Lithium battery pack consistency scoring method based on fuzzy analytic hierarchy process

Abstract

The invention provides a lithium battery pack consistency scoring method based on fuzzy analytic hierarchy process, which belongs to the technical field of battery evaluation, and comprises the following steps: S1, obtaining consistency evaluation indexes of each battery monomer in a lithium battery pack, including an outlier factor, a standard deviation, a correlation coefficient between the battery monomers and outlier times; s2, performing fuzzy analytic hierarchy process to obtain a consistency score of the to-be-evaluated battery monomers; and S3, performing fuzzy analytic hierarchy process according to the consistency evaluation indexes of all the battery monomers in the lithium battery pack to obtain a consistency score of the lithium battery pack, determining a fault monomer according to the consistency score of each battery monomer, and evaluating a fault level. The method has the beneficial effects that through the fuzzy analytic hierarchy process, the outlier factor, the standard deviation, the correlation coefficient and the outlier frequency are adopted as consistency evaluation indexes, the performance of the single battery and the performance of the whole lithium battery pack are quantitatively evaluated, and the accuracy of fault judgment of the single battery is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery evaluation, and in particular to a lithium battery pack consistency scoring method based on fuzzy hierarchical analysis. Background Art

[0002] As the world pays more attention to renewable energy and electric transportation, lithium batteries have been widely used as an efficient and environmentally friendly energy storage solution. However, due to slight differences in the manufacturing process, material inhomogeneities, and changes in operating conditions, the performance and life of lithium batteries can vary greatly. This variability is called the battery consistency problem, which directly affects the overall performance, safety, and reliability of the battery pack.

[0003] The Chinese patent application with publication number CN115792678A is "A method for evaluating the performance of automotive power batteries based on fuzzy hierarchical analysis method". This invention uses fuzzy hierarchical analysis method to evaluate the overall performance, charging performance, discharging performance and safety performance of the battery according to factors such as the battery's charging and discharging speed, life, and temperature, and then combines these four performances to give a power battery performance score.

[0004] However, scoring power battery performance only ensures that the battery can enter the market, and currently there is no way to evaluate the performance of the battery during use. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a lithium battery pack consistency scoring method based on fuzzy hierarchical analysis.

[0006] The technical problem solved by the present invention can be achieved by adopting the following technical solutions:

[0007] A lithium battery pack consistency scoring method based on fuzzy hierarchical analysis, comprising:

[0008] Step S1, obtaining consistency evaluation indicators of each battery cell in a lithium battery pack, wherein the consistency evaluation indicators include an outlier factor, a standard deviation, a correlation coefficient between battery cells, and a number of outlier times;

[0009] Step S2, taking any battery cell as the battery cell to be evaluated, performing fuzzy hierarchical analysis processing according to the consistency evaluation index of the battery cell to be evaluated, and obtaining the consistency score of the battery cell to be evaluated;

[0010] Step S3, performing fuzzy hierarchical analysis processing according to the consistency evaluation indexes of all the battery cells in the lithium battery pack to obtain the consistency score of the lithium battery pack, and determining the faulty cell according to the consistency score of each battery cell, and assessing the fault level.

[0011] Preferably, in step S1, the outlier factor is a local outlier factor;

[0012] The local outlier factor is a ratio of an average local reachable density of all the battery cells within a k-distance neighborhood of the battery cell to be evaluated to the local reachable density of the battery cell to be evaluated, where k is a positive integer.

[0013] Preferably, the kth distance of the k-distance neighborhood is the kth in the order of distances between the battery cell to be evaluated and other battery cells in the lithium battery pack.

[0014] Preferably, in step S1, the standard deviation is calculated using the following formula:

[0015]

[0016] in, represents the average voltage of the jth battery cell; u nj represents the voltage value of the nth sampling point of the jth battery cell; std represents the standard deviation.

[0017] Preferably, in step S1, the correlation coefficient between the battery cells is calculated using the following formula:

[0018]

[0019] Where r represents the correlation coefficient between the x-th battery cell and the y-th battery cell; cov(x,y) is the covariance between the x-th battery cell and the y-th battery cell; σx represents the variance of the x-th battery cell; σy represents the variance of the y-th battery cell; represents the mean value of the x-th battery cell; represents the mean value of the y-th battery cell; x i represents the voltage value of the ith sampling point of the xth battery cell; y i represents the voltage value of the ith sampling point of the yth battery cell; n represents the length of the cell voltage data.

