A lithium-ion battery consistency identification method based on voltage curve shape

Abstract

The present invention relates to a method for identifying lithium-ion battery consistency based on voltage curve shape, comprising the following steps: obtaining voltage information for each battery; obtaining a shape dissimilarity matrix based on the derivative dynamic time warping distance based on the voltage information; obtaining a non-metric multidimensional scaling result based on the dissimilarity matrix; and implementing multi-parameter consistency visualization based on the non-metric multidimensional scaling fitting composition based on the non-metric multidimensional scaling result. Compared with existing technologies, the present invention can directly utilize voltage data to achieve multi-parameter consistency identification for lithium-ion batteries and is applicable to various operating conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of battery packs, and in particular to a lithium-ion battery consistency identification method based on voltage curve shape. Background Art

[0002] Lithium-ion batteries, with their high energy density, long cycle life, and low self-discharge, have become widely used in new energy vehicles, smart grids, and other fields. However, due to fluctuations in the operating environment during battery manufacturing, parameter inconsistencies are common among lithium-ion batteries. These inconsistencies can amplify as the battery system operates, leading to decreased system performance, shortened service life, and even safety issues. Therefore, developing fast and reliable battery consistency identification methods is crucial for improving the performance and safety of battery packs and extending their service life.

[0003] External battery parameters usually refer to voltage, current, and temperature; internal parameters mainly refer to capacity, internal resistance, state of charge, coulombic efficiency, etc., among which capacity, internal resistance, and state of charge are of greatest concern. Existing feature-based methods mainly use the following two methods for battery consistency identification: one is to use voltage as a consistency feature parameter and directly make battery consistency judgments based on differences in voltage values; the other is to first convert voltage operating data into differential capacity curves, capacity increment curves, and other forms, and then extract characteristic parameters such as peak height and peak width to make capacity and internal resistance consistency judgments. The former focuses on differences in external battery characteristics and cannot deeply explore the source of internal battery consistency; the latter relies on characteristic constant-rate operating conditions, which makes it difficult to meet operational requirements and can only distinguish between changes in capacity and internal resistance caused by aging. Therefore, how to use external voltage operating data to identify the consistency of key internal parameters such as capacity, internal resistance, and state of charge requires the development of new methods.

[0004] Chinese Invention Publication No. CN201710308848.4 discloses a consistency evaluation method for series-structured battery packs. Based on the measured voltage curves of individual cells, a distance similarity index calculation method based on a dynamic time warping strategy was developed. This method can adaptively eliminate factors influencing differences in cell voltage due to different SOCs, enabling estimation of differences in operating characteristics between cells during the charge and discharge process. Compared to existing consistency extraction methods based on voltage differences at discrete moments, this method is more helpful in extracting intrinsic battery information. Furthermore, a statistical identification method for abnormal cells based on a stepwise iteration strategy was developed, enabling consistency evaluation of series-structured battery packs. However, this method uses a dynamic time warping strategy to represent similarity, primarily based on voltage differences. It directly evaluates consistency overall based on a similarity index matrix and eliminates the influence of the state of charge (SOC). This method is unable to distinguish between individual differences in SOC, capacity, or internal resistance.

[0005] In summary, existing battery identification methods have the following shortcomings:

[0006] (1) Most existing methods evaluate abnormal batteries based on the overall consistency of the battery. The object of evaluation is the overall battery performance, and it is impossible to distinguish the differences between the various battery parameters;

[0007] (2) Some existing methods use similarity indicators to sort batteries and use the threshold of abnormal indicators to achieve the purpose of abnormal battery identification. However, batteries generally have consistency problems, and the existence of consistency problems does not mean that they are abnormal batteries, and the evaluation of batteries is inaccurate. Summary of the Invention

[0008] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a lithium-ion battery consistency identification method based on the voltage curve shape that can distinguish the consistency of single parameters such as state of charge, capacity, and internal resistance.

