CN111308379B - Battery health state estimation method based on local constant voltage charging data - Google Patents

Abstract

The invention discloses a battery state-of-health estimation method based on local constant-voltage charging data, comprising the following steps: S1. Carrying out cyclic charge-discharge experiments on the battery under the conditions of different temperatures and charging rates, and measuring the current and terminal voltage of the battery in real time; S2. Under the conditions of different temperatures and CC (constant current) charging rates, fit the linear relationship between the CV (constant voltage) charging capacity and the CCCV (constant current and then constant voltage) charging capacity, and establish a parameter mapping database for the linear model; S3 . Establish a charging current prediction model in the CV stage, use the local CV charging data, identify the model parameters, and predict the entire CV charging data; S4. According to the actual charging process temperature and CC charging rate, select the corresponding linear model from the parameter mapping database, Calculate the SOH (State of Health) of the battery based on the estimated CV charge capacity. The invention has low calculation cost, can overcome the limitation of the traditional method requiring complete charging data, and can estimate the SOH of the battery with high accuracy only through the partial CV charging data.

Description

Battery health state estimation method based on local constant voltage charging data

Technical Field

The present invention relates to state of health estimation of lithium ion batteries, and more particularly, to a battery state of health (SOH) estimation method based on local Constant Voltage (CV) charging data.

Background

Lithium ion batteries have the advantages of high energy density, long service life and the like, and are now the main energy storage tools of electric automobiles. However, as the number of cycles increases, the internal physical and chemical processes of the lithium ion battery change, and the capacity and performance of the lithium ion battery are continuously attenuated along with various side reactions, and even become invalid in severe cases, so that safety accidents are caused. The performance attenuation of the lithium ion battery system seriously influences the endurance mileage of the electric automobile, and greatly hinders the further popularization and promotion of the electric automobile. Therefore, accurately detecting the SOH of the battery is very important to the reliability of the electric vehicle.

Currently, methods for estimating SOH of a lithium ion battery are mainly classified into three types. The first is electrochemical measurement, which is mainly to measure impedance information of a battery through Electrochemical Impedance Spectroscopy (EIS) and map the impedance information to SOH; the method can describe the impedance more accurately, but the measuring process is complicated, a professional measuring instrument is needed, and the method is difficult to be applied practically. The second method is an estimation method based on a model, an electric model of the lithium ion battery is established by relying on theoretical supports such as Butler-Volmer law, kirchhoff law and the like, and parameters such as a plurality of states, capacity and the like are estimated by adopting filter algorithms such as extended Kalman filtering and the like; the method has the advantages of high estimation precision and strong robustness, has good universal applicability to different batteries, but has high requirements on model precision and high calculation cost, and has challenges in online application. The third method is an estimation method based on data driving, and utilizes battery voltage and current information which is convenient to measure to extract characteristic parameters with high correlation with battery aging so as to estimate the SOH of the battery; the method does not need a complex model, and reduces the calculation cost while ensuring the estimation precision. However, such methods require charging or discharging data in the full state of charge (SOC) range, and most of the practical applications of batteries can only obtain fragment data, so the environmental adaptability of such methods needs to be enhanced.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provides a battery health state estimation method based on local constant voltage charging data. Calibrating a linear relation between the CV charging capacity and the CCCV charging capacity, predicting current change characteristics of the residual section CV charging by adopting a battery second-order RC equivalent circuit model according to local CV charging data of the battery aiming at a typical scene that the CV charging stage is incomplete, further extracting a health factor and estimating the SOH of the battery. The method provided by the invention has low calculation complexity, overcomes the limitation that the traditional method needs complete charging data, and can realize high-precision estimation of the SOH of the battery only according to the charging data of the local CV.

The purpose of the invention is realized by the following technical scheme: a battery state of health estimation method based on local constant voltage charging data includes the following steps:

s1, carrying out a cyclic charge-discharge experiment on a battery under different temperature and charge multiplying power conditions, and measuring the current and terminal voltage of the battery in real time by adopting a current sensor and a voltage sensor;

s2, extracting the CV charging capacity by adopting an ampere-hour integration method, fitting a linear relation between the CV charging capacity and the CCCV charging capacity under different temperatures and CC charging multiplying power conditions, and establishing a parameter mapping database of a linear model;

wherein CV represents constant voltage, CC represents constant current, CCCV represents constant current first and then constant voltage;

s3, establishing a second-order RC equivalent circuit model, establishing a CV stage charging current prediction model on the basis of the second-order RC equivalent circuit model, identifying model parameters by using local CV charging data, and predicting the whole CV charging data;

and S4, according to the temperature and the CC charging rate in the actual charging process, selecting a corresponding linear model from the parameter mapping database in the step S2, and calculating the SOH of the battery according to the estimated CV charging capacity, wherein the SOH represents the state of health.

The invention has the beneficial effects that: the invention provides a battery SOH estimation method based on local CV charging data_CV-Q_CCCVIs calculated from the linear relationship of_CCCVAnd then estimating the SOH of the battery. The method of the invention has three advantages: first, the health factor Q can be directly extracted_CVThe conversion of a complex capacity increment curve, a differential voltage curve and the like is not needed, so that the calculation cost can be reduced; second, Q in the method of the present invention_CV-Q_CCCVHas good linear relation, does not need complex machine learning algorithm, and can realize the SOH of the battery only through a simple linear modelHigh-precision estimation; and thirdly, the method only needs partial data of the CV charging stage, and has better practical application prospect.

Drawings

FIG. 1 is a flow chart of a method of the present invention;

FIG. 2 shows example Q_CVAnd Q_CCCVThe fitting curve of (1);

FIG. 3 is a schematic circuit diagram of a second-order RC equivalent circuit model in the embodiment;

FIG. 4 is a predicted curve of CV of a battery according to a second-order RC equivalent circuit model in an embodiment;

FIG. 5 shows the prediction error of the CV prediction curves in the examples.

Detailed Description

The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.

As shown in fig. 1, a method for estimating SOH of a battery based on local CV charge data includes the steps of:

s1, carrying out a cyclic charge-discharge experiment on a battery under different temperature and charge multiplying power conditions, and measuring the current and terminal voltage of the battery in real time by adopting a current sensor and a voltage sensor;

the step S1 includes the following sub-steps:

s101, charging the battery to the state of charge (SOC) of 100% by adopting a constant-current-constant-voltage charging method, and monitoring the current I of the battery in real time in the charging process_LAnd terminal voltage U_lAnd storing the voltage U at the CC stage_lThe current I at the CV stage is up to the preset cut-off value_LReaching a preset cut-off value;

s102, performing a discharge experiment by adopting a CC discharge method until the terminal voltage is reduced to a lower limit cut-off voltage, wherein the CC discharge multiplying power and the CC charge multiplying power do not need to be kept consistent;

s103, repeating S101-S102 until the rated capacity of the battery is reduced to the end-of-life standard, and the end voltage U at the CV stage in different cycles_lNeed to be kept consistent, integrate the current and terminal voltage information collected during the process, establish a specific temperature and CC charge rateA battery aging database;

s104, repeating S101-S103 at different temperatures and different CC charging multiplying factors to obtain battery aging databases at different temperatures and different CC charging multiplying factor variables.

In the embodiment of the present application, after step S103 or S104 is finished, data preprocessing for missing value padding and error value deletion needs to be performed on the battery aging database. The error value comprises a value with overlarge deviation compared with the previous and next data (the deviation exceeds a preset threshold value compared with the previous and next data), and can be directly deleted during preprocessing; the missing value includes the condition that the current value and the voltage value at a certain moment are vacant, and the preprocessing mode includes but is not limited to padding by the average value of the previous data and the next data.

S2, extracting the CV charging capacity by adopting an ampere-hour integration method, fitting a linear relation between the CV charging capacity and the CCCV charging capacity under different temperatures and CC charging multiplying power conditions, and establishing a parameter mapping database of a linear model;

wherein CV represents constant voltage, CC represents constant current, CCCV represents constant current first and then constant voltage;

s201, selecting CV stage charging data of a plurality of cycles under the same temperature and CC charging rate from a battery aging database, and extracting CV charging capacity Q of each selected cycle by adopting an ampere-hour integration method_CVAs a health factor;

s202, calculating the CCCV charging capacity Q of each extracted cycle by adopting an ampere-hour integration method_CCCV；

The calculation method of the charge capacity described in steps S201 and S202 is as follows:

wherein m is the initial moment of CV charging, n is the moment when the CV charging current reaches the preset cut-off value, I_L(k) For the current value at the k-th time, Δ t load current sampleA time interval.

S203, fitting Q off line by adopting a least square method_CV-Q_CCCVThe formula is as follows:

Q_CCCV＝aQ_CV+b

wherein a and b are fitting parameters obtained by least square fitting. The fitting method comprises the following specific steps:

according to Q_CVThe measured values of (A) were as follows: q_CCCV＝aQ_CV+ b calculates the corresponding Q_CCCVIs marked as

Calculating the measured value Q directly obtained from the experimental data_CCCVAnd a calculated value obtained by the calculation

The square sum R of the dispersion of (a), the formula is as follows:

substituting the linear equation to be fitted into the formula:

wherein Q is_CCCViAnd Q_CViRespectively represent the ith Q_CCCVAnd Q_CVAnd (4) data. Then, the fitting parameters a and b are respectively subjected to partial derivation:

will measure the value Q_CCCVAnd Q_CVCalculated by substituting the formula when

And when the sum is 0, the values of a and b are obtained. In this embodiment, Q_CV-Q_CCCVThe linear function of (a) is:

Q_CCCV＝-3.5853Q_CV+3.5344

the fitted curve is shown in FIG. 2;

s204, repeating S201-S203, obtaining the linear relation of the CV charging capacity and the CCCV charging capacity under different temperature and CC charging multiplying factor variables, namely a plurality of groups of [ a, b ] parameter sets under different temperature and CC charging multiplying factors, and forming a parameter mapping database of a linear model.

S3, establishing a battery second-order RC equivalent circuit model, establishing a CV stage charging current prediction model on the basis of the battery second-order RC equivalent circuit model, identifying model parameters by using local CV charging data, and predicting the whole CV charging data through a battery model;

s301, acquiring current and terminal voltage data in the CCCV charging process in real time by adopting a current and voltage sensor under an actual charging working condition; defining the time when the terminal voltage reaches the upper limit cut-off voltage and the current starts to drop as the algorithm starting time;

s302, establishing a second-order RC equivalent circuit model of the battery, wherein a circuit schematic diagram of the second-order RC equivalent circuit model is shown in FIG. 3;

the circuit equation of the second-order RC equivalent circuit model of the battery is as follows:

U_OC+R₀I_L+U_p1+U_p2＝U_l

wherein, U_p1、U_p2To polarize the voltage, U_lIs terminal voltage, U_ocBeing batteriesOpen Circuit Voltage (OCV), I_LIs a current, R_p1、R_p2、C_p1、C_p2And R₀Model parameters to be identified, specifically: r_p1、R_p2Is a polarization resistance, C_p1、C_p2Is a polarization capacitance, R₀Ohmic internal resistance;

s303, establishing a prediction equation of the CV charging curve based on a battery second-order RC equivalent circuit model:

firstly, performing laplace transform on an equation of a second-order RC equivalent circuit model:

where s is the complex variable of the laplace transform. After a relational expression of the difference value of the current, the terminal voltage and the open circuit voltage is established, inverse Laplace transform is carried out on the equation:

wherein a ═ R₀R_p1R_p2C_p1C_p2，b＝R₀R_p1C_p1+R₀R_p2C_p2+C_p1R_p1R_p2+C_p2R_p1R_p2，c＝R₀+R_p2+R_p1. The calculated CV current is expressed as:

wherein, I_LThe current value at time t (positive value during charging and negative value during discharging) is shown as (t). Open circuit voltage U of battery due to CV charging stage_ocIs generally close to the battery terminal voltage U_lResult in (U)_l-U_oc) The value of (c) is small. For simple calculation, in the CV stage charging current prediction model, a/c (U) is ignored_l-U_oc) To I_L(t), i.e., establishing a CV-stage charging current prediction model as follows:

wherein, I_L1、I_L2、τ_eq,1、τ_eq,2In order to predict the parameters to be identified in the equation, the identification method adopts a Levenberg-Marquardt (LM) algorithm, and comprises the following steps:

setting the calculated value of current calculated according to the prediction equation as

Calculating the measured value I directly obtained from the experimental data_LAnd the calculated value

The square sum R of the dispersion of (a), the formula is as follows:

calculate the Jacobian matrix J (x):

wherein x is_iRepresenting the parameters of the ith row in vector x. Subsequently, the gradient of R is calculated as:

g＝R′(x)＝J(x)^Tf(x)

then, solving an iteration step h, wherein the formula is as follows:

wherein, J_k＝J(x_k)，f_k＝f(x_k)，x_kFor the value of x after iterating k times, I is an identity matrix, u is a positive number, and the function of shortening the iteration step length is realized in the formula, and the determination method is as follows:

where ρ is a gain ratio, and when ρ is greater than 0, the formula is as follows:

when ρ is equal to or less than 0, the formula is as follows:

u_k+1＝u_k×v_k

v_k+1＝2v_k

in this embodiment, the initial value of v is 2, and the initial value of u is calculated as follows:

A₀＝J(x₀)^TJ(x₀)

u₀＝τ×max{a_ii}

wherein, a_iiIs a matrix A₀The elements on the diagonal. Repeating the iterative process:

x_k+1＝x_k-h

when the iteration process meets one of the following conditions, exiting the iteration;

g≤ε₁

‖h‖≤ε₂(‖x‖+ε₂)

wherein | represents the matrix norm, ε₁、ε₂Any small value can be selected as the termination condition for the preset parameter, and the value range can be 10^-8～10^-12The concrete adjustment can be made according to the actual situation; when the iteration is judged to meet the termination condition, I at the moment_L1、I_L2、τ_eq,1、τ_eq,2The value of (a) is the identification result, and the parameter identification result in this embodiment is: i is_L1＝0.9545、I_L2＝0.5455、τ_eq,1＝769.23、τ_eq,2＝2564.1；

S304, if the CV charging is not cut off in advance, namely the charging is finished until the current is reduced to the CV cut-off current, recording the complete CV charging capacity according to an ampere-hour integration method; if the CV charging is cut off in advance, that is, the current does not decrease to the CV cut-off current at the end of CV, the change of the future current is calculated according to the charging current prediction model in S303 until the CV cut-off current is reached, and the complete CV charging capacity is calculated according to an ampere-hour integration method.

S4, according to the temperature and the CC charging multiplying power in the actual charging process, selecting a corresponding linear model from the parameter mapping database in the S2, and calculating the SOH of the battery according to the estimated CV charging capacity, wherein the specific implementation scheme is as follows:

selecting a corresponding linear model in the parameter mapping database in S204 according to the actual temperature and the CC charging multiplying power; q acquired in S304_CVSubstitution into a Linear model to calculate Q_CCCVAnd further calculating the SOH of the battery at the current moment, wherein the formula is as follows:

Q_CCCV＝-3.5853Q_CV+3.5344

where S is the estimation of battery SOH, Q_ratedIs the nominal capacity of the battery.

And estimating the health state of the battery in real time according to the steps, comparing the health state of the battery with the calculated health state of the battery according to a preset service life ending standard, and judging whether the current health state of the battery meets the safety standard or not.

In the embodiment of the patent, 18650 lithium ion batteries with nominal capacity of 2Ah are taken as experimental objects, the method S1 is carried out at different temperatures, and battery aging databases with different temperatures and CC discharge rates are established. Next, Q is calculated by the ampere-hour integration method_CVAnd Q_CCCVFitting Q by least squares_CVAnd Q_CCCVThe linear relationship of (a) to (b) constitutes a parameter mapping database. Aiming at a typical scene that CV charging is incomplete, a second-order RC equivalent circuit model is built, and on the basis, a CV stage charging current prediction model is established. And then, estimating the SOH of the battery to be tested. The upper limit cut-off voltage of the terminal voltage of the battery to be tested is 4.2A, CV, the lower limit cut-off current of the current is 0.05A, the constant current of the discharge experiment is 2A, and the lower limit cut-off voltage is 2.5V. When the current data at the CV stage is complete, the ampere-hour integration method can be directly adopted to calculate Q_CV(ii) a And when the current of the battery to be measured in the CV stage does not fall to the CV lower limit cut-off current, fitting local CV current data by using the established charging current prediction model, and predicting the residual CV current until the CV cut-off current is reached. The CV prediction curve of the charging current prediction model in the embodiment is shown in fig. 4, and the prediction error is shown in fig. 5. Finally, the complete CV capacity was calculated to be 0.5507Ah according to the ampere-hour integration method. Selecting corresponding linear model, and calculating to obtain Q_CCCVIt was 1.56 Ah. Finally pass through Q_CCCVThe ratio of the calibrated capacity determines the SOH of the cell to be 78%. And Q obtained by actual measurement_CCCV1.6109Ah, the estimated value differs from 0.0509Ah with a relative error of 3.1%. The actual battery SOH was 80.55%, and the relative error of the estimate was within 3%. Therefore, the method can estimate the health state of the battery based on the local CV charging data, and has low calculation cost and small data requirement under the condition of meeting the precision requirement.

