Battery comprehensive utilization method and system based on multi-parameter collaborative optimization

Abstract

The application belongs to the technical field of battery sorting, and provides a battery comprehensive utilization method and system based on multi-parameter collaborative optimization, which obtains capacity and internal resistance parameters of a battery, and performs enhancement to generate an enhanced data set, analyzes the enhanced data set, removes abnormal fault batteries according to data distribution density, divides healthy batteries into high-consistency clusters through clustering of an initial center; determines a battery module reconstruction scheme according to a connection mode and a parameter distribution mode of single batteries in a battery pack; establishes a module capacity and internal resistance analysis formula based on the obtained relationship and the divided battery clusters, calculates the dispersion degree of capacity and internal resistance between single batteries in a battery module under different connection modes and parameter distribution combinations, constructs a consistency evaluation index, analyzes the inhibition effect of each battery module reconstruction scheme on the wood barrel effect, and determines an optimal battery module reconstruction scheme. The application improves the accuracy of battery sorting.

Description

Technical Field

[0001] This invention belongs to the field of battery comprehensive utilization technology, specifically relating to a battery comprehensive utilization method and system based on multi-parameter collaborative optimization. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Long-term use of batteries leads to gradual degradation of battery parameters. When their capacity decays to below 80% of their rated capacity, their range and performance decline significantly, necessitating retirement. Failure to effectively utilize these batteries not only results in economic losses but also triggers serious environmental problems. Reusing retired batteries not only maximizes their remaining value but also effectively reduces resource waste and environmental impact, making it a crucial path to achieving sustainable development in the new energy industry.

[0004] Lithium-ion battery sorting, which involves eliminating or clustering significantly different battery cells through difference analysis, is the foundation for comprehensive utilization. Faulty batteries pose safety hazards to retired battery packs, thus requiring an abnormal battery screening step before sorting. Existing comprehensive utilization frameworks simply involve random combination of batteries within the same cluster, resulting in disordered distribution of individual cells within the battery pack. This makes them susceptible to the "weakest link" effect, limiting performance and creating safety hazards. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a battery comprehensive utilization method and system based on multi-parameter collaborative optimization, which improves the accuracy of battery comprehensive utilization and enables comprehensive evaluation and reconstruction.

[0006] According to some embodiments, the present invention adopts the following technical solution:

[0007] A method for comprehensive battery utilization based on multi-parameter collaborative optimization includes the following steps:

[0008] Obtain the battery's capacity and internal resistance parameters, enhance them, and generate an enhanced dataset;

[0009] The augmented dataset is analyzed, and abnormal faulty batteries are removed based on the data distribution density. Healthy batteries are divided into highly consistent clusters by optimizing the clustering of the initial centers.

[0010] The battery module reconfiguration scheme is determined based on the connection method and parameter distribution pattern of the individual cells within the battery pack.

[0011] Based on the relationships obtained from the analysis and the divided battery clusters, analytical formulas for module capacity and internal resistance are established to quantitatively evaluate the performance of each battery module reconfiguration scheme in terms of total module capacity and equivalent internal resistance.

[0012] The dispersion of capacity and internal resistance among individual cells in a battery module under different connection methods and parameter distribution combinations is calculated. A consistency evaluation index is constructed, the suppression effect of each battery module reconfiguration scheme on the barrel effect is analyzed, and the optimal battery module reconfiguration scheme is determined.

[0013] As an alternative implementation method, the process of obtaining the battery's capacity and internal resistance parameters and generating enhanced data using a two-dimensional Weibull distribution includes: obtaining the static capacity and DC internal resistance of retired lithium-ion batteries as basic feature parameters, introducing a two-dimensional Weibull distribution model to fit the joint distribution of the battery's capacity and internal resistance, and generating an enhanced dataset that conforms to the distribution characteristics of the original data through Monte Carlo simulation.

[0014] As an alternative implementation method, the process of eliminating abnormal faulty batteries based on data distribution density includes: using a density-based outlier detection algorithm to identify and eliminate faulty batteries with abnormal characteristics such as micro-short circuits, lithium plating, or capacity jumps.

