to form the principal components V. k ;

[0305] S5045: Multiply the original dataset by the principal component V k The features in the dataset are projected into a low-dimensional space to obtain the dimensionality-reduced data, which is the contribution rate of the features in the principal component space.

[0306] The formula is expressed as:

[0307] W k =X·V k ;

[0308] In the formula: W k The features are represented by X after dimensionality reduction; X represents the original dataset.

[0309] S5046: Select the feature with the highest contribution rate as the fusion feature after dimensionality reduction.

[0310] VI. Support Vector Regression Model

[0311] In this embodiment, the support vector machine model selected is the support vector regression model (SVR).

[0312] The working logic of the support vector regression model can be represented as follows:

[0313] 1) The original dataset is mapped to a high-dimensional feature space using a mapping function Φ(x), and linear regression is performed in the feature space:

[0314] f(x) = wΦ(x) + b;

[0315] In the formula: w represents the hyperplane normal vector; b represents the hyperplane parameter;

[0316] 2) Define the ε-linear insensitive loss function:

[0317]

[0318] In the formula: f(x) represents the predicted value of battery health status; y represents the actual value of battery health status; ε is a threshold that determines the model's tolerance to error;

[0319] 3) Based on the idea of ​​maximizing the margin and minimizing the loss of the support vector machine, the support vector regression model is expressed as:

[0320]

[0321] In the formula: C represents the penalty factor; ζ i and These represent the extent to which each sample exceeds the upper and lower intervals, respectively; m represents the number of samples.

[0322] 4) Introducing the Lagrange function and converting it to its dual form, we obtain the regression function as follows:

[0323]

[0324] In the formula: a i and Dual variable, x is the bias value. i Let x represent the eigenvector. j The center of the kernel function, K(x) i ,x j ) represents the radial basis function kernel function; σ represents the width parameter of the function, which controls the radial range of the kernel function.

[0325] Unlike other complex neural network models, this invention does not require a large dataset for training, nor does it require a deep network architecture or high-performance computing resources. It can achieve efficient and accurate estimation of battery health status using a simple machine learning algorithm.

[0326] VII. Model Training

[0327] In the specific implementation process, combined with Figure 8 As shown, the support vector regression model is trained using the k-fold cross-validation algorithm. The training steps include:

[0328] S601: Obtain a dataset containing input features and target variables; divide the dataset into k non-overlapping subsets;

[0329] S602: For each of the k subsets, perform the following steps:

[0330] S6021: Use the current subset as the test set and the other k-1 subsets as the training set;

[0331] S6022: Train a support vector regression model on the training set;

[0332] S6023: Evaluate the performance of the support vector regression model on the test set and record the performance metrics;

[0333] S6024: Repeat S6021 to S6023, average the performance metrics obtained from k iterations, and obtain an overall evaluation of the performance of the support vector regression model.

[0334] S603: Based on the results of k-fold cross-validation, compare the average performance indicators under different parameter combinations; select the parameter combination with the best performance indicators (such as minimum MSE and maximum R2) as the optimal parameters for the support vector regression model;

[0335] S604: The optimal parameters of the support vector regression model were determined by training the entire dataset until the model converged.

[0336] Specifically, the performance of the support vector regression model is evaluated using root mean square error (RMSE) and mean absolute error (MAE).

[0337] The formula is expressed as:

[0338]

[0339] In the formula, N represents the number of samples; This represents the predicted battery health status output by the support vector regression model; y m This indicates the true value of the battery's health status.