[0020] Preferably, in step S1, the method for obtaining the number of outliers includes:

[0021] The density-based clustering algorithm determines whether each battery cell is an outlier at each sampling point:

[0022] If so, the outlier number of the battery cell is 1;

[0023] If not, the outlier number of the battery cell is 0;

[0024] All sampling point moments are traversed, and the outlier number of the battery cell is determined according to the outlier number at all moments.

[0025] Preferably, the density-based clustering algorithm is a density-based noise-applied spatial clustering algorithm.

[0026] Preferably, step S2 comprises:

[0027] Step S21, constructing a fuzzy judgment matrix P based on the outlier factor, standard deviation, correlation coefficient between battery cells and outlier times of the battery cells to be evaluated; the fuzzy judgment matrix P is:

[0028]

[0029] Where xa,xb represent consistency evaluation indicators respectively, a,b=1,2,...,m, m represents the total number of consistency evaluation indicators; xab represents the relative importance measure of xa to xb, and xab+xba=1 is satisfied;

[0030] Step S22, converting the sum of each row element in the fuzzy judgment matrix by a conversion formula to obtain a conversion matrix; the conversion formula is:

[0031] x′ ab =(θ a -θ b ) / 2m+0.5

[0032] Among them, θ a ,θ b Respectively represent the sum of the elements in the ath and bth rows of the fuzzy judgment matrix; x′ ab represents the converted value of the relevant importance measure xab in the fuzzy judgment matrix;

[0033] Step S23, according to the conversion matrix, obtain the weight vector of each consistency evaluation index in the entire hierarchy The calculation formula is:

[0034]

[0035] Step S24, the weight vector Normalize to obtain a single-layer weight vector W, where each component W of the weight vector W is a The calculation method is:

[0036]

[0037] Step S25: Based on the pre-built battery consistency scoring system, determine the indicator interval corresponding to each consistency evaluation indicator of the battery cell to be evaluated and the indicator score corresponding to the indicator interval, and calculate the components W of the weight vector W according to the determined indicator score and step S24. a A weighted sum is performed to obtain the consistency score of the battery cell.

[0038] Preferably, in step S25, the method for constructing the battery consistency scoring system includes:

[0039] Determine the indicator interval and the indicator score corresponding to each indicator interval according to the threshold value corresponding to each consistency evaluation indicator, and construct the battery consistency scoring system;

[0040] The threshold value corresponding to the consistency evaluation index is determined as follows:

[0041] T a =μ a +3σ a

[0042] Among them, T a represents the threshold of the ath consistency evaluation indicator; σ represents the standard deviation; μ represents the mean of the standard deviation.

[0043] Preferably, step S3 comprises:

[0044] Step S31, performing mean processing on each consistency evaluation index of all the battery cells to obtain the mean value of each consistency evaluation index;

[0045] Step S32: using the mean of the consistency evaluation index as the consistency evaluation index of the lithium battery pack, and determining the consistency score of the lithium battery pack based on the battery consistency scoring system.

[0046] The advantages or beneficial effects of the technical solution of the present invention are:

[0047] The present invention adopts fuzzy hierarchical analysis method, adopts outlier factor, standard deviation, correlation coefficient and outlier number as consistency evaluation indexes, quantitatively evaluates the performance of a single battery cell and the entire lithium battery pack, and improves the accuracy of fault judgment of battery cells. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of a flow chart of a lithium battery pack consistency scoring method based on fuzzy hierarchical analysis in a preferred embodiment of the present invention;

[0049] Figure 2 A schematic diagram of the overall algorithm flow in a preferred embodiment of the present invention;

[0050] Figure 3 A schematic diagram of a flow chart of a local outlier factor algorithm in a preferred embodiment of the present invention;

[0051] Figure 4 A schematic diagram of a flow chart of a density-based clustering algorithm in a preferred embodiment of the present invention;

[0052] Figure 5 A schematic diagram of the consistency score of a lithium battery pack in a preferred embodiment of the present invention;

[0053] Figure 6 A schematic diagram of a hierarchical structure model of a fuzzy hierarchical classification method in a preferred embodiment of the present invention; Figure 7 FIG. 4 is a flow chart of step S3 in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0056] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0057] See also Figure 1 In a preferred embodiment of the present invention, based on the above problems existing in the prior art, a lithium battery pack consistency scoring method based on fuzzy hierarchical analysis is provided, comprising:

[0058] Step S1, obtaining consistency evaluation indicators of each battery cell in the lithium battery pack, the consistency evaluation indicators including outlier factor, standard deviation, correlation coefficient between battery cells and outlier number;

[0059] Step S2, taking any battery cell as the battery cell to be evaluated, performing fuzzy hierarchical analysis processing according to the consistency evaluation index of the battery cell to be evaluated, and obtaining the consistency score of the battery cell to be evaluated;

[0060] Step S3, performing fuzzy hierarchical analysis processing according to the consistency evaluation index of all battery cells in the lithium battery pack to obtain the consistency score of the lithium battery pack, and determining the faulty cell according to the consistency score of each battery cell, and assessing the fault level.

[0061] Specifically, in response to the consistency failure problem of lithium battery packs during use, the embodiment of the present invention proposes a lithium battery pack consistency scoring method based on fuzzy analytic hierarchy process (FAHP), such as Figure 2 As shown in the figure, the output results of the four algorithms, namely the local outlier factor (LOF) algorithm, the standard deviation algorithm, the correlation coefficient method and the density-based clustering, are used to build a consistency scoring model through the fuzzy hierarchical analysis method, and then the consistency score of the battery cell is obtained. The consistency fault cell is analyzed according to the score threshold, and the fault level is judged by the consistency score.

[0062] Furthermore, by inputting the mean of the four algorithm results of all battery cells into the fuzzy hierarchical analysis model for scoring, the consistency score of the entire lithium battery pack can be obtained, and then the overall performance of the battery can be analyzed.

[0063] The specific implementation process of the LOF algorithm, standard deviation algorithm, correlation coefficient method, density-based clustering algorithm, and fuzzy hierarchical analysis method used in the embodiments of the present invention is exemplarily described below:

[0064] The LOF algorithm is an algorithm for anomaly detection. It is a density-based outlier detection method. The core idea is that the density of outliers is less than the density of other points in the neighborhood, that is, all other points in the neighborhood are relatively concentrated, while outliers are relatively discrete. LOF reflects the degree of discreteness of data points by detecting relative density. If the relative density of a data point is extremely low, it is considered that the probability that the data point is an outlier is high. The higher the outlier factor, the greater the degree of outlier.

[0065] As a preferred embodiment, in step S1, the outlier factor is a local outlier factor;

[0066] The local outlier factor is the ratio of the average local reachable density of all battery cells in the k-distance neighborhood of the battery cell to be evaluated to the local reachable density of the battery cell to be evaluated, where k is a positive integer.

[0067] like Figure 3 As shown, the LOF algorithm includes the following steps:

[0068] S101, inputting a data set; the data set includes voltage curves of each battery cell. The performance of the battery during use can be reflected from these voltage curves.

[0069] S102, the distance between any two monomers;

[0070] Specifically, the voltage curves of any two battery cells in the data set are used to calculate the distance between the two cells. The distance is usually calculated using the Euclidean distance, and the calculation formula is as follows:

[0071]

[0072] Among them, X and Y are the voltage curves corresponding to the two battery cells respectively;

[0073] x i ,y i Respectively represent the corresponding voltage values ​​on the two voltage curves;

[0074] X={x1,...,xn},Y={y1,...,yn} is the set of all data points on the two voltage curves, i=1,...,n, n represents the total number of sampling points on the voltage curve;

[0075] dist(X,Y) represents the Euclidean distance between voltage curve X and voltage curve Y.

[0076] In addition to the commonly used Euclidean distance, other distance metrics can also be used, such as the Hamming distance.

[0077] S103, kth distance;

[0078] Specifically, in a given data set, for any point x, the k-th distance of point x refers to the distance from the k-th nearest neighbor of point x to point x, also referred to as k-distance.

[0079] As a preferred implementation, the k-th distance in the k-distance neighborhood is the k-th in the order of distances between the battery cell to be evaluated and other battery cells in the lithium battery pack.

[0080] Specifically, according to the Euclidean distance calculated in step S102, the distances between the point (eg, the battery cell to be evaluated) and other points in the data set are sorted in ascending order, and the kth corresponding Euclidean distance in the sorting is the kth distance.

[0081] k is a positive integer and can be set according to actual application requirements and is not limited here.

[0082] S104, k-distance neighborhood;

[0083] Specifically, the k-distance neighborhood refers to a set of all points whose distance to point x is less than the kth distance in a given data set for any point x.

[0084] The distances from all points in the k-distance neighborhood to point x are less than or equal to the k-th distance, and the set of all points in the neighborhood is recorded as N(x,k).

[0085] S105, local reachable distance;

[0086] Specifically, the local reachable distance dist(x,y) is:

[0087] dist(x,y)=max{d k ,d(x,y)}

[0088] Where dist(x,y) represents the local reachable distance of point x, y∈N(x,k);

[0089] d k represents the kth distance;

[0090] d(x,y) represents the actual distance, and the above-mentioned Euclidean distance can be used.

[0091] The local reachable distance dist(x,y) is the maximum value between the kth distance and the actual distance, that is:

[0092] If the actual distance from other points to point x is less than its kth distance, then the local reachable distance dist(x,y) is the kth distance;

[0093] If the actual distance from other points to point x is greater than its kth distance, then the local reachable distance dist(x,y) is the actual distance.

[0094] S106, local reachability density;

[0095] Specifically, the local reachability density is defined as the reciprocal of the average local reachability distance from a point in the k-distance neighborhood to point x, and its calculation formula is:

[0096]

[0097] Wherein, denisty(x,k) represents the local reachability density of the k-distance neighborhood of the x-th battery cell;

[0098] N(x,k) represents the set of all battery cells within the k-distance neighborhood of the x-th battery cell; the battery cell x refers to the battery cell to be evaluated, which is any battery cell to be evaluated in the lithium battery pack.

[0099] |N(x,k)| represents the number of battery cells in the set N(x,k), and battery cell x itself is not included in the set.

[0100] S107, local outlier factor.

[0101] Specifically, the local outlier factor is equal to the average of the local reachability density of the points in the k-distance neighborhood divided by the local reachability density of point x. The local outlier factor is calculated using the following formula:

[0102]

[0103] Where denisty(y,k) represents the local reachability density of the k-distance neighborhood of the y-th battery cell;

[0104] LOF(x,k) represents the outlier factor of the k-distance neighborhood of the x-th battery cell.

[0105] Through the above steps, the LOF outlier factor of each battery cell is calculated.

[0106] As a preferred embodiment, in step S1, the standard deviation is calculated using the following formula:

[0107]

[0108] in, represents the average voltage value of the jth battery cell, that is, the average voltage value of the voltage curve of battery cell j in the data set; u nj represents the voltage value of the nth sampling point of the jth battery cell; std represents the standard deviation.

[0109] Furthermore, by calculating the Pearson correlation coefficient between the voltages of two adjacent battery cells in the lithium battery pack, the linear correlation between the voltages of the two cells is measured, thereby comparing the voltage consistency of the battery pack. As a preferred embodiment, in step S1, the correlation coefficient between the battery cells is calculated using the following formula:

[0110]

[0111] Where r represents the correlation coefficient between the x-th battery cell and the y-th battery cell; cov(x,y) is the covariance between the x-th battery cell and the y-th battery cell; σx represents the variance of the x-th battery cell; σy represents the variance of the y-th battery cell; represents the mean value of the x-th battery cell; represents the mean value of the y-th battery cell; x i represents the voltage value of the ith sampling point of the xth battery cell; y i represents the voltage value of the ith sampling point of the yth battery cell; n represents the length of the cell voltage data.

[0112] As a preferred embodiment, in step S1, the method for obtaining the number of outliers includes:

[0113] The density-based clustering algorithm determines whether each battery cell is an outlier at each sampling point:

[0114] If so, the number of outliers of the battery cell is 1;

[0115] If not, the number of outliers of the battery cell is 0;

[0116] Traverse all sampling points and determine the number of outliers of the battery cell according to the number of outliers at all times. The number of outliers of each battery cell is the number of times it is determined to be an outlier at all times.

[0117] Specifically, in this embodiment, the number of outliers is obtained based on a density-based clustering algorithm.

[0118] As a preferred implementation, the density-based clustering algorithm is a Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm.

[0119] Specifically, in this embodiment, DBSCAN is a typical density-based clustering algorithm that describes the compactness of a sample set based on a set of neighborhoods (ε, MinPts).

[0120] The data set of the DBSCAN algorithm is the voltage values ​​of all battery cells at the same time, such as Figure 4 As shown, the following steps are included:

[0121] S141, setting parameters MinPts and ε;

[0122] Among them, ε describes the neighborhood distance threshold of a certain sample, and MinPts describes the threshold of the number of samples in the neighborhood of a certain sample with a distance of ε.

[0123] S142, determine whether it is a core sample:

[0124] Among them, a circle is drawn with sample point P as the center and ε as the radius. If there are at least MinPts samples (including P itself) in the circle, P is called a core sample; otherwise, P is called a non-core sample.

[0125] S143, determine whether it is reachable:

[0126] Among them, all points in the circular domain are directly reachable from point P.

[0127] For example, if core sample P1 is directly reachable to core sample P2, and core sample P2 is directly reachable to core sample P3, then core sample P3 is reachable to core sample P1.

[0128] If a point is not a core sample or is unreachable from other core sample points, the point is recorded as an outlier.

[0129] The DBSCAN algorithm traverses all moments, and at each moment it determines whether each battery cell is an outlier: if it is an outlier, the outlier count is 1; if it is not an outlier, the outlier count is 0. In this way, the full voltage data sequence becomes a sequence of only 0 and 1, and then the sum of the columns can be obtained to obtain the total number of times the battery cell is diagnosed as an outlier, which is the final outlier count of the battery cell.

[0130] The embodiment of the present invention uses the calculation results of the LOF outlier factor algorithm, the standard deviation algorithm, the correlation coefficient method, and the DBSCAN clustering algorithm as the input of the fuzzy analytic hierarchy process (FAHP) method, and outputs the consistency scores of the battery cells and the entire lithium battery pack.

[0131] FAHP method decomposes the decision problem into multiple levels, each level contains several criteria or schemes, and obtains the final decision result by fuzzy comprehensive evaluation of each level. The hierarchical structure of FAHP method includes target level and criterion level.

[0132] First, determine the decision hierarchy. The purpose of the embodiment of the present invention is to score the consistency of the battery, so there is only one target layer, namely, battery consistency; the score is scored based on four factors: LOF outlier factor, standard deviation, correlation coefficient, and number of outliers, so the criterion layer includes the above four judgment elements.

[0133] like Figure 5 As shown, the fuzzy hierarchical analysis in step S2 specifically includes the following steps:

[0134] Step S21, constructing a fuzzy judgment matrix P based on the outlier factor, standard deviation, correlation coefficient between battery cells and outlier times of the battery cells to be evaluated; the fuzzy judgment matrix P is:

[0135]

[0136] Where xa,xb represent consistency evaluation indicators, i.e., judgment factors at the criterion level, a,b=1,2,...,m, m represents the total number of consistency evaluation indicators; xab represents the relative importance measure of xa to xb, and xab+xba=1 is satisfied;

[0137] Exemplarily, in the embodiment of the present invention, the judgment elements are outlier factor, standard deviation, correlation coefficient and outlier number, which include 4 in total, namely m=4; therefore, the fuzzy judgment matrix P is:

[0138]

[0139] x1 represents the LOF outlier factor;

[0140] x2 represents the standard deviation;

[0141] x3 represents the correlation coefficient;

[0142] x4 represents the number of outliers;

[0143] Among them, the elements on the main diagonal (i.e., the diagonal from the upper left corner to the lower right corner of the matrix) in the fuzzy judgment matrix P are 1. For example, x11=x22=x33=x44=1, so:

[0144]

[0145] In the matrix P, x12 represents a measure of the relative importance of x1 to x2; x12 is not necessarily equal to x21, but x12+x21=1.

[0146] Step S22, converting the sum of each row element in the fuzzy judgment matrix by a conversion formula to obtain a conversion matrix;

[0147] Among them, the row summation of the fuzzy judgment matrix is ​​as follows:

[0148]

[0149] Among them, θ a represents the sum of the elements in the ath row of the fuzzy judgment matrix, that is, θ a =xa1+xa2+...+xam;

[0150] Exemplarily, θ1 represents the sum of the elements in the first row of the fuzzy judgment matrix, that is, θ1=x11+x12+x13+x14. According to this formula, the sum of the elements in all rows, that is, θ1, θ2, θ3, θ4, can be calculated.

[0151] The conversion formula is:

[0152] x′ ab =(θ a -θ b ) / 2m+0.5

[0153] Among them, θ a ,θ b Respectively represent the sum of the elements in the ath and bth rows of the fuzzy judgment matrix; x′ ab Represents the converted value of the relevant importance measure xab in the fuzzy judgment matrix;

[0154] According to the above conversion formula, the values ​​of each row and column element in the conversion matrix P' can be calculated. The conversion matrix P' is:

[0155]

[0156] Exemplarily, x'1m=(θ1-θm) / 2m+0.5.

[0157] Step S23, according to the conversion matrix, obtain the weight vector of each consistency evaluation index in the entire hierarchy The calculation formula is:

[0158]

[0159] in, Represents the weight vector of the ath judgment factor (LOF outlier factor, standard deviation, correlation coefficient and number of outliers);

[0160] Specifically, by operating the fuzzy judgment matrix, the weight vector of each judgment factor in the entire hierarchy is obtained. Weight vector It can be used to evaluate the importance of each factor and serve as the basis for multi-criteria decision analysis.

[0161] For example, for row 1 in the transformation matrix, the weight vector of the LOF outlier factor is for:

[0162]

[0163] In this way, the weight vector of all rows in the transformation matrix can be calculated, that is, the weight vector of the LOF outlier factor Weight vector for standard deviation Weight vector of correlation coefficients Weight vector of outlier counts

[0164] Step S24: weight vector Normalize to get the weight vector W of a single layer, and the components W of the weight vector W a The calculation method is:

[0165]

[0166] Among them, W a Represents the normalized weight vector of the ath judgment factor (LOF outlier factor, standard deviation, correlation coefficient, and number of outliers), a=1,...,m.

[0167] Exemplarily, the normalized weight vector W1 of the LOF outlier factor is:

[0168]

[0169] By analogy, the weight vector W2 of the normalized standard deviation, the weight vector W3 of the normalized correlation coefficient, and the weight vector W4 of the normalized number of outliers can be calculated.

[0170] Step S25: Based on the pre-built battery consistency scoring system, determine the indicator interval corresponding to each consistency evaluation indicator of the battery cell to be evaluated and the indicator score corresponding to the indicator interval, and calculate the components W of the weight vector W according to the determined indicator score and step S24. a Perform a weighted sum to obtain the consistency score of the battery cell.

[0171] Specifically, after obtaining the weight vector W of each judgment factor a After that, the threshold of each judgment factor is obtained, and then based on the pre-constructed battery consistency scoring system, the consistency index is determined in which index range, the corresponding index score is obtained, multiplied by the corresponding weight vector, and then all the consistency index scores are added together to obtain the total consistency score of the battery cell.

[0172] As a preferred embodiment, in step S25, the method for constructing a battery consistency scoring system includes:

[0173] According to the threshold value corresponding to each consistency evaluation indicator, the indicator interval and the indicator score corresponding to each indicator interval are determined to construct a battery consistency scoring system;

[0174] The threshold value corresponding to the consistency evaluation indicator is determined as follows:

[0175] T a =μ a +3σ a

[0176] Among them, T a represents the threshold of the ath consistency evaluation indicator; σ a Represents the standard deviation of the battery cell to be evaluated; μ a Represents the mean standard deviation of all battery cells.

[0177] Specifically, each algorithm corresponds to a threshold. Taking the outlier factor algorithm as an example, the threshold T1 of the outlier factor algorithm is solved as follows:

[0178] First, the outlier factor of each battery cell is calculated according to the outlier factor algorithm; then, the mean μ1 and standard deviation σ1 of the outlier factors of all cells are calculated, and then the threshold T1 = μ1 + 3σ1.

[0179] In this way, the thresholds corresponding to the various consistency evaluation indicators can be calculated, namely, the LOF outlier factor corresponding threshold T1, the standard deviation corresponding threshold T2, the correlation coefficient corresponding threshold T3, and the outlier number corresponding threshold T4.

[0180] For example, the constructed battery consistency scoring system is shown in Table 1 below:

[0181] Table 1 Battery consistency scoring system

[0182]

[0183] It can be seen from Table 1 above that in the battery consistency scoring system, each indicator (such as LOF outlier factor, standard deviation, correlation coefficient, number of outliers) corresponds to a threshold (T1, T2, T3, T4), and the indicator interval is determined according to the threshold. Each indicator interval corresponds to an indicator score range.

[0184] The indicator interval is divided into multiple intervals according to the threshold. Generally speaking, if the indicator is 0, the corresponding indicator score is 100; if the indicator is greater than the corresponding threshold, the corresponding indicator score is 0.

[0185] like Figure 6 A schematic diagram of the consistency scores of all battery cells in a lithium battery pack is shown, with the horizontal axis being the cell number and the vertical axis being the corresponding consistency score.

[0186] The consistency scores of all battery cells in the lithium battery pack were statistically analyzed, and the statistical results are shown in Table 2 below. The consistency score results of some batteries are shown in Table 3 below.

[0187] Table 2 Statistical results of battery consistency scores based on FAHP

[0188]

[0189] Table 3. Results of some battery consistency scores

[0190]

[0191]

[0192] According to the consistency score of the battery cells, it is possible to analyze whether there is a faulty cell, and judge the fault level according to the size of the score. As can be seen from Table 3 above, the consistency score of cell #195 is 0, which obviously has failed. The consistency scores of cells #9 and #101 are also low, indicating that there is a fault.

[0193] As a preferred embodiment, wherein Figure 7 As shown, step S3 includes:

[0194] Step S31, performing mean processing on each consistency evaluation index of all battery cells to obtain the mean value of each consistency evaluation index;

[0195] Step S32, using the mean of the consistency evaluation index as the consistency evaluation index of the lithium battery pack, and determining the consistency score of the lithium battery pack based on the battery consistency scoring system.

[0196] Specifically, in this embodiment, the consistency evaluation indicators of all battery cells are averaged, that is, the outlier factor mean, standard deviation mean, correlation coefficient mean, and outlier number mean are obtained. These means are used as consistency evaluation indicators of lithium battery packs and input into the above-mentioned battery consistency scoring system to obtain the consistency score of the lithium battery pack.

[0197] In the consistency evaluation process of the entire lithium battery pack, the outlier factor mean, standard deviation mean, correlation coefficient mean and outlier count mean are used as consistency evaluation indicators of the lithium battery pack. Based on the fuzzy analytic hierarchy process, the weight vector of each consistency evaluation indicator is determined in the same way as mentioned above; the indicator score is determined according to the indicator interval where the consistency evaluation indicator of the lithium battery pack is located, and the determined indicator score is weightedly summed with the weight vector of the corresponding indicator to obtain the consistency score of the lithium battery pack, thereby quantitatively evaluating the performance of the entire lithium battery pack.

[0198] For the consistency assessment of lithium battery packs, the fuzzy hierarchical analysis method is similar to the process of a single battery cell, except that:

[0199] In the consistency evaluation process of battery cells, the input of fuzzy hierarchical analysis (i.e., the judgment factors of the criterion layer) is the outlier factor, standard deviation, correlation coefficient and number of outliers of a single battery cell (i.e., the battery cell to be evaluated).

[0200] In the consistency evaluation process of the entire lithium battery pack, the judgment factors of the criterion layer of the fuzzy hierarchical analysis are the mean of the outlier factor, the mean of the standard deviation, the mean of the correlation coefficient and the mean of the number of outliers of all battery cells.

[0201] The advantages or beneficial effects of adopting the above technical solution are: the present invention adopts the fuzzy hierarchical analysis method, uses the outlier factor, standard deviation, correlation coefficient and number of outliers as consistency evaluation indicators, quantitatively evaluates the performance of a single battery cell and the entire lithium battery pack, and improves the accuracy of battery cell fault judgment.

[0202] The above are only preferred embodiments of the present invention, and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the contents of this specification and illustrations should be included in the protection scope of the present invention.

Claims

1. A lithium battery pack consistency scoring method based on fuzzy hierarchical analysis, It is characterized in that include: Step S1, obtaining consistency evaluation indicators of each battery cell in a lithium battery pack, wherein the consistency evaluation indicators include an outlier factor, a standard deviation, a correlation coefficient between battery cells, and a number of outlier times; Step S2, taking any battery cell as the battery cell to be evaluated, performing fuzzy hierarchical analysis processing according to the consistency evaluation index of the battery cell to be evaluated, and obtaining the consistency score of the battery cell to be evaluated; Step S3, performing fuzzy hierarchical analysis processing according to the consistency evaluation indexes of all the battery cells in the lithium battery pack to obtain the consistency score of the lithium battery pack, and determining the faulty cell according to the consistency score of each battery cell, and assessing the fault level.

2. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 1, It is characterized in that In the step S1, the outlier factor is a local outlier factor; The local outlier factor is a ratio of an average local reachable density of all the battery cells within a k-distance neighborhood of the battery cell to be evaluated to the local reachable density of the battery cell to be evaluated, where k is a positive integer.

3. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 2, It is characterized in that The k-th distance of the k-distance neighborhood is the k-th distance in the order of the distances between the battery cell to be evaluated and other battery cells in the lithium battery pack.

4. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 1, It is characterized in that In step S1, the standard deviation is calculated using the following formula: in, represents the average voltage of the jth battery cell; u nj represents the voltage value of the nth sampling point of the jth battery cell; std represents the standard deviation.

5. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 1, It is characterized in that In the step S1, the correlation coefficient between the battery cells is calculated using the following formula: Where r represents the correlation coefficient between the x-th battery cell and the y-th battery cell; cov(x,y) is the covariance between the x-th battery cell and the y-th battery cell; σx represents the variance of the x-th battery cell; σy represents the variance of the y-th battery cell; represents the mean value of the x-th battery cell; represents the mean value of the y-th battery cell; x i represents the voltage value of the ith sampling point of the xth battery cell; y i represents the voltage value of the ith sampling point of the yth battery cell; n represents the length of the cell voltage data.

6. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 1, It is characterized in that In step S1, the method for obtaining the number of outliers includes: The density-based clustering algorithm determines whether each battery cell is an outlier at each sampling point: If so, the outlier number of the battery cell is 1; If not, the outlier number of the battery cell is 0; All sampling point moments are traversed, and the outlier number of the battery cell is determined according to the outlier number at all moments.

7. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 6, It is characterized in that The density-based clustering algorithm is a density-based noise application spatial clustering algorithm.

8. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 1, It is characterized in that The step S2 comprises: Step S21, constructing a fuzzy judgment matrix P based on the outlier factor, standard deviation, correlation coefficient between battery cells and outlier times of the battery cells to be evaluated; the fuzzy judgment matrix P is: Where xa,xb represent consistency evaluation indicators respectively, a,b=1,2,...,m, m represents the total number of consistency evaluation indicators; xab represents the relative importance measure of xa to xb, and xab+xba=1 is satisfied; Step S22, converting the sum of each row element in the fuzzy judgment matrix by a conversion formula to obtain a conversion matrix; the conversion formula is: x′ ab =(θ a -θ b ) / 2m+0.5 Among them, θ a ,θ b Respectively represent the sum of the elements in the ath and bth rows of the fuzzy judgment matrix; x′ ab represents the converted value of the relevant importance measure xab in the fuzzy judgment matrix; Step S23, according to the conversion matrix, obtain the weight vector of each consistency evaluation index in the entire hierarchy The calculation formula is: Step S24, the weight vector Normalize to obtain a single-layer weight vector W, where each component W of the weight vector W is a The calculation method is: Step S25: Based on the pre-built battery consistency scoring system, determine the indicator interval corresponding to each consistency evaluation indicator of the battery cell to be evaluated and the indicator score corresponding to the indicator interval, and calculate the components W of the weight vector W according to the determined indicator score and step S24. a A weighted sum is performed to obtain the consistency score of the battery cell.

9. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 8, It is characterized in that In step S25, the method for constructing the battery consistency scoring system includes: Determine the indicator interval and the indicator score corresponding to each indicator interval according to the threshold value corresponding to each consistency evaluation indicator, and construct the battery consistency scoring system; The threshold value corresponding to the consistency evaluation index is determined as follows: T a =μ a +3s a Among them, T a represents the threshold of the ath consistency evaluation indicator; σ represents the standard deviation; μ represents the mean of the standard deviation.

10. The lithium battery pack consistency scoring method based on fuzzy hierarchical analysis according to claim 8, It is characterized in that The step S3 comprises: Step S31, performing mean processing on each consistency evaluation index of all the battery cells to obtain the mean value of each consistency evaluation index; Step S32, using the mean of the consistency evaluation index as the consistency evaluation index of the lithium battery pack, and determining the consistency score of the lithium battery pack based on the battery consistency scoring system.

Images