Lithium battery SOH estimation method based on constant voltage charging time ratio

Through a method based on the proportion of constant voltage charging time, combined with charge and discharge cycle testing and SVR model, the problems of high calculation cost and measurement error in SOH estimation of lithium batteries are solved, and efficient and accurate SOH estimation is achieved, which is suitable for real-time applications.

CN119024210APending Publication Date: 2024-11-26GUANGDONG UNIV OF TECH
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411233022.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-04
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing lithium battery SOH estimation method has high calculation cost in some application scenarios, and voltage and current measurement errors affect the accuracy, making it difficult to achieve real-time and accurate SOH estimation.

Method used

Using a method based on the proportion of constant voltage charging time, the lithium battery is charged and discharge cycle test, the total time and constant voltage charging time of the charging process are collected, the charging time ratio is calculated, and the SVR model is used for training and estimation, which reduces the Calculate the amount of errors.

Benefits of technology

It greatly reduces the calculation amount of SOH estimation, improves estimation accuracy, is suitable for real-time applications, and reduces the impact of voltage and current measurement errors.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119024210A_ABST
    Figure CN119024210A_ABST
Patent Text Reader

Abstract

The invention discloses a lithium battery SOH estimation method based on a constant-voltage charging time ratio, and the method comprises the steps: firstly carrying out the constant-current charging and then constant-voltage charging of a battery until the battery is fully charged, and collecting the total charging time T and the constant-voltage charging time Tcv during the period; and dividing the constant-voltage charging time Tcv of each cycle of each battery by the total time T of the whole charging process to obtain a constant-voltage charging time ratio Tper of each cycle of each battery as a characteristic for estimating the SOH of the battery. Characteristic extraction is carried out on the charging time, and the calculation amount of complex characteristics is reduced, so that the SOH estimation precision of the lithium battery under some complex conditions is guaranteed, and meanwhile, the SOH estimation precision is not influenced by voltage and current measurement errors. According to the method, the characteristics which are low in calculation cost and easy to obtain are provided, and the problems that an existing characteristic extraction method is high in cost and high in complexity in some application scenes under an existing method, and errors of battery SOH estimation caused by voltage and current measurement errors are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery energy storage, and in particular to a lithium battery SOH estimation method based on constant voltage charging time proportion. Background Art

[0002] As my country's new energy vehicle industry enters an accelerated development stage, the safety of new energy vehicles has become particularly important, especially the life and health of its batteries. SOH is a key parameter of lithium batteries. As lithium batteries are used, their life will gradually decrease. Therefore, by accurately estimating the SOH of lithium batteries, lithium batteries can be replaced in time to reduce the risk of driving.

[0003] Existing methods for estimating the SOH of lithium batteries are generally data-driven. First, features are extracted from information related to lithium battery aging, such as current, voltage, temperature, charging time, etc., and then the features are input into a machine learning or deep learning network for learning before estimating the SOH of the lithium battery. Existing methods generally extract features from aging information such as the charging current, charging voltage, temperature in some specified intervals, and charging time of laboratory lithium batteries. However, in some application scenarios, high computational cost is a problem that hinders its real-time application, and voltage and current measurement errors also affect the accuracy of SOH estimation. Summary of the invention

[0004] In order to overcome the impact of high calculation cost and voltage and current measurement errors on SOH estimation accuracy in certain application scenarios, the present invention proposes a lithium battery SOH estimation method based on the proportion of constant voltage charging time.

[0005] The primary purpose of the present invention is to solve the above technical problems. The technical solutions of the present invention are as follows:

[0006] S1: Perform charge and discharge cycle test on N lithium batteries. In each cycle, first perform constant current charging, then constant voltage charging until the lithium battery is fully charged. Then collect the total time T of the entire charging process of N lithium batteries and the constant voltage charging time T cv , then discharge at a constant current to the discharge cut-off voltage, and calculate the SOH of the lithium battery in the current cycle according to the discharge capacity;

[0007] S2: For all lithium batteries in step S1, set the constant voltage charging time T of each lithium battery per cycle. cv Divide it by the total time of the entire charging process T, and get the constant voltage charging time ratio T of each lithium battery in each cycle per , and use it as a feature, and use the SOH value of each lithium battery after each cycle obtained in step S1 as a label to form a feature data set T perh ;

[0008] S3: Feature data set T obtained in step S2 perh Randomly select 80% of the data as training set T train , and then use the features of the remaining lithium batteries as the test set T test ;

[0009] S4: The training set T in step S3 train Input into the SVR model for training and testing. If the performance of the model meets the requirements, save the model. If not, modify the model parameters until the model training and testing meet the accuracy requirements.

[0010] S5: When the lithium battery is actually running, extract the constant voltage charging time T of the lithium battery according to the method of step S1 during the constant voltage charging stage. cv The total time of the whole charging process is T, and then the constant voltage charging time is T cv Divide by the total time T of the entire charging process to get the constant voltage charging time ratio T of each lithium battery charging process per , and then the constant voltage charging time ratio T per The data is input into the SVR model described in step S4 to estimate the SOH of the lithium battery.

[0011] 2. A lithium battery SOH estimation method based on constant voltage charging time ratio according to claim 1, characterized in that the total time T of the entire charging process and the constant voltage charging time ratio T in step S1 are per The construction method is as follows:

[0012] S1-1: Perform charge and discharge cycle tests on N lithium batteries. In each cycle, first perform constant current charging until the lithium battery terminal voltage reaches the charging cut-off voltage, and then perform constant voltage charging until the lithium battery is fully charged. In each cycle, collect the total charging time and constant voltage charging time of the nth lithium battery in the ath cycle, and calculate the constant voltage charging time ratio and arrange them in a row. The specific form is as follows:

[0013] T=[T 11 ,T 12 ,...,T 1a ,T 21 ,T 22 ,...,T 2a ,...,T n1 ,T n2 ,...,T na ] T

[0014] T cv =[T cv11 ,T cv12 ,...,T cv1a ,T cv21,T cv22 ,...,T 2a ,...,T cvn1 ,T cvn2 ...,T cvna ] T

[0015] T per =[T per11 ,T per12 ,...,T per1a ,T per21 ,T per22 ,...T per2a ,...,T pern1 ,T pern2 ...,T perna ] T

[0016] Among them, T na represents the total charging time of the nth lithium battery in the ath cycle; T cvna represents the constant voltage charging time in the ath cycle of the nth lithium battery; T perna It represents the proportion of the constant voltage charging time in the ath cycle of the nth lithium battery in the total charging time;

[0017] S1-2: In each discharge cycle, the SOH label value of the lithium battery is calculated based on the total discharge capacity. When the SOH of the lithium battery is lower than 70%, the cycle is stopped. The calculation formula of SOH is as follows:

[0018]

[0019] Among them, C now Represents the current cycle discharge capacity of the lithium battery, C initial Represents the initial capacity of the lithium battery.

[0020] Arrange the obtained SOH into a column to form a label sequence as follows:

[0021] SOH=[SOH 11 ,SOH 12 ,...,SOH 1a ,SOH 21 ,SOH 22 ,...SOH 2a ,...SOH n1 ,SOH n2 ...,SOH na ] T

[0022] Among them, SOH na represents the SOH in the ath cycle of the nth lithium battery;

[0023] S1-3: Construct a feature dataset T according to steps S1-1 and S1-2 perh as follows:

[0024]

[0025] 3. The method for estimating SOH of a lithium battery based on constant voltage charging time ratio according to claim 1 is characterized in that the data set division obtained in S3 and the training and estimation of the SVR model are specifically performed as follows:

[0026] S4-1: Feature dataset T obtained in S2 perh Randomly select 80% of the data as training set T train , and then use the features of the remaining lithium batteries as the test set T test ;

[0027] S4-2: Create a machine learning SVR, use rbf as the kernel function, set the penalty factor, epsilon and other parameters of the error term in the SVR, and perform training;

[0028] S4-3: Input the test set in step S3 into the trained model in S4-2 for estimation;

[0029] S4-4: Compare the estimated value of the lithium battery SOH output by the model in step S4-3 with the true value, and use the mean square error MSE and the mean absolute error MAE as the performance indicators of SVR. The calculation formulas of the mean square error MSE and the mean absolute error MAE are as follows:

[0030]

[0031] Among them, SOH real (a) represents the true value of SOH in the ath cycle, SOH pre (a) represents the estimated value of SOH in the ath cycle;

[0032] S4-5: If the mean square error MSE and mean absolute error MAE calculated in step S4-4 are within an acceptable range, the SVR model meets the accuracy requirements, otherwise repeat step S4 to retrain the model. Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0033] The present invention proposes a lithium battery SOH estimation method based on the proportion of constant-voltage charging time. Only the constant-voltage charging time and the total charging time are needed to estimate the SOH of the lithium battery, which greatly reduces the calculation amount of SOH estimation, is conducive to real-time application, and at the same time makes the SOH estimation accuracy not affected by voltage and current measurement errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The present invention provides an overall flow chart of a lithium battery SOH estimation method based on constant voltage charging time ratio. DETAILED DESCRIPTION

[0035] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0037] In a specific implementation case, Figure 1 As shown, a lithium battery SOH estimation method based on constant voltage charging time ratio is provided, and the specific steps are as follows:

[0038] S1: Perform charge and discharge cycle test on N lithium batteries. In each cycle, first perform constant current charging, then constant voltage charging until the lithium battery is fully charged. Then collect the total time T of the entire charging process of N lithium batteries and the constant voltage charging time T cv , then discharge at a constant current to the discharge cut-off voltage, and calculate the SOH of the lithium battery in the current cycle according to the discharge capacity;

[0039] S2: For all lithium batteries in step S1, set the constant voltage charging time T of each lithium battery per cycle. cv Divide it by the total time of the entire charging process T, and get the constant voltage charging time ratio T of each lithium battery in each cycle per , and use it as a feature, and use the SOH value of each lithium battery after each cycle obtained in step S1 as a label to form a feature data set T perh ;

[0040] S3: Feature data set T obtained in step S2 perh Randomly select 80% of the data as training set T train , and then use the features of the remaining lithium batteries as the test set T test ;

[0041] S4: The training set T in step S3 train Input into the SVR model for training and testing. If the performance of the model meets the requirements, save the model. If not, modify the model parameters until the model training and testing meet the accuracy requirements.

[0042] S5: When the lithium battery is actually running, extract the constant voltage charging time T of the lithium battery according to the method of step S1 during the constant voltage charging stage.cv The total time of the whole charging process is T, and then the constant voltage charging time is T cv Divide by the total time T of the entire charging process to get the constant voltage charging time ratio T of each lithium battery charging process per , and then the constant voltage charging time ratio T per The data is input into the SVR model described in step S4 to estimate the SOH of the lithium battery.

[0043] 2. A lithium battery SOH estimation method based on constant voltage charging time ratio according to claim 1, characterized in that the total time T of the entire charging process and the constant voltage charging time ratio T in step S1 are per The construction method is as follows:

[0044] S1-1: Perform charge and discharge cycle tests on N lithium batteries. In each cycle, first perform constant current charging until the lithium battery terminal voltage reaches the charging cut-off voltage, and then perform constant voltage charging until the lithium battery is fully charged. In each cycle, collect the total charging time and constant voltage charging time of the nth lithium battery in the ath cycle, and calculate the constant voltage charging time ratio and arrange them in a row. The specific form is as follows:

[0045] T=[T 11 ,T 12 ,...,T 1a ,T 21 ,T 22 ,...,T 2a ,...,T n1 ,T n2 ,...,T na ] T

[0046] T cv =[T cv11 ,T cv12 ,...,T cv1a ,T cv21 ,T cv22 ,...,T 2a ,...,T cvn1 ,T cvn2 ...,T cvna ] T

[0047] T per =[T per11 ,T per12 ,...,T per1a ,T per21 ,T per22 ,...T per2a ,...,T pern1 ,T pern2 ...,Tperna ] T

[0048] Among them, T na represents the total charging time of the nth lithium battery in the ath cycle; T cvna represents the constant voltage charging time in the ath cycle of the nth lithium battery; T perna It represents the proportion of the constant voltage charging time in the ath cycle of the nth lithium battery in the total charging time;

[0049] S1-2: In each discharge cycle, the SOH label value of the lithium battery is calculated based on the total discharge capacity. When the SOH of the lithium battery is lower than 70%, the cycle is stopped. The calculation formula of SOH is as follows:

[0050]

[0051] Among them, C now Represents the current cycle discharge capacity of the lithium battery, C initial Represents the initial capacity of the lithium battery.

[0052] Arrange the obtained SOH into a column to form a label sequence as follows:

[0053] SOH=[SOH 11 ,SOH 12 ,...,SOH 1a ,SOH 21 ,SOH 22 ,...SOH 2a ,...SOH n1 ,SOH n2 ...,SOH na ] T

[0054] Among them, SOH na represents the SOH in the ath cycle of the nth lithium battery;

[0055] S1-3: Construct a feature dataset T according to steps S1-1 and S1-2 perh as follows:

[0056]

[0057] 3. The method for estimating SOH of a lithium battery based on constant voltage charging time ratio according to claim 1 is characterized in that the data set division obtained in S3 and the training and estimation of the SVR model are specifically performed as follows:

[0058] S4-1: Feature dataset T obtained in S2 perh Randomly select 80% of the data as training set T train, and then use the features of the remaining lithium batteries as the test set T test ;

[0059] S4-2: Create a machine learning SVR, use rbf as the kernel function, set the penalty factor, epsilon and other parameters of the error term in the SVR, and perform training;

[0060] S4-3: Input the test set in step S3 into the trained model in S4-2 for estimation;

[0061] S4-4: Compare the estimated value of the lithium battery SOH output by the model in step S4-3 with the true value, and use the mean square error MSE and the mean absolute error MAE as the performance indicators of SVR. The calculation formulas of the mean square error MSE and the mean absolute error MAE are as follows:

[0062]

[0063] Among them, SOH real (a) represents the true value of SOH in the ath cycle, SOH pre (a) represents the estimated value of SOH in the ath cycle;

[0064] S4-5: If the mean square error MSE and the mean absolute error MAE calculated in step S4-4 are within an acceptable range, the SVR model meets the accuracy requirements, otherwise repeat step S4 to retrain the model.

Claims

1. A lithium battery SOH estimation method based on constant voltage charging time ratio, characterized in that: The specific steps are as follows: S1: Perform charge and discharge cycle test on N lithium batteries. In each cycle, first perform constant current charging, then constant voltage charging until the lithium battery is fully charged. Then collect the total time T of the entire charging process of N lithium batteries and the constant voltage charging time T cv , then discharge at a constant current to the discharge cut-off voltage, and calculate the SOH of the lithium battery in the current cycle according to the discharge capacity; S2: For all lithium batteries in step S1, set the constant voltage charging time T of each lithium battery per cycle. cv Divide it by the total time of the entire charging process T, and get the constant voltage charging time ratio T of each lithium battery in each cycle per , and use it as a feature, The SOH value of each lithium battery after each cycle obtained in step S1 is used as a label to form a feature data set T perh ; S3: Feature data set T obtained in step S2 perh Randomly select 80% of the data as training set T train , and then use the features of the remaining lithium batteries as the test set T test ; S4: The training set T in step S3 train Input into the SVR model for training and testing. If the performance of the model meets the requirements, save the model. If not, modify the model parameters until the model training and testing meet the accuracy requirements. S5: When the lithium battery is actually running, extract the constant voltage charging time T of the lithium battery according to the method of step S1 during the constant voltage charging stage. cv The total time of the whole charging process is T, and then the constant voltage charging time is T cv Divide by the total time T of the entire charging process to get the constant voltage charging time ratio T of each lithium battery charging process per , and then the constant voltage charging time ratio T per The data is input into the SVR model described in step S4 to estimate the SOH of the lithium battery.

2. A lithium battery SOH estimation method based on constant voltage charging time ratio according to claim 1, characterized in that: The total time T of the entire charging process described in step S1 and the constant voltage charging time ratio T per The construction method is as follows: S1-1: Perform charge and discharge cycle tests on N lithium batteries. In each cycle, first perform constant current charging until the lithium battery terminal voltage reaches the charging cut-off voltage, and then perform constant voltage charging until the lithium battery is fully charged. In each cycle, collect the total charging time and constant voltage charging time of the nth lithium battery in the ath cycle, and calculate the constant voltage charging time ratio and arrange them in a row. The specific form is as follows: T=[T 11 ,T 12 ,...,T 1a ,T 21 ,T 22 ,...,T 2a ,...,T n1 ,T n2 ,...,T na ] T T cv =[T cv11 ,T cv12 ,...,T cv1a ,T cv21 ,T cv22 ,...,T 2a ,...,T cvn1 ,T cvn2 ...,T cvna ] T T per =[T per11 ,T per12 ,...,T per1a ,T per21 ,T per22 ,...T per2a ,...,T pern1 ,T pern2 ...,T perna ] T Among them, T na represents the total charging time of the nth lithium battery in the ath cycle; T cvna represents the constant voltage charging time in the ath cycle of the nth lithium battery; T perna It represents the proportion of the constant voltage charging time in the ath cycle of the nth lithium battery in the total charging time; S1-2: In each discharge cycle, the SOH label value of the lithium battery is calculated based on the total discharge capacity. When the SOH of the lithium battery is lower than 70%, the cycle is stopped. The calculation formula of SOH is as follows: Among them, C now Represents the current cycle discharge capacity of the lithium battery, C initial Represents the initial capacity of the lithium battery. Arrange the obtained SOH into a column to form a label sequence as follows: SOH=[SOH 11 ,SOH 12 ,...,SOH 1a ,SOH 21 ,SOH 22 ,...SOH 2a ,...SOH n1 ,SOH n2 ...,SOH na ] T Among them, SOH na represents the SOH in the ath cycle of the nth lithium battery; S1-3: Construct a feature dataset T according to steps S1-1 and S1-2 perh as follows:

3. A lithium battery SOH estimation method based on constant voltage charging time ratio according to claim 1, characterized in that: The specific steps for data set partitioning obtained by S3 and training and estimating the SVR model are as follows: S4-1: Feature dataset T obtained in S2 perh Randomly select 80% of the data as training set T train , and then use the features of the remaining lithium batteries as the test set T test ; S4-2: Create a machine learning SVR, use rbf as the kernel function, set the penalty factor, epsilon and other parameters of the error term in the SVR, and perform training; S4-3: Input the test set in step S3 into the trained model in S4-2 for estimation; S4-4: Compare the estimated value of the lithium battery SOH output by the model in step S4-3 with the true value, and use the mean square error MSE and the mean absolute error MAE as the performance indicators of SVR. The calculation formulas of the mean square error MSE and the mean absolute error MAE are as follows: Among them, SOH real (a) represents the true value of SOH in the ath cycle, SOH pre (a) represents the estimated value of SOH in the ath cycle; S4-5: If the mean square error MSE and the mean absolute error MAE calculated in step S4-4 are within an acceptable range, the SVR model meets the accuracy requirements, otherwise repeat step S4 to retrain the model.

Citation Information

Cited By

  • Lithium battery health state prediction method based on CPO and SVR

    CN121186638A