[0340] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Those skilled in the art should understand that any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A method for estimating the state of energy (SOH) of an energy storage battery based on electrochemical models and machine learning, characterized in that, include: S1: Construct an electrochemical model for energy storage batteries; S2: Parameter identification of the electrochemical model under initial aging cycles; S3: Extracting model features related to battery aging based on the electrochemical model after parameter identification; In step S3, model features related to battery aging are extracted through the following steps: S3 01: Input experimental data under different aging cycles and different operating conditions into the electrochemical model after parameter identification; S302: Calculate the simulated SOC value of the energy storage battery using an electrochemical model; calculate the SOC error by comparing the simulated and actual SOC values. Extracting SOC error The mean and integral are used as model features. and model features ; The formula is expressed as: ； ； S303: Calculate the simulated voltage value of the energy storage battery using an electrochemical model; calculate the voltage error by comparing the simulated voltage value with the actual voltage value. Extraction terminal voltage error The mean and integral are used as model features. and model features ; The formula is expressed as: ； ； S4: Extract data features related to battery aging based on the acquired experimental data; In step S4, data features related to battery aging are extracted through the following steps: S4 01: Obtain experimental data; S402: Extracting discharge time as a data feature from experimental data ; The formula is expressed as: ； In the formula: Indicates the discharge end time; Indicates the discharge start time; S403: Extracting net discharge energy as a data feature from experimental data. ; The formula is expressed as: ； In the formula: and These represent the voltages corresponding to the discharge current and the charging current, respectively. and These represent the discharge current and the charging current, respectively. and These represent the discharge time and charging time, respectively. S404: Extracting voltage change rate as a data feature from experimental data ; The formula is expressed as: ； In the formula: Indicates the voltage at the end of the discharge; This indicates the voltage at the start of discharge; This indicates the time from the start to the end of the discharge. S5: Perform correlation analysis, feature dimensionality reduction, and feature fusion on model features and data features to obtain fused features; S6: Input the fused features into the trained support vector machine model and output the corresponding battery health status prediction value; In step S6, the support vector machine model is a support vector regression model; The support vector regression model is trained using the k-fold cross-validation algorithm. The training steps include: S601: Obtain a dataset containing input features and target variables; divide the dataset into k non-overlapping subsets; S602: For each of the k subsets, perform the following steps: S6021: Use the current subset as the test set and the other k-1 subsets as the training set; S6022: Train a support vector regression model on the training set; S6023: Evaluate the performance of the support vector regression model on the test set and record the performance metrics; S6024: Repeat S6021 to S6023, average the performance metrics obtained from k iterations, and obtain an overall evaluation of the performance of the support vector regression model. S603: Based on the results of k-fold cross-validation, compare the average performance indicators under different parameter combinations; select the parameter combination with the best performance indicators as the optimal parameters for the support vector regression model; S604: The optimal parameters of the support vector regression model were determined by training the entire dataset until the model converged.

2. The SOH estimation method for energy storage batteries based on electrochemical models and machine learning as described in claim 1, characterized in that: In step S1, the electrochemical model of the energy storage battery is constructed through the following steps: S101: The relationship between the lithium-ion flux generated by the electrochemical reaction on the surface of the active particles of the negative and positive electrodes of the energy storage battery and the external current is calculated as follows: ； ； In the formula: This indicates the relationship between the lithium-ion flow rate generated by the electrochemical reaction on the surface of the negative electrode active particles of an energy storage battery and the external current. This indicates the relationship between the lithium-ion flow rate generated by the electrochemical reaction on the surface of the positive electrode active particles of an energy storage battery and the external current. Indicates external current; Indicates the thickness of the negative electrode; This indicates the area of ​​the negative electrode plate; This represents the specific surface area of ​​the negative electrode active particles; Denotes Faraday's constant; Indicates the thickness of the positive electrode; This indicates the area of ​​the positive electrode plate; This represents the specific surface area of ​​the positive electrode active particles; S102: The lithium-ion concentration on the particle surface of the positive and negative electrodes was obtained using the second-order Runge-Kutta method. ； In the formula: and These represent the lithium-ion concentrations on the particle surfaces of the positive and negative electrodes, respectively. and Indicates the solid-phase diffusion coefficient of the positive and negative electrodes; S103: Calculate the electrode utilization rate of the positive and negative electrodes based on the lithium ion concentration on the particle surface, and then calculate the open circuit voltage; The formula for calculating electrode utilization rate is as follows: ； ； In the formula: , These represent the electrode utilization rates of the positive and negative electrodes, respectively. and These represent the maximum solid-phase lithium-ion concentrations of the positive and negative electrodes, respectively. The formula for the positive open-circuit voltage is: ； The formula for the open-circuit voltage of the negative terminal is expressed as: ； S104: The formula for defining liquid phase potential is: ； In the formula: Represents the liquid phase potential; Indicates the electrode surface area; Indicates effective conductivity; , and These represent the thicknesses of the positive electrode, negative electrode, and separator, respectively. S105: The steps in S101... and Substituting into the Butler-Volmer equation, we get: ； ； In the formula: and Represents the electrochemical reaction rate constants at the negative and positive electrodes; Indicates the concentration of lithium ions in the liquid phase. and Indicates the lithium-ion concentration in the solid-liquid phase; Indicates the transmission coefficient; Represents the gas constant; Represents the temperature constant; and These represent the overpotentials at the positive and negative terminals, respectively. Among them, auxiliary variables and for: ； ； The magnitude of the overpotential generated by the positive and negative electrode reactions is calculated as follows: ； S106: Establish an electrochemical model to estimate the battery's terminal voltage and state of charge (SOC); The formula for calculating the terminal voltage in the electrochemical model is expressed as follows: ； In the formula: Indicates terminal voltage; Indicates the internal resistance of the ohm; Indicates the open-circuit voltage of the positive terminal; Indicates the open-circuit voltage of the negative terminal; This indicates the overpotential generated by the positive electrode reaction; This indicates the overpotential generated by the negative electrode reaction; I represents the liquid phase potential; I represents the current. The state of charge (SOC) in the electrochemical model is expressed using the electrode utilization rate of the positive electrode as follows: ； In the formula: Indicates the SOC value; , These represent the utilization rates of the spherical particle surface electrodes when the battery is fully discharged and fully charged, respectively. This indicates the electrode utilization rate of the positive electrode.