[0015] As an alternative implementation, the process of dividing healthy batteries into highly consistent clusters by optimizing the initial center clustering includes: applying an initial center optimized clustering algorithm to divide healthy batteries with highly consistent performance into several clusters based on the multidimensional characteristics of battery capacity and internal resistance.

[0016] As a further defined implementation, the process of dividing healthy batteries with highly consistent performance into several clusters includes: randomly selecting a point as the cluster center; calculating the shortest distance between each sample point in the dataset and the current cluster center; calculating the probability that each sample will be selected as the next cluster center; using a roulette wheel method to select the next cluster center; until a predetermined number K cluster centers are selected; and calculating the distance from each sample point to the current cluster center. K The distance between each cluster center is calculated, and the nearest cluster is assigned to each cluster. The average value of all points within each cluster is then recalculated as the new cluster center, and this process continues until the cluster center is no longer changed.

[0017] As an alternative implementation method, in the process of determining the battery module reconfiguration scheme based on the connection method and parameter distribution pattern of the individual cells in the battery pack, the connection method of the individual cells in the battery pack includes two methods: parallel connection followed by series connection and series connection followed by parallel connection. The parameter distribution pattern includes various arrangement methods based on different sorting methods according to capacity or internal resistance.

[0018] As an alternative implementation method, the process of determining the battery module reconfiguration scheme based on the connection method and parameter distribution pattern of the individual cells within the battery pack specifically includes: reconfiguring the basic components from... n A module consisting of 10 units connected in series is defined as OnS, and its basic components are... m A module consisting of units connected in parallel is defined as OmP;

[0019] Setting the battery pack n × m The batteries are arranged in ascending order of capacity or resistance. As the battery index increases, its capacity or resistance increases accordingly. The distribution pattern of the individual cells is divided into two categories based on the direction of increase of battery parameters. The first category is that the battery parameters increase horizontally from left to right in the inter-group OmP and OnS modules, showing an H distribution. The second category is that the battery parameters increase in an S-shaped pattern in the inter-group OmP and OnS modules, showing an S distribution. Based on the two distribution patterns and the connection patterns of mPnS and nSmP, four battery reconfiguration schemes are obtained, which are denoted as H-mPnS, H-nSmP, S-mPnS, and S-nSmP, respectively.

[0020] The capacity and resistance of the battery pack are calculated as follows:

[0021] The capacity and resistance of the OmP module are as follows:

[0022] ;

[0023] in, C OmP,i and R OmP,i of i Representing the first i The capacity and resistance of each OmP module, C i,j and R i,j They represent the first and second modules respectively. j The capacity and resistance of each unit;

[0024] The resistance and capacitance of an OnS module are expressed as follows:

[0025] ;

[0026] in, C OnS,j and R OnS,j Representing the first j The capacitance and resistance of each OnS module;

[0027] mPnS battery pack is composed of n It consists of several OmP modules connected in series, and its capacity and resistance are denoted as follows: C mPnS and R mPnS :

[0028] ;

[0029] Depend on m The capacity and resistance of an nSmP battery pack composed of several OnS modules connected in parallel are expressed as follows:C nSmP and R nSmP :

[0030] ;

[0031] The capacity and resistance of a battery pack's connection mode can be obtained from the parameters of its battery cells.

[0032] As an alternative implementation method, the process of establishing analytical formulas for module capacity and internal resistance based on the analyzed relationships and the divided battery clusters includes: For the H-mPnS reconfiguration scheme, taking the capacity-increasing order (i.e., the capacity of a single cell increases with its label) as an example, the battery pack capacity is expressed as:

[0033] ;

[0034] In the formula, Q1, Q2…Q nm These are the capacities of the corresponding labeled battery cells in a module with n×m battery cells;

[0035] Q H-mPnS The capacity of the battery pack reconstructed using the H-mPnS scheme;

[0036] Q H-nSmP The capacity of the battery pack reconstructed using the H-nSmP scheme;

[0037] Q S-mPnS The capacity of the battery pack reconstructed using the S-mPnS scheme;

[0038] Q S-nSm P represents the capacity of the battery pack reconstructed using the S-nSmP scheme.