[0009] The purpose of the present invention can be achieved by the following technical solutions:

[0010] One aspect of the present invention provides a method for identifying the consistency of lithium-ion batteries based on the shape of a voltage curve, comprising the following steps: obtaining voltage information of each battery; obtaining a shape dissimilarity matrix based on the derivative dynamic time warping distance based on the voltage information; obtaining a non-metric multidimensional scaling result based on the dissimilarity matrix; and realizing multi-parameter consistency visual identification based on the non-metric multidimensional scaling fitting composition based on the non-metric multidimensional scaling result.

[0011] As a preferred technical solution, the step of acquiring the voltage information of each battery includes: measuring the voltage of each single battery during the charge and discharge process, and obtaining the online voltage of each battery in multiple measurement stages.

[0012] As a preferred technical solution, the process of obtaining the dissimilarity matrix includes the following steps: obtaining a voltage derivative matrix based on the voltage information; and obtaining the dissimilarity matrix based on the voltage derivative matrix.

[0013] As a preferred technical solution, the process of obtaining the voltage derivative matrix includes the following steps: obtaining the voltage derivative matrix by calculating the derivative value of each voltage point based on the online voltage measured by each battery in multiple measurement stages.

[0014] As a preferred technical solution, the process of obtaining the dissimilarity matrix includes the following steps: adopting a dynamic regularization algorithm to obtain the dissimilarity matrix according to the voltage derivative matrix.

[0015] As an optimal technical solution, the process of obtaining the non-metric multidimensional scaling results includes the following steps: using the non-metric multidimensional scaling method to obtain the coordinate information of each battery in the p-dimensional space according to the non-similarity matrix; normalizing the coordinate information to obtain the non-metric multidimensional scaling results.

[0016] As an optimal technical solution, the process of obtaining the coordinate information of the p-dimensional space includes the following steps: obtaining the stress coefficient for non-metric multidimensional scaling based on the non-similarity matrix; using the direct gradient descent method to solve the minimization problem of the stress coefficient for non-metric multidimensional scaling to obtain the coordinate information of the p-dimensional space.

[0017] As a preferred technical solution, the implementation process of the multi-parameter consistency visual recognition based on the non-metric multidimensional scaling fitting composition includes the following steps: representing the non-metric multidimensional scaling results in a two-dimensional space fitting composition, clustering single cells with the same internal resistance, and forming equal SOC lines and equal capacity lines within each cluster; according to the positional relationship of each point in the p-dimensional space fitting composition, the farther the distance between the two points, the greater the difference in the corresponding single cells, thereby realizing multi-parameter consistency visual recognition based on the non-metric multidimensional scaling fitting composition.

[0018] Another aspect of the present invention provides an electronic device comprising: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the above-mentioned lithium-ion battery consistency identification method based on voltage curve shape.

[0019] Another aspect of the present invention provides a computer-readable storage medium comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the above-mentioned lithium-ion battery consistency identification method based on voltage curve shape.

[0020] Compared with the prior art, the present invention has the following advantages:

[0021] (1) Since the influence of different consistency parameters on voltage is not only reflected in the magnitude but also in the shape of the voltage curve, the present invention decouples the differences in different consistency parameters from the shape of the voltage curve and adopts a derivative dynamic time warping strategy to represent similarity, thereby achieving the distinction of the consistency of the single parameters of state of charge, capacity, and internal resistance;

[0022] (2) It can directly measure voltage data online and visualize the inconsistencies between individual cells, significantly reducing test time, battery modeling costs, and dependence on characteristic operating conditions. It has a wide range of applications and is highly user-friendly.

[0023] (3) The non-metric multidimensional scaling algorithm can effectively deal with the problem of missing data and is applicable to various non-similarity measurement indicators and is universal. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 Flowchart of a lithium-ion battery consistency identification method based on voltage curve shape in an embodiment;

[0025] Figure 2 2 is a diagram showing the effect of implementing the lithium-ion battery consistency identification method based on the voltage curve shape in the embodiment. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0027] Example 1

[0028] like Figure 1 As described above, this embodiment provides a lithium-ion battery consistency identification method based on voltage curve shape, comprising the following steps:

[0029] Step S1, online measurement of the voltage of each single battery during the charge and discharge process;

[0030] Step S2, calculating the shape dissimilarity matrix based on the derivative dynamic time warping distance;

[0031] Step S3, calculating based on the non-metric multidimensional scaling method of the shape dissimilarity matrix;

[0032] Step S4, normalization of the non-metric multidimensional scaling results;

[0033] Step S5: Visual recognition of multi-parameter consistency of the non-metric multi-dimensional scaling fitting composition.