3. The SOH estimation method for energy storage batteries based on electrochemical models and machine learning as described in claim 1, characterized in that: In step S2, the parameters of the electrochemical model are identified using a particle swarm optimization algorithm under the initial aging cycle. The processing steps of the particle swarm optimization algorithm include: S201: Determine the model parameters that need to be identified; S202: Define each particle as a set of model parameters to be identified; S203: Initialize particle position and velocity: ； ； In the formula: and These represent the maximum solid-phase lithium-ion concentrations of the negative and positive electrodes, respectively. and These represent the radii of the solid particles at the negative and positive electrodes, respectively. and These represent the electrochemical reaction rate constants at the negative and positive electrodes, respectively. and These represent the particle's position and velocity, respectively. S204: Define the minimum root mean square error of velocity as the fitness function; The formula is expressed as: ； In the formula: Represents the fitness function. and These are represented as the simulated terminal voltage and the measured terminal voltage, respectively. S205: Calculate the fitness value of each particle; compare the fitness of the current particle position with its historical best position; if the current position is better, update the individual's best position. S206: Find the particle with the best fitness among all particles and update the global optimal position; S207: Update the velocity and position of each particle based on inertial weights, acceleration constants, and individual and global best positions; The formula is expressed as: ； In the formula: Indicates the particle at the 1st The position vector in the next iteration; Indicates the particle at the 1st The velocity vector in the next iteration; Indicates the particle in the first... The historical best position in the next iteration; Indicates the group in the th The historical best position in the next iteration; Indicates inertia weight; and These represent individual learning factors and group learning factors, respectively. and These are random numbers within the range [0,1]. S208: Repeat steps S205 to S207 until the preset number of iterations is reached or the convergence condition is met. Output the model parameters corresponding to the optimal particle as the parameter identification result of the electrochemical model.

4. The SOH estimation method for energy storage batteries based on electrochemical models and machine learning as described in claim 3, characterized in that: In step S201, by using the controlled variable method, while ensuring that other parameters of the electrochemical model remain unchanged, the value of a certain parameter is changed to analyze the battery voltage characteristics of the electrochemical model, and several parameters with the highest parameter sensitivity are selected as model parameters to be identified.

5. The SOH estimation method for energy storage batteries based on electrochemical models and machine learning as described in claim 1, characterized in that: In step S5, the fusion features are generated through the following steps: S5 01: Put all model features and data features into one dataset; S502: Normalize all features in the dataset; The formula is expressed as: ； In the formula: Represents the normalized features; Indicates original characteristics; and These represent the maximum and minimum values ​​of the features in the dataset, respectively. S503: Perform correlation analysis on the normalized features of the dataset using the Pearson correlation coefficient; The formula for correlation analysis is: ； In the formula: This represents the Pearson correlation coefficient; Represents two consecutive variables and The covariance matrix, and They represent and Standard deviation; Expressing expectations; and These are the means of X and Y, respectively; S504: Perform principal component analysis to reduce the dimensionality of several target features with the highest correlation coefficients, and select the feature with the largest contribution rate as the fusion feature after dimensionality reduction; The steps involved in principal component analysis dimensionality reduction include: S5041: Centralize the target features; The formula is expressed as: ； In the formula: Representing input features The mean; , ,… This represents the target characteristics after centralized processing. S5042: Calculate the covariance matrix based on the target features after centralization; The formula is expressed as: ； In the formula: The expression calculates the covariance; S5043: Solving for the eigenvalues ​​and eigenvectors of the covariance matrix through eigenvalue decomposition; S5044: Arrange the eigenvectors into a matrix from top to bottom according to the magnitude of their corresponding eigenvalues, and take the first few... Columns form principal components ; S5045: Multiply the original dataset by the principal components. The features in the dataset are projected into a low-dimensional space to obtain the dimensionality-reduced data, which is the contribution rate of the features in the principal component space. The formula is expressed as: ； In the formula: This represents the data after dimensionality reduction; Represents the original dataset; S5046: Select the feature with the highest contribution rate as the fusion feature after dimensionality reduction.

6. The SOH estimation method for energy storage batteries based on electrochemical models and machine learning as described in claim 1, characterized in that: In step S6023, the performance of the support vector regression model is evaluated by the root mean square error and the mean absolute error. The formula is expressed as: ； ； In the formula: Indicates the number of samples; This represents the predicted battery health status output by the support vector regression model. This indicates the true value of the battery's health status.

Images