[0039] For the H-nSmP reconfiguration scheme, the battery pack capacity is expressed as:

[0040] ;

[0041] For the S-mPnS reconfiguration scheme, the battery pack capacity is expressed as:

[0042] ;

[0043] For the S-nSmP reconfiguration scheme, the battery pack capacity is expressed as:

[0044] ;

[0045] Assuming that battery capacity increases with battery index, the capacity ranking of various reconfiguration schemes is as follows: Q S-mPnS > Q S-nSmP> Q H-mPnS = Q H-nSmP .

[0046] As an alternative implementation method, the process of calculating the dispersion of capacity and internal resistance among individual cells within a battery module under different connection methods and parameter distribution combinations, and constructing a consistency evaluation index, includes: combining the dispersion coefficient and the cumulative distribution function to conduct a comprehensive evaluation of the consistency of battery capacity and internal resistance; for each reconfiguration scheme, calculating the capacity variation coefficient based on the capacity sequence of the battery module; calculating the dispersion coefficient values ​​of various reconfiguration schemes based on the sequential distribution of the capacity of the individual cells constituting each series-parallel module; and sorting the dispersion coefficient values.

[0047] As an alternative implementation method, the process of analyzing the suppression effect of each battery module reconfiguration scheme on the barrel effect and determining the optimal battery module reconfiguration scheme includes: screening the preferred arrangement that maximizes module capacity and minimizes internal resistance, and comprehensively considering the ranking of the discrete coefficient values, selecting the arrangement with a discrete coefficient value less than a set value.

[0048] A battery comprehensive utilization system based on multi-parameter collaborative optimization includes:

[0049] The battery data testing module is configured to acquire and enhance the battery's capacity and internal resistance parameters, generating an enhanced dataset.

[0050] The two-stage sorting module is configured to analyze the augmented dataset, remove abnormal faulty batteries based on data distribution density, and divide healthy batteries into highly consistent clusters by optimizing the clustering of the initial centers.

[0051] The reconfiguration scheme design module is configured to determine the battery module reconfiguration scheme based on the connection method and parameter distribution pattern of the individual cells in the battery pack.

[0052] The performance quantification analysis module is configured to establish analytical formulas for module capacity and internal resistance based on the relationships obtained from the analysis and the divided battery clusters, so as to quantitatively evaluate the performance of each battery module reconfiguration scheme in terms of total module capacity and equivalent internal resistance.

[0053] The multi-index comprehensive evaluation module is configured to calculate the dispersion of capacity and internal resistance among individual cells in the battery module pack under different connection methods and parameter distribution combinations, construct consistency evaluation indicators, analyze the suppression effect of each battery module reconfiguration scheme on the barrel effect, and determine the optimal battery module reconfiguration scheme by combining the battery pack capacity and equivalent internal resistance value.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0055] This invention constructs a two-stage sorting system. First, abnormal and faulty batteries are removed. Then, healthy batteries are divided into highly consistent clusters by optimizing the initial center clustering. This accurately obtains multiple groups of retired batteries with excellent stability, providing performance and safety guarantees for module reconfiguration. For batteries within the same cluster after sorting, the electrical connection method is incorporated into the optimization, proposing multiple topologies such as parallel-to-series and series-to-parallel. Considering the distribution of individual cells within the group, H-type and S-type distributions are proposed, combining to form various reconfiguration topologies. Analytical calculation formulas for the total module capacity and equivalent internal resistance under each topology are established. Numerical simulations are used to compare different topologies and parameter distributions. This invention addresses the performance differences under different battery configurations, selecting the optimal arrangement that maximizes module capacity and minimizes internal resistance. Simultaneously, it calculates dispersion indices such as the coefficient of variation of capacity and internal resistance for each scheme, constructs a comprehensive consistency evaluation coefficient, and compares the suppression effect of different reconstruction methods on the "weakest link" effect. A comprehensive parameter and consistency evaluation system is established to provide a decision-making basis for selecting the optimal reconstruction scheme. This invention solves the technical problems existing in the comprehensive utilization methods of retired batteries, such as insufficient removal of faulty battery interference, sensitivity of clustering algorithms to initial centers, and low capacity utilization and high safety risks due to neglecting the connection methods and parameter distribution patterns within the battery pack.