In summary, the invention provides a battery SOH estimation method based on local CV charging data, and the high-precision estimation of the health state of a battery can be realized only according to the local CV charging data through a model prediction method, so that the limitation that the traditional method needs complete charging data is overcome. The method only involves a voltage sensor and a current sensor, and does not need complex experimental equipment. Q in the process_CVAnd Q_CCCVThe calculation cost is low, and the conversion of a complex capacity increment curve, a differential voltage curve and the like is not needed. Q of the process_CVAnd Q_CCCVThe method has good linear relation, does not need a complex machine learning algorithm, and can realize the high-precision estimation of the SOH of the battery only through a simple linear model.

The foregoing is a preferred embodiment of the present invention, it is to be understood that the invention is not limited to the form disclosed herein, but is not to be construed as excluding other embodiments, and is capable of other combinations, modifications, and environments and is capable of changes within the scope of the inventive concept as expressed herein, commensurate with the above teachings, or the skill or knowledge of the relevant art. And that modifications and variations may be effected by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (6)

1. A method for estimating battery state of health based on local constant voltage charging data, characterized in that it comprises the following steps: S1. Under the conditions of different temperatures and CC charging rates, the battery is subjected to a cyclic charge-discharge experiment of constant current and then constant voltage, and the current sensor and voltage sensor are used to measure the current and terminal voltage of the battery in real time; S2. Use the ampere-hour integration method to extract the CV charging capacity, fit the linear relationship between the CV charging capacity and the CCCV charging capacity under the conditions of different temperatures and CC charging rates, and establish a parameter mapping database of the linear model; Among them, CV means constant voltage, CC means constant current, and CCCV means first constant current and then constant voltage; S3. Establish a second-order RC equivalent circuit model, based on which, establish a CV stage charging current prediction model, use local CV charging data, identify model parameters, and predict the entire CV charging data; S4. According to the temperature of the actual charging process and the CC charging rate, select the corresponding linear model from the parameter mapping database described in step S2, estimate the CV charging capacity from the entire CV charging data predicted in step S3, and estimate the CV charging capacity according to the estimated CV charging capacity. Calculate the SOH of the battery, where SOH represents the state of health. 2. A method for estimating battery state of health based on local constant voltage charging data according to claim 1, wherein the step S1 comprises the following sub-steps: S101. Use the charging method of constant current and then constant voltage to charge the battery until the SOC reaches 100%, monitor the current _IL and terminal voltage _Ul of the battery in real time during the charging process, and save them, and the terminal voltage _Ul in the CC stage reaches the preset value Until the set cut-off value, the CV stage current _IL reaches the preset cut-off value; S102. Use the CC discharge method to conduct the discharge experiment until the terminal voltage drops to the lower cut-off voltage, and the CC discharge rate and CC charge rate do not need to be consistent; S103. Repeat S101-S102 until the rated capacity of the battery drops to the end-of-life standard. The terminal voltage U _l in the CV stage needs to be consistent in different cycles. The current and terminal voltage information collected during the process is integrated to establish the battery at a specific temperature and CC charging rate. aging database; S104. Repeat S101-S103 at different temperatures and CC charging rates to obtain a battery aging database under different temperature and CC charging rate variables. 3 . The method for estimating battery state of health based on local constant voltage charging data according to claim 2 , wherein the step S1 further comprises: after the end of step S103 or S104 , updating the battery aging database on the battery aging database. 4 . Data preprocessing for missing value imputation and error value removal. 4. The method for estimating battery state of health based on local constant voltage charging data according to claim 2, wherein the step S2 includes the following sub-steps: S201. Select the CV stage charging data of several cycles under the same temperature and CC charging rate from the battery aging database, and use the ampere-hour integration method to extract the selected CV charging capacity Q _CV of each cycle as a health factor; S202. adopt the ampere-hour integration method to calculate the CCCV charging capacity Q _CCCV of each cycle selected; S203. Use the least squares method to fit the linear relationship of Q _CV -Q _CCCV off-line. The formula is as follows: Q _CCCV = aQ _CV +b Among them, a and b are the fitting parameters obtained by the least squares fitting. The specific process of fitting is as follows: According to the measured value of Q _CV , calculate the corresponding Q _CCCV according to: Q _CCCV =aQ _CV +b, denoted as

Calculate the measured value Q _CCCV directly obtained from the experimental data and the calculated value obtained from the calculation

The sum of squares of the dispersion R, the formula is as follows:

Substitute the linear equation to be fitted into the formula:

Among them, Q _CCCVi and Q _CVi represent the i-th Q _CCCV and Q _CV data, respectively; then, the partial derivatives of the fitting parameters a and b are calculated respectively:

Substitute the measured values Q _CCCV and Q _CV into the above formula to calculate, when

When both are 0, the values of a and b are the required values; S204. Repeat S201-S203 to obtain the linear relationship between the CV charging capacity and the CCCV charging capacity under different temperature and CC charging rate variables, that is, multiple sets of [a, b] parameters under different temperatures and CC charging rates to form a linear model The parameter mapping database. 5. A method for estimating battery state of health based on local constant voltage charging data according to claim 4, wherein the step S3 includes the following sub-steps: S301. Use current and voltage sensors to collect real-time current and terminal voltage data during CCCV charging under actual charging conditions; define the time when the terminal voltage reaches the upper limit cut-off voltage and the current begins to drop as the algorithm start time; S302. Establish a second-order RC equivalent circuit model of the battery; The circuit equation of the battery second-order RC equivalent circuit model is as follows:

U _OC +R ₀ I _L +U _p1 +U _p2 =U _l Among them, U _p1 and U _p2 are polarization voltages, U _l is the terminal voltage, U _oc is the open circuit voltage of the battery, _IL is the current, and R _p1 , R _p2 , C _p1 , C _p2 and R ₀ are the models to be identified Parameters, specifically: R _p1 and R _p2 are polarization resistances, C _p1 and C _p2 are polarization capacitors, and R ₀ is ohmic internal resistance; S303. Based on the second-order RC equivalent circuit model of the battery, establish the prediction equation of the CV charging curve:

Among them, t is the time, I _L1 , I _L2 , τ _eq,1 , τ _eq,2 are the parameters to be identified in the prediction equation, and the identification method adopts the Levenberg-Marquardt algorithm, and the steps are as follows: Assuming that the calculated current calculated according to the prediction equation is

Calculate the measured value _IL and the calculated value obtained directly from the experimental data

The sum of squares of the dispersion R, the formula is as follows:

Among them, x is the parameter vector to be identified; in order to minimize the square sum of dispersion and find the best function matching of the data, the parameter fitting process of the Levenberg-Marquardt algorithm is as follows: First, calculate the gradient g of R(x): g=R'(x)=J(x) ^T f(x) where J(x) is the Jacobian matrix of R(x), and second, solve for the iteration step size h:

Among them, J _k =J(x _k ), f _k =f(x _k ), x _k is the value of x after k iterations, I is the identity matrix, and u is a positive number, which shortens the iteration step in the formula , then update the value of x _k : x _k+1 = x _k -h Repeat the iterative process. When one of the following conditions is satisfied during the iterative process, the required iteration accuracy is achieved, and the iteration is exited; g≤ε ₁ ‖h‖≤ε ₂ (‖x‖+ε ₂ ) Among them, ‖·‖ represents the matrix norm, ε ₁ , ε ₂ are preset parameters, as termination conditions; when it is judged that the iteration meets the termination conditions, at this time I _L1 , I _L2 , τ _eq,1 , τ _eq,2 The value of is the identification result; S304. If the CV charging is not terminated in advance, that is, charging until the current drops to the CV cut-off current, the complete CV charging capacity is recorded according to the ampere-hour integration method; if the CV charging is terminated early, that is, the current does not drop to the CV cut-off current at the end of the CV , then the future current change is estimated according to the prediction equation of the CV charging curve described in S303 until the CV cut-off current is reached, and then the complete CV charging capacity is calculated according to the ampere-hour integration method. 6. A method for estimating battery state of health based on local constant voltage charging data according to claim 5, wherein in step S4, the SOH of the battery is estimated according to the estimated CV charging capacity, specifically comprising: According to the actual temperature and the CC charging rate, the corresponding linear model is selected in the parameter mapping database described in S204; the whole section of VC charging data is predicted by the prediction equation in step S303, and the CV charging capacity Q _CV is estimated according to step S304, and the S304 The Q _CV obtained in , is substituted into the linear model to calculate Q _CCCV , and then the SOH of the battery at the current moment is calculated. The formula is as follows:

Among them, S is the estimated result of the battery SOH, and Q _rated is the rated capacity of the battery; According to the above steps, the state of health of the battery is estimated in real time, and whether the current state of health of the battery complies with the safety standard is compared and determined according to the preset end-of-life standard and the calculated state of health of the battery.

Images