[0034] The detailed description of each step is as follows:

[0035] Step S1, online measurement of the voltage of each single cell during the charge and discharge process: for each charge or discharge stage, the voltage value of each single cell at different times is collected and recorded by the measuring equipment. The voltage value of the i-th cell at the k-th sampling point is V i,k , the voltage curves measured on N single cells during each charging or discharging stage can be expressed as:

[0036]

[0037] Among them, V i =[V i,1 ,…,V i,M ] T , i = 1, 2, ..., N is the voltage curve of the i-th single cell in the charging or discharging stage, N is the number of single cells, M is the number of measurement points in the charging or discharging stage, and T represents the transpose operator.

[0038] Step S2, shape dissimilarity matrix calculation based on derivative dynamic time warping distance: for the battery voltage curve V obtained in step S1 i =[V i,1 ,V i,2 ,…,V i,k ,…,V i,M ] T , use the following formula to calculate the derivative value of each voltage point:

[0039]

[0040] Since the derivative values ​​at the starting point and the end point cannot be calculated according to the above formula, the voltage derivative matrix can be expressed as:

[0041]

[0042] For the i-th cell voltage derivative time series v i =[v i,1 ,v i,2 ,…,v i,k …,v i,M-2 ] T and the jth voltage derivative time series v j =[v j,1 ,v j,2 ,…,v j,l …,v j,M-2 ] T , construct a matrix D, where the element d i,j (k,l),1≤k,l≤M-2 means v i The kth element in v j The Euclidean distance between the lth elements in . Using the dynamic regularization algorithm, calculate v i and v j The distance r i,j , the specific formula is as follows:

[0043]

[0044] where r i,j(k, l) represents the cumulative distance on the path from (1, 1) to (k, l) in matrix D. According to the above formula, starting from (1, 1) and ending at (M-2, M-2), the final cumulative distance is the derivative dynamic time warping distance ddtw i,j .

[0045] Calculate N(N-1) / 2 times and you can get the following dissimilarity matrix:

[0046]

[0047] Step S3, calculation based on non-metric multidimensional scaling of shape dissimilarity matrix: using non-metric multidimensional scaling, the dissimilarity matrix DDTW is mapped to p-dimensional space, and the coordinates X1, X2, ... X of each single cell in the p-dimensional space are calculated. N The non-metric multidimensional scaling uses the stress coefficient Stress1:

[0048]

[0049] Among them, β i,j for ddtw i,j The equivalent distance in the p-dimensional space is calculated using the following Euclidean distance:

[0050]

[0051] By using the direct gradient descent method to solve the Stress1 minimization problem, the coordinates of each single battery in the p-dimensional space can be obtained. In this embodiment, p=2 is selected.

[0052] Step S4, normalization of the non-metric multidimensional scaling results: based on the coordinates X1, X2, ... X obtained in step 4 N , perform normalization to [-1,1] on the sth dimension, where s=1,2,…,p. and obtain the coordinates Y1,Y2,…Y N , where element Y i Calculate according to the following formula:

[0053]

[0054] Step S5, multi-parameter consistency visualization recognition based on non-metric multi-dimensional scaling fitting composition: coordinates Y1, Y2, ...Y NThe coordinates of N individual cells in p-dimensional space correspond to N discrete points in two-dimensional space. Using visualization, these N discrete points are represented within a two-dimensional fitted graph. Cells with identical internal resistance are clustered in the graph, and within each cluster, lines of equal SOC and equal capacity are formed. The positional relationship of each point within the p-dimensional fitted graph allows the source of parameter differences to be identified. The greater the distance between two points, the greater the difference between the corresponding cells.

[0055] The above-mentioned lithium-ion battery consistency identification method based on voltage curve shape is used to visually identify the multi-parameter consistency of 484 lithium-ion single batteries under the constant discharge condition which is common in both offline and online applications. The implementation effect is as follows: Figure 2 shown.

[0056] This embodiment recognizes that different battery internal parameters have different effects on the shape of the voltage curve. Therefore, based on the curve shape characteristics, the derivative dynamic time warping distance is used to construct a voltage shape dissimilarity matrix, and the non-metric multidimensional scaling method is used to transform this shape dissimilarity into a low-dimensional space, thereby realizing the identification of internal parameter differences such as capacity, internal resistance, and SOC between single cells.

[0057] Example 2

[0058] This embodiment provides an electronic device, including a processor and a memory, wherein the memory stores instructions for executing the lithium-ion battery consistency identification method based on voltage curve shape as in Embodiment 1.

[0059] Example 3

[0060] This embodiment provides a computer-readable storage medium including a program for execution by a processor of an electronic device, wherein the program includes instructions for executing the lithium-ion battery consistency identification method based on voltage curve shape as in Embodiment 1.

[0061] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A lithium-ion battery consistency identification method based on voltage curve shape, characterized in that: The steps include: Get the voltage information of each battery; Obtaining a shape dissimilarity matrix based on a derivative dynamic time warping distance according to the voltage information; According to the non-similarity matrix, a non-metric multidimensional scaling method is used to obtain the coordinate information of each battery in the p-dimensional space and obtain a non-metric multidimensional scaling result; According to the non-metric multidimensional scaling results, a multi-parameter consistency visualization recognition based on the non-metric multidimensional scaling fitting composition is realized, The implementation process of the multi-parameter consistency visual recognition based on the non-metric multi-dimensional scaling fitting composition includes the following steps: Representing the non-metric multidimensional scaling results in a two-dimensional spatial fitting composition, clustering cells with the same internal resistance, and forming equal SOC lines and equal capacity lines within each cluster; According to the positional relationship of each point in the p-dimensional space fitting composition, the farther the distance between two points, the greater the difference in the corresponding single batteries, thus realizing multi-parameter consistency visualization recognition based on non-metric multidimensional scaling fitting composition.

2. The method for identifying consistency of lithium-ion batteries based on voltage curve shape according to claim 1, characterized in that: The step of obtaining the voltage information of each battery includes: The voltage of each single battery is measured during the charge and discharge process to obtain the online voltage of each battery in multiple measurement stages.

3. The method for identifying consistency of lithium-ion batteries based on voltage curve shape according to claim 1, wherein: The process of obtaining the dissimilarity matrix includes the following steps: Obtaining a voltage derivative matrix according to the voltage information; The dissimilarity matrix is ​​obtained according to the voltage derivative matrix.

4. The method for identifying consistency of lithium-ion batteries based on voltage curve shape according to claim 3, wherein: The process of obtaining the voltage derivative matrix includes the following steps: According to the online voltage of each battery measured in multiple measurement stages, the voltage derivative matrix is ​​obtained by calculating the derivative value of each voltage point.

5. The method for identifying consistency of lithium-ion batteries based on voltage curve shape according to claim 3, wherein: The process of obtaining the dissimilarity matrix includes the following steps: A dynamic regularization algorithm is adopted to obtain the dissimilarity matrix according to the voltage derivative matrix.

6. The method for identifying consistency of lithium-ion batteries based on voltage curve shape according to claim 1, characterized in that: The process of obtaining the non-metric multidimensional scaling results includes the following steps: Using a non-metric multidimensional scaling method, according to the dissimilarity matrix, to obtain the coordinate information of each battery in the p-dimensional space; The coordinate information is normalized to obtain the non-metric multidimensional scaling result.

7. The method for identifying consistency of lithium-ion batteries based on voltage curve shape according to claim 6, characterized in that: The process of obtaining the coordinate information of the p-dimensional space includes the following steps: Obtaining a stress coefficient for non-metric multidimensional scaling according to the dissimilarity matrix; A direct gradient descent method is used to solve the minimization problem of the stress coefficient for non-metric multidimensional scaling to obtain the coordinate information of the p-dimensional space.

8. An electronic device, characterized in that: include: One or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the lithium-ion battery consistency identification method based on voltage curve shape as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that The method comprises one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the lithium-ion battery consistency identification method based on voltage curve shape as described in any one of claims 1 to 7.

Images