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0057] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0058] Figure 1 A flowchart illustrating an implementation method of one embodiment;

[0059] Figure 2 A structural diagram of a module reconfiguration solution for one embodiment;

[0060] Figure 3 This is an example of a classification and evaluation system for the comprehensive utilization of retired batteries. Detailed Implementation

[0061] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0062] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0063] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0064] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0065] Example 1

[0066] A method for comprehensive battery utilization based on multi-parameter collaborative optimization includes the following steps:

[0067] Step S1: Collect capacity and internal resistance parameters, and use a two-dimensional Weibull distribution to generate augmented data to expand the sample space and improve the model's generalization ability;

[0068] The capacity and resistance parameters of retired batteries conform to a Weibull distribution. Therefore, this invention uses a two-dimensional Weibull distribution to construct multiple samples of retired lithium-ion batteries. This distribution clearly describes the nonlinearity and correlation of the battery capacity and resistance distributions, providing a more reliable statistical basis for the comprehensive utilization of retired batteries. The following is the construction of the capacity sequence ( X ) and resistance sequence ( Y The specific process of the two-dimensional Weibull distribution.

[0069] First, let's denote the capacitance and resistance as... x and y The joint survival function of the retired batteries is:

[0070] (1)

[0071] in η C , β C , η R and β R These are the shape and scale parameters of the Weibull distributions for capacitance and resistance, respectively. Related parameters θ The correlation between them is described. When θ= At 0 o'clock, X and Y They are mutually independent and conform to a one-dimensional Weibull distribution; when θ> At 0 o'clock, X and Y Positive correlation; when θ< At 0 o'clock,X and Y Negative correlation. Then S(x,y) right y The partial derivatives are:

[0072] (2)

[0073] The joint probability density function shown below can be obtained. f X,Y ( x , y This function represents the two-dimensional Weibull distribution of the capacity and resistance of retired lithium-ion batteries:

[0074] (3)

[0075] Step S2: Construct a two-stage sorting system. First, remove abnormal and faulty batteries, and then divide healthy batteries into highly consistent clusters by optimizing the clustering of the initial centers, providing a basis for reconstruction.

[0076] In the first-level sorting stage, sufficiently dense areas are divided into clusters to effectively identify and distinguish feature data from noise. It has two initial parameters: neighborhood radius. Eps The minimum number of points that must exist in the neighborhood. Minpts These are used to describe the density of sample distribution in the neighborhood. The algorithm can cluster dense data points and remove outliers that do not belong to any cluster.

[0077] The following definition is required: Assume the dataset W There are two data points, p and q .if p and q The distance between them is less than s, then data points q It is a point p The neighborhood of is called the direct density reachable. This is shown in the equation below:

[0078] (4)

[0079] in d ( p , q )express p and q The distance between them is chosen to be the Euclidean distance in this paper. Furthermore, Indicates the sample at the data point p The neighborhood on is shown by the following equation:

[0080] (5)

[0081] If data points p The number of neighborhood data points is greater than the density threshold Then that point p This is called the core point, meaning it exists:

[0082] (6)

[0083] In addition, if multiple points exist , ..., ,in , ,and It can be directly from If the density is reached, then q From p Achievable density. Clustering of density clustering. C It consists of a core point and its density-reachable relational samples. After all retired cells have passed through, points that do not belong to any cluster are considered anomalous cells.

[0084] The second-stage clustering method uses a probabilistic approach to select initial centroids, ensuring a uniform distribution of initial cluster centers. If the dataset contains n samples... B ={ , ,..., } ,b i = (x i ,y i ) represents the i-th battery sample, divided into K For each cluster, the algorithm steps can be described in detail.

[0085] Randomly select a point as B Cluster center in Repeat the following steps.

[0086] First, calculate the relationship between each sample point in the dataset and the current cluster center. The shortest distance between:

[0087] (7)

[0088] Next, calculate the probability that each sample will be selected as the next cluster center:

[0089] (8)

[0090] Finally, the roulette wheel method is used to select the next cluster center, until the first cluster center is selected. K Repeat the steps for each cluster center. Calculate the results for each sample point. arrive KThe distance between each cluster center is used to divide the cluster into the nearest clusters. Then, the average value of all points within each cluster is recalculated as the new cluster center, and this process continues until the cluster centers no longer need to be changed.

[0091] By overcoming the sensitivity of randomly initialized centroids, the proposed sorting method can further improve the clustering accuracy and convergence efficiency of retired battery sorting. Therefore, based on the proposed two-stage sorting, multiple stable retired battery clusters can be accurately obtained, providing performance and safety guarantees for reconstruction.

[0092] Step S3: Incorporate the connection method and parameter distribution pattern of individual cells within the battery pack into the module reconfiguration optimization, and propose multiple comprehensive utilization schemes.

[0093] This invention conducts an in-depth study on the impact of battery capacity and resistance distribution on battery pack performance. For those with... n × m The two most common connection modes for battery packs are "parallel connection followed by series connection" and "series connection followed by parallel connection". The basic components of an OnS (On-Side System) consist of... n OmP is a module composed of interconnected units. m A module consisting of units connected in parallel.

[0094] Secondly, setting the battery pack n × m Batteries are arranged in ascending order of capacity or resistance. As the battery index increases, its capacity or resistance increases accordingly. The individual cell distribution patterns are divided into two categories based on the direction of parameter increase. The first category is where battery parameters increase horizontally from left to right within the OmP and OnS modules, hence the name "H" distribution. The second category is where battery parameters increase in an S-shaped pattern within the OmP and OnS modules, hence the name "S" distribution. These two distribution patterns, combined with the mPnS and nSmP connection patterns, yield four battery reconfiguration schemes, denoted as H-mPnS, H-nSmP, S-mPnS, and S-nSmP, respectively.

[0095] The capacity and resistance of the battery pack can be calculated as follows.

[0096] First, all cells in the OmP module share the same terminal voltage, and the module's current is the sum of the currents of all cells. Therefore, the capacitance and resistance of the OmP module are as follows:

[0097] (9)

[0098] in, C OmP,i and R OmP,i of i Representing the first iThe capacity and resistance of each OmP module. C i,j and R i,j They represent the first and second modules respectively. j The capacity and resistance of each unit.

[0099] For an OnS module, all its cells have the same current. Therefore, the capacity of an OnS module is determined by its lowest capacity cell, and the OnS module will stop working when the charge / discharge cutoff voltage is reached. The resistance of an OnS series module is the sum of the resistances of all its cells. Therefore, the resistance and capacity of an OnS module can be expressed as:

[0100] (10)

[0101] in C OnS,j and R OnS,j Representing the first j The capacitance and resistance of each OnS module.

[0102] mPnS battery pack is composed of n It consists of several OmP modules connected in series. Based on the above description, its capacity and resistance are denoted as follows: C mPnS and R mPnS :

[0103] (11)

[0104] Similarly, by m The capacity and resistance of an nSmP battery pack composed of several OnS modules connected in parallel are expressed as follows: C nSmP and R nSmP :

[0105] (12)

[0106] In summary, the capacity and resistance of a battery pack with two common connection modes can be obtained from the parameters of its battery cells. These parameters are also used as performance parameters to evaluate sorting and reconfiguration methods for retired batteries.

