CN115219902B - A method and system for quickly testing the life of a power battery - Google Patents

Abstract

The present invention belongs to the field of power batteries, and discloses a method and system for rapid testing of power battery life, including: obtaining a variable curve of a power battery to be tested, and obtaining a horizontal and vertical variable sequence based on the obtained variable curve; using the horizontal and vertical variable sequence as independent variables, fitting the independent variables to the battery cycle life based on the kernel canonical correlation analysis method, and introducing a recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variable and the battery cycle life; combining the sequence corresponding to the battery cycle life with a maximum correlation coefficient higher than a threshold with a machine learning model to obtain a power battery life test result. The present invention avoids the requirement for prior knowledge, and compared with traditional testing methods, it shows good effectiveness and improved robustness in battery cycle life testing, and greatly shortens the testing time.

Description

A method and system for quickly testing the life of a power battery

Technical Field

The present invention belongs to the field of power batteries, and in particular relates to a method and system for quickly testing the life of a power battery.

Background Art

Battery life refers to the number of charge and discharge cycles until the available capacity drops to 80% of the initial value. Cycle life testing of batteries can further understand battery characteristics, prove whether the battery meets the design goals, achieve better control during battery use, and evaluate the battery's ability to meet the needs of different application scenarios.

At present, the cycle life test of lithium-ion batteries mainly adopts the national standard GB/T 31484-2015 "Requirements and Test Methods for Cycle Life of Power Storage Batteries for Electric Vehicles". The standard stipulates that when conducting a standard battery cycle life test, the battery needs to be continuously charged and discharged until its life end condition, that is, "test to the end". The number of charge and discharge cycles of lithium-ion batteries can reach thousands or even more than 10,000 times, and the test time can reach 1 year or even longer. This test is usually very time-consuming. The prior art discloses a battery testing method: controlling the battery to perform N rounds of charge and discharge operations, each round of charge and discharge operations includes multiple charging operations and multiple discharging operations, multiple charging operations are performed based on at least two current specifications, and multiple discharging operations are performed based on at least two current specifications; after controlling the battery to perform N rounds of charge and discharge operations, the number of tests is recorded plus 1; re-execute the step of determining the battery capacity value of the battery until the battery capacity value is less than the preset capacity value. The prior art discloses a battery testing method: obtaining a current charge and discharge curve, dividing the voltage interval based on the upper and lower limits of the battery voltage, and performing cycle tests on the same battery sample in parallel in different voltage intervals to obtain the cycle attenuation of the different voltage intervals; and fusing the cycle attenuation of the different voltage intervals to obtain the total cycle attenuation. However, the above method requires full-cycle, long-term, and high-cost testing to obtain the battery cycle life.

The prediction-based rapid battery life test method is an excellent solution to the above problems, that is, "estimate instead of test". By replacing the full cycle test with prediction, a lot of time can be saved, and the efficient and accurate evaluation of the battery cycle life can be achieved. The prior art discloses a prediction method, which performs a short-term cycle performance test on the battery with different cycle times, records the cycle times and capacity retention rate, and then disassembles the battery after different cycle times, uses X-ray diffraction to test the graphitization degree of the graphite negative electrode material, and tests according to the three data of cycle times, capacity retention rate and graphitization degree. This method requires disassembly and destruction of the battery, and has certain limitations. The prior art discloses a test method for the cycle life of a lithium-ion battery, installs a pressure sensor on the surface of the battery, records the capacity information of the battery within a certain number of cycles, and fits the cycle times, discharge capacity and voltage data in the pressure sensor to achieve the test of the battery life. This method requires an additional pressure sensor, which will increase the cost in practice and has certain limitations. The prior art also discloses a rapid test method for the cycle life of a lithium-ion battery, which performs 500 cycle tests on the battery at different rates, and obtains a test equation to achieve the battery life test by fitting the equation. This method does not require sophisticated testing equipment and complex theoretical calculations, but it still requires a long cycle test, so it is not very practical. Battery life test methods are helpful for battery production, optimization and development, but power batteries are a strongly nonlinear system that is highly sensitive to the environment, and cycle life testing is extremely difficult. The new energy vehicle industry requires that power batteries be able to quickly iterate technology and upgrade products, but the lack of rapid cycle life test methods for existing batteries seriously restricts the rapid and high-quality development of the new energy vehicle industry. Therefore, how to shorten the life test time and quickly, accurately and efficiently test and evaluate the battery cycle life has become an important means to break through the key technical bottlenecks in the rapid development of power batteries and related industries.

There are many model testing methods for lithium battery cycle life testing, which can generally be divided into mechanism models, empirical models and machine learning models. Many mechanism models currently used for power battery life testing are based on pseudo-two-dimensional electrochemical models, which take into account capacity degradation caused by side reactions, but they cannot take into account battery inconsistencies. Unlike mechanism models, empirical models directly fit the capacity decline trajectory and ignore the internal mechanism of the battery; in order to update model parameters according to the latest battery status and working conditions, Kalman filters (KF) and local filters (PF) are widely used; however, since these empirical models fit the collected aging curves, they can usually only be applied under similar conditions, which means that the generalization ability is poor. Machine learning models are tested by mapping a set of extracted battery features to the battery cycle life based on a large amount of data. However, most machine learning models usually have fewer input features, which may lead to weak robustness in practical applications; in addition, these features are usually concentrated on a few points in the entire data set, and the rest of the data set is not fully utilized, resulting in reduced accuracy; in addition, manually finding a suitable data feature set usually takes a lot of effort.

Summary of the invention

In order to solve the technical problems of long cycle and high cost of traditional testing methods, the present invention proposes a rapid testing method and system for power battery life, which obtains horizontal and vertical variable sequences based on the acquired variable curves; takes the horizontal and vertical variable sequences as independent variables, fits the independent variables to the battery cycle life based on the kernel canonical correlation analysis method, introduces the recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variable and the battery cycle life and the corresponding kernel correlation coefficient analysis solution, which shows good effectiveness in battery cycle life testing and greatly shortens the test time.

In order to achieve the above object, the present invention adopts the following technical solution:

The first aspect of the present invention provides a method for quickly testing the life of a power battery, which adopts a method of estimating instead of testing, and comprises the following steps:

Obtaining a variable curve of the power battery to be tested;

Based on the obtained variable curves, horizontal and vertical variable sequences are obtained;

The horizontal and vertical variable sequences are used as independent variables, and the independent variables are fitted to the battery cycle life based on the kernel canonical correlation analysis method. The recursive kernel canonical correlation analysis method is introduced to obtain the maximum correlation coefficient between the independent variable and the battery cycle life.

The sequence corresponding to the battery cycle life with the maximum correlation coefficient higher than the threshold is tested with the machine learning model to obtain the power battery life test result.

A second aspect of the present invention provides a power battery life rapid testing system, comprising:

A data acquisition module, used to acquire a variable curve of the power battery to be tested;

A sequence building module is used to obtain horizontal and vertical variable sequences based on the obtained variable curves;

A data fitting module is used to take the horizontal and vertical variable sequences as independent variables, fit the independent variables to the battery cycle life based on the kernel canonical correlation analysis method, and introduce the recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variable and the battery cycle life;

The battery life test module combines the sequence corresponding to the battery cycle life with the maximum correlation coefficient higher than the threshold with the machine learning model to obtain the power battery life test result.

A third aspect of the present invention provides a computer-readable storage medium.

A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in the method for quickly testing the life of a power battery as described above.

A fourth aspect of the present invention provides a computer device.

A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for rapid testing the life of a power battery as described above are implemented.

Compared with the prior art, the present invention has the following beneficial effects:

(1) Compared with traditional testing methods, the present invention adopts a method of replacing testing with estimation, which shows good effectiveness in battery cycle life testing and greatly shortens the testing time. The life test can be completed by relying on accurate prediction, which greatly reduces the life test time and accelerates battery upgrading.

(2) The present invention tests the cycle life by comparing the states of different batteries in the same cycle and predicts the cycle life by the changing trends of different batteries at the same point. The kernel canonical correlation analysis method is introduced to solve the problem of overfitting by fitting the variable sequence to an almost perfect cycle life. The method can be combined with traditional features and shows good effectiveness and improved robustness in battery cycle life testing compared with traditional testing methods.

(3) The present invention fits the independent variable to the battery cycle life based on the kernel canonical correlation analysis method, introduces the kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variable and the battery cycle life, and proposes a recursive canonical correlation analysis solution algorithm to improve the solution operation speed; the sequence corresponding to the battery cycle life with a maximum correlation coefficient higher than the threshold is combined with the machine learning model to obtain the power battery life test result, avoiding the requirement for prior knowledge.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

FIG1 is a schematic diagram of the overall process of a method for rapidly testing the life of a power battery according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of the transverse and longitudinal variable sequences of a battery according to an embodiment of the present invention.

FIG. 3 is a test result obtained using the first 50 cycle data according to an embodiment of the present invention.

FIG. 4 is a test result obtained using the first 100 cycle data according to an embodiment of the present invention.

DETAILED DESCRIPTION

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and/or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and/or combinations thereof.

The overall idea of the present invention is as follows: First, a horizontal sequence set and a vertical sequence set are derived from each selected battery variable, and these sequences are fitted to the battery cycle life through kernel canonical correlation analysis. Recursive kernel canonical correlation analysis is used to reduce the computational load; in addition, a recursive canonical correlation analysis solution algorithm is proposed to improve the solution operation speed. Through the above analysis method, appropriate fitting results are generated as optional inputs to the machine learning model. Finally, a threshold of the correlation coefficient is specified to select high-quality machine learning algorithm inputs.

Embodiment 1

As shown in FIG1 , this embodiment provides a method for quickly testing the life of a power battery, which uses estimation instead of a method to be tested, and includes the following steps:

S1: Obtaining a variable curve of the power battery to be tested;

As one or more embodiments, in S1, this embodiment selects a data set collected by the Massachusetts Institute of Technology (MIT).

It should be noted that the number of batteries and the variable curves of the batteries in the data set selected in this embodiment can be set according to those skilled in the art.

For example, in this embodiment, a total of 123 batteries and 7 battery variable curves are selected;

The variable curves of the power battery to be tested include: discharge capacity-voltage curve (Q(V)), discharge temperature-voltage curve (T(V)), voltage increment-capacity (IC), discharge voltage-time curve (V(t)), discharge temperature-time curve (T(t)), discharge voltage/current-time curve (V/I(t)) and the charging current interpolation curve (ICC) during the entire CV (abbreviation for constant voltage, i.e. the constant voltage stage of charging) step.

Among them, ICC is standardized using linear interpolation.

S2: Obtain horizontal and vertical variable sequences based on the obtained variable curves;

As shown in FIG2 , the sequence of battery variables that change within one cycle in the variable curve is taken as the horizontal variable sequence, and the sequence of battery variables that change at the same sampling point in different cycles is taken as the vertical variable sequence.

The curves of different horizontal variable series are more obvious in the later cycles, so they usually perform better.

In contrast, in early cycles, where differences between cells are small and irregular, the performance of longitudinal series is related to the location of the sampling points, generally performing better where the variables vary more widely and more regularly.

For example, based on the data obtained by S1, a set of horizontal and vertical variable sequences is derived for each variable, that is, a set of horizontal and vertical variable sequences is derived from each selected variable.

In this way, 14 variable sequences can be generated as independent variables for subsequent input of the recursive kernel canonical correlation analysis method.

S3: Taking the horizontal and vertical variable sequences as independent variables, fitting the independent variables to the battery cycle life based on the kernel canonical correlation analysis method, and introducing the recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variables and the battery cycle life;

Kernel Canonical Correlation Analysis (KCCA) can test the cycle life of batteries by comparing the states of different batteries in the same cycle through a horizontal variable sequence, and can test the cycle life of batteries by comparing the change trends of different batteries at the same point through a vertical variable sequence.

Since KCCA can fit the variable sequence to almost perfect cycle life in the training step, the problem of overfitting must be addressed. The hyperparameters for which KCCA performs well on the training set may not be suitable for the test set.

For example, if a small η is simply chosen in the training step to make the correlation coefficient larger, the fitting results may not be satisfactory in the test set.

To address this issue, this example creates a KCCA validation group to test and select KCCA hyperparameters. Considering the differences in calendar aging and rest time between batches, approximately the same number of cells in each batch are allocated to the training, validation, and test groups.

As shown in Figure 2, the dataset is divided into three groups, KCCA training group, KCCA validation group and KCCA test group, each with 41 batteries. The KCCA training group and the KCCA validation group both constitute the machine learning training group.

The KCCA training group trains the variable sequence using different hyperparameter sets, which are effectively identified by the grid search method.

This embodiment proposes a recursive KCCA (RKCCA) solution algorithm based on kernel canonical correlation analysis (KCCA) to facilitate calculation.

Traditional canonical correlation analysis (CCA) aims to find the linear relationship between two sets of multidimensional variables.

Two sets of random variables, of the form (x, y), The mean is zero. Suppose there is a sample (x, y) of a power battery data instance S = ((x ₁ , y ₁ ), …, (x _n , _yn )), where x represents a sequence composed of horizontal and vertical variable sequences of the power battery, that is, the independent variable, and y represents the life of the corresponding power battery, that is, the dependent variable.

Here, X is represented by (x ₁ , …, x _n ) ^T , and Y is represented by (y ₁ , …, _yn ) ^T .

Assume that n ≥ rank(X), rank(Y), and that both X and Y have zero mean. Define a new vector w, named the loading vector of x and project x into that direction, and do the same for y by choosing a loading vector v to obtain the projection of y.

Define the projections t=Xw and u=Yv as the score vectors of X and Y, respectively. CCA determines w and v to maximize the correlation coefficient between t=Xw and u=Yv.

In this embodiment, all ||·|| represent Euclidean norms.

The method of using the horizontal and vertical variable sequences as independent variables and fitting the independent variables to the battery cycle life based on the kernel canonical correlation analysis method specifically includes:

S301: Using kernel function to map independent variables to high-dimensional space includes:

Map each _xi into a high-dimensional vector space F with infinite dimensions:

Wherein, _xi represents each element in the independent variable sequence, and Φ( _xi ) represents the result of mapping each element in the independent variable sequence to a high-dimensional space;

Φ(X)=(Φ(x ₁ ), Φ(x ₂ ),…, Φ(x _n )) ^T (2)

Among them, the mean of Φ(X) should also be zero:

Avoid solving the specific Φ( _xi ) by applying the kernel trick:

K(X) _i,j =Φ(x _i )Φ(x _j ) ^T =k(x _i ,x _j ) (4)

K(X)＝Φ(X)Φ(X) ^T (5)

Among them, the kernel function k( _xi , _xj ) is a symmetric function. Therefore, K(X) is a symmetric matrix, which represents the independent variable kernel matrix.

S302: Centralization

Usually k( _xi , _xj ) cannot ensure that the mean of Φ(X) is zero, so centering is required:

S303: Calculate the projection of the sequence and battery cycle life corresponding to the independent variable mapped to the high-dimensional space; and construct an optimization function with the goal of maximizing the correlation coefficient between the above projections.

It should be noted that, for convenience, in the rest of this embodiment, Φ and K are used to replace Φ(X) and K(X) respectively.

The optimization problem becomes maximizing the correlation coefficient ρ ₁ :

Among them, Φ and Φ(X) have the same meaning, representing the independent variable matrix mapped to the high-dimensional space, Y represents the dependent variable matrix, w ₁ represents the projection vector of the independent variable matrix mapped to the high-dimensional space, and v ₁ represents the projection vector of the dependent variable matrix.

Φ is usually full rank, so _wi is always a linear combination of Φ( _xi ):

w _i =Φ ^T α _i (8)

Similar to CCA, we can get:

(K ^T K) ^-1 K ^T Y(Y ^T Y) ^-1 Y ^T Kα ₁ =ρ ₁ ² α ₁ (9)

(Y ^T Y) ^-1 Y ^T K(K ^T K) ^-1 K ^T Yv ₁ =ρ ₁ ² v ₁ (10)

However, K is symmetric, which leads to:

(K ^T K) ^-1 K ^T Y(Y ^T Y) ^-1 Y ^T K＝I (11)

Where I is the identity matrix, and all eigenvalues are equal to 1. When α _i is arbitrarily selected, |ρ _i ||＝1 always holds, and a perfect correlation can always be formed.

Regularization is required in kernel methods, and the term ηα ₁ ^T Kα ₁ is added to the optimization equation:

where η is a manually chosen hyperparameter, usually small, and ^{α T} Kα is a term in the optimization problem of the kernel partial least squares (KPLS) method.

The regularization problem can be viewed as a transitional form of non-regularized KCCA and KPLS.

After this step, it is guaranteed that the equation is solvable and that the matrix eigenvalues are not always 1:

(K ^T K+η ₁ I) ^-1 K ^T Y(Y ^T Y) ^-1 Y ^T Kα ₁ =λ ₁ ² α ₁ : (13)

(Y ^T Y) ^-1 Y ^T K(K ^T K+η ₁ I) ^-1 K ^T Yv ₁ =λ ₁ ² v ₁ (14)

It is important to emphasize that λ ₁ ² is not exactly the same as ρ ₁ ² because the restrictions have changed, but their differences should be small.

The maximum correlation coefficient between the independent variable and the battery cycle life obtained by introducing the recursive kernel canonical correlation analysis method includes:

Based on the power method, based on a given diagonalizable matrix and a random non-zero vector, an iterative solution method is used to find the maximum eigenvalue of the diagonalizable matrix, and the maximum correlation coefficient corresponding to the maximum eigenvalue is obtained according to the relationship between the eigenvalue and the correlation coefficient.

Similar to CCA, α and v are the corresponding eigenvectors of the solution (K ^T K+ηI) ^-1 K ^T Y(Y ^T Y) ^-1 Y ^T Kα ₁ and (Y ^T Y) ^-1 Y ^T K(K ^T K+η ₁ I) ^-1 K ^T Yv ₁ .

Since we only need to find the largest eigenvalue, using the power method is a better approach to accomplish this task.

Given a diagonalizable matrix D and a random non-zero vector b ₀ , when k→∞, the eigenvector b _k corresponding to the maximum eigenvalue of D can be obtained by the following recursive relation:

The relationship between α and v can be obtained:

v∝(YY ^T ) ^-1 YK ^T α (16)

α∝(KK ^T +ηI) ^-1 KY ^T v (17)

In this embodiment, according to the properties of canonical correlation analysis CCA, the relationship between the eigenvalue and the correlation coefficient is: the square of the eigenvalue is the corresponding correlation coefficient.

In view of this feature, this embodiment uses an iterative solution method to find the maximum eigenvalue of the diagonalizable matrix based on a given diagonalizable matrix and a random non-zero vector, including:

(1) Select a random non-zero vector as the initial value of α, usually the first column of K (alpha normalize the weight vector of the kernel matrix);

Then perform the following iterations:

(2) The score vector of Φ: t = Kα, (The score vector t of the independent variable matrix mapped to the high-dimensional space at convergence is the product of the independent variable kernel matrix K and the weight vector alpha, and t is standardized)

(3) Y load vector: v = (Y ^T Y) ^-1 Y ^T t, (The relationship between the load vector v of the dependent variable and the dependent variable matrix Y and the score vector t of the independent variable matrix mapped to the high-dimensional space at convergence, and v is standardized)

(4) Score vector of Y: u = Yv; (When converged, the score vector u of the dependent variable is the product of the dependent variable matrix Y and the dependent variable loading vector v)

(5) Weight vector: α = (KK ^T + ηI) ^-1 KY ^T v, (The relationship between the weight vector alpha of the kernel matrix and the kernel matrix, the dependent variable matrix Y, and the load vector v of the dependent variable at convergence, and the weight vector alpha of the kernel matrix is standardized)

In the above manner, this embodiment can find the maximum eigenvalue and the corresponding eigenvector in a more convenient way at the cost of a slight loss of accuracy, rather than performing a complete eigenvector decomposition on the two matrices.

In the KCCA testing algorithm, x is considered as an independent variable and y as a dependent variable. The goal is to test Y _new with a new given independent variable X _new based on the previously captured relationship. The kernel matrix K _new of X _new is preprocessed as follows:

K _new ＝Φ(X _new )Φ(X) ^T (18)

Then centralize K _new :

The test Y _new is performed by assuming that the score matrices of X _new and Y _new are the same. The score matrix of X _new is:

T _new = K _new A (20)

Think of T _new as the score matrix of Y _new :

The estimated Y _new can be obtained:

S4: The sequence corresponding to the battery cycle life with a maximum correlation coefficient higher than the threshold is tested in combination with the machine learning model to obtain the power battery life test result.

This embodiment takes into account that the quality of the data input to the test model is more important than the quantity, and the performance of the deep learning method is better than the traditional machine learning method (GPR and Elastic net).

In this embodiment, four representative machine learning models are used to test the sequences corresponding to the fitting results with correlation coefficients higher than the threshold, and the machine learning models include artificial neural network (ANN), random forest (RF), Gaussian process regression (GPR) and elastic net.

Using the data from the first 100 cycles, the best mean absolute percentage errors (MAPE) for ANN, RF, GPR, and Elastic net were 8.4%, 8.6%, 11.1%, and 14.0%, respectively. Using the data from the first 50 cycles, the best results ranged from 9.5% to 14.9%, with RF achieving the best performance. The above experiments verified that the performance of ANN and RF is better than that of traditional machine learning algorithms (GPR and Elastic network).

Regarding threshold setting, a higher correlation coefficient threshold is not always better, especially for deep learning methods. A lower threshold means that more fitting results can be input into the machine learning model, which is also helpful for testing.

Therefore, the threshold is set according to the actual selected machine learning model.

For example, generally for ANN, as the threshold decreases, MAPE decreases first and then increases, with the lowest point occurring when the threshold is set to 0.8, where there is a balance between the quantity and quality of these inputs. For GPR and Elastic net, their best performance occurs when the threshold is high.

In order to verify the effectiveness of the method in this embodiment, the following experiments were conducted:

Table 1 shows the experimental results based on the MIT data set, FIG. 3 shows the test results obtained by the embodiment of the present invention using the first 50 cycle data, and FIG. 4 shows the test results obtained by the embodiment of the present invention using the first 100 cycle data.

The experimental results are evaluated by introducing MAPE and RMSE, where MAPE is the mean absolute percentage error and RMSE is the root mean square error.

The final experimental results show that compared with the traditional full life cycle test which requires 2,000 times, the present invention can complete the life test at the expense of a small amount of accuracy by relying on precise testing. It only takes 50 to 300 cycles to complete the estimation, without the need for comprehensive testing, shortening the test cycles by 1,950 and increasing the speed by 40 times, which can greatly reduce the life test time and accelerate battery upgrades.

Table 1 Experimental results

Number of cycles MAPE(%) RMSE(loop) Shorten the number of cycles 50(2.5%) 9.1 122 1950 100(5%) 8.2 98 1900 200(10%) 8.0 93 1800 300(15%) 7.7 89 1700

Embodiment 2

This embodiment provides a power battery life rapid testing system, including:

A data acquisition module, used to acquire a variable curve of the power battery to be tested;

A sequence building module is used to obtain horizontal and vertical variable sequences based on the obtained variable curves;

A data fitting module is used to take the horizontal and vertical variable sequences as independent variables, fit the independent variables to the battery cycle life based on the kernel canonical correlation analysis method, and introduce the recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variable and the battery cycle life;

The battery life test module combines the sequence corresponding to the battery cycle life with the maximum correlation coefficient higher than the threshold with the machine learning model to obtain the power battery life test result.

Embodiment 3

This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps in the method for quickly testing the life of a power battery as described above are implemented.

Embodiment 4

This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the method for rapid testing the life of a power battery as described above are implemented.

Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program codes.

The present invention is described with reference to the flowchart and/or block diagram of the method, device (system), and computer program product according to the embodiment of the present invention. It should be understood that each process and/or box in the flowchart and/or block diagram, as well as the combination of the process and/or box in the flowchart and/or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims (9)

1. The rapid power battery life test method is characterized by adopting an estimated replacement test method and comprises the following steps:

Acquiring a variable curve of a power battery to be tested;

the variable curve of the power battery to be tested comprises a discharge capacity-voltage curve, a discharge temperature-voltage curve, a voltage increment-capacity curve, a discharge voltage-time curve, a discharge temperature-time curve, a discharge voltage/current-time curve and a charge current interpolation curve during the constant voltage stage step of the whole charge;

obtaining a transverse variable sequence and a longitudinal variable sequence based on the obtained variable curve;

The transverse and longitudinal variable sequences obtained based on the obtained variable curves are as follows:

taking a sequence formed by battery variables changing in one period in a variable curve as a transverse variable sequence, and taking a sequence formed by battery variables changing in the same sampling point in different periods as a longitudinal variable sequence;

Using the transverse and longitudinal variable sequences as independent variables, fitting the independent variables to the cycle life of the battery based on a kernel canonical correlation analysis method, and introducing a recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variables and the cycle life of the battery;

And combining the sequence corresponding to the battery cycle life with the maximum correlation coefficient higher than the threshold value with a machine learning model to test so as to obtain a power battery life test result.

2. The rapid power cell life test method of claim 1, wherein said fitting the independent variables to the battery cycle life based on a kernel-based canonical correlation analysis method using the sequence of lateral and longitudinal variables as the independent variables comprises:

Mapping the sequence corresponding to the independent variable to a high-dimensional space by adopting a kernel function solution;

Calculating projections of the sequence and the battery cycle life corresponding to the high-dimensional space independent variable;

And constructing an optimization function by maximizing the correlation coefficient between the projections as a target.

3. A rapid power battery life test method according to claim 2, wherein the term in the optimization problem of the hyper-parametric and the kernel-biased least squares method is added to the objective function by regularization when the optimization function is constructed.

4. The rapid power battery life test method of claim 1, wherein said introducing a recursive kernel canonical correlation analysis method to obtain a maximum correlation coefficient between the independent variable and the battery cycle life comprises: based on the exponentiation method, based on a given diagonable matrix and a random non-zero vector, the maximum eigenvalue of the diagonable matrix is found by adopting an iterative solution method, and the maximum correlation coefficient corresponding to the maximum eigenvalue is obtained according to the relation between the eigenvalue and the correlation coefficient.

5. A rapid power battery life test system, characterized in that it is based on the rapid power battery life test method according to any one of claims 1 to 4, further comprising:

The data acquisition module is used for acquiring a variable curve of the power battery to be tested;

The sequence construction module is used for obtaining transverse and longitudinal variable sequences based on the obtained variable curves;

The transverse and longitudinal variable sequences obtained based on the obtained variable curves are as follows:

taking a sequence formed by battery variables changing in one period in a variable curve as a transverse variable sequence, and taking a sequence formed by battery variables changing in the same sampling point in different periods as a longitudinal variable sequence;

the data fitting module is used for taking the transverse and longitudinal variable sequences as independent variables, fitting the independent variables to the cycle life of the battery based on a kernel canonical correlation analysis method, and introducing a recursive kernel canonical correlation analysis method to obtain the maximum correlation coefficient between the independent variables and the cycle life of the battery;

And the battery life test module is used for testing the sequence corresponding to the battery cycle life with the maximum correlation coefficient higher than the threshold value by combining the machine learning model to obtain a power battery life test result.

6. The rapid power cell life test system of claim 5 wherein said fitting the independent variables to the battery cycle life based on a kernel-based canonical correlation analysis method using the transverse and longitudinal variable sequences as independent variables comprises:

Mapping the sequence corresponding to the independent variable to a high-dimensional space by adopting a kernel function solution;

Calculating projections of the sequence and the battery cycle life corresponding to the high-dimensional space independent variable;

And constructing an optimization function by maximizing the correlation coefficient between the projections as a target.

7. The rapid power battery life test system of claim 5 wherein said introducing a recursive kernel canonical correlation analysis method results in a maximum correlation coefficient between the independent variable and the battery cycle life comprising: based on the exponentiation method, based on a given diagonable matrix and a random non-zero vector, the maximum eigenvalue of the diagonable matrix is found by adopting an iterative solution method, and the maximum correlation coefficient corresponding to the maximum eigenvalue is obtained according to the relation between the eigenvalue and the correlation coefficient.

8. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of a power battery life fast test method according to any one of claims 1-4.

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps of a method for rapid testing of the life of a power battery as claimed in any one of claims 1 to 4.