[0107] Step S4: Based on the topology designed in Step S3 and the battery clusters sorted in Step S2, establish analytical formulas for module capacity and internal resistance, and quantitatively evaluate the performance of each scheme in terms of total module capacity and equivalent internal resistance.

[0108] The theoretical analysis of the impact of different reconfiguration schemes on capacity and resistance distribution is shown below. First, the retired lithium-ion battery pack based on H-mPnS consists of adjacent parallel modules connected in series. The capacity of a parallel module is approximately the sum of the capacities of its individual cells. The capacity of the battery pack is determined by the smallest parallel module. If the capacity increases with increasing cell serial number, the following equation can be derived:

[0109] (13)

[0110] For battery packs using the H-nSmP reconfiguration scheme, the capacity of the basic series modules depends on the smallest cell. The capacity of the battery pack can be approximated as the sum of the capacities of the series-connected modules, i.e.:

[0111] (14)

[0112] Similarly, the capacity of S-mPnS-based battery packs is the smallest among parallel modules. However, since the internal cells are distributed in a cross-capacity manner, the capacities of their parallel modules are similar, and therefore their capacity can be approximated by the following formula:

[0113] (15)

[0114] For the S-nSmP reconfiguration scheme, the battery pack capacity can be expressed as:

[0115] (16)

[0116] Assuming that battery capacity increases with battery index, the capacity ranking of various reconfiguration schemes is as follows: Q S-mPnS > Q S-nSmP > Q H-mPnS = Q H-nSmP Unlike capacity, the internal resistance of a battery pack is not determined by the worst-performing individual cell, but rather by the combined internal resistance of all individual cells within the pack. Therefore, after sorting, the differences in internal resistance among the individual cells are small, and the impact of the reconfiguration scheme on the overall battery pack's internal resistance is negligible, ultimately resulting in similar internal resistance parameters.

[0117] In summary, when sorted by increasing capacity, the S-mPnS reconfiguration scheme can significantly improve the capacity of each retired battery pack while maintaining relatively low internal resistance. This fully demonstrates its superiority and robustness in improving battery pack performance.

[0118] Unlike the capacity-based sorting distribution scheme, when sorting by increasing single-cell internal resistance, the capacities of H-mPnS and S-mPnS-based battery packs are relatively similar and high across all reconfiguration scheme groups. However, the capacity distribution of the S-mPnS reconfiguration scheme is still the largest. The reasons for this result are as follows:

[0119] The capacity distribution of retired batteries is negatively correlated with their resistance distribution, with a correlation coefficient of only -0.6795. This indicates that, in battery packs sorted by increasing resistance, the battery capacity generally decreases with increasing battery index. However, some batteries deviate from this trend, with relatively large or small capacities. Therefore, referring to formulas (13)-(16), the following analysis can be performed. First, the battery pack capacity based on mPnS can be expressed as: n The minimum capacity of parallel modules, where the capacity of each parallel module is m The sum of the capacities of individual cells, arranged in an H-shape or S-shape. For an nSmP-based battery pack, its capacity is... m The capacity of a battery pack using the mPnS connection method is the sum of the minimum capacities of the cells in each series module. Therefore, the capacity of a battery pack using the mPnS connection method will be significantly larger than that of a battery pack based on the nSmP connection method. Furthermore, the cells in the parallel modules based on the H-mPnS grouping scheme are numbered consecutively, while the cell numbers in the parallel modules based on the S-mPnS grouping scheme are discretely distributed. Therefore, since capacity and resistance are negatively correlated, and resistance increases with cell number, the parallel module capacity of the H-mPnS battery pack is smaller, resulting in a smaller total battery pack capacity. In summary, the battery pack based on the S-mPnS arrangement has superior capacity performance. The resistance of the battery packs using the four reconfiguration schemes is roughly the same, with the resistance of the S-mPnS scheme being slightly lower than the other three schemes.

[0120] In summary, compared with the H-mPnS, H-nSmP, and S-nSmP schemes, the S-mPnS reconfiguration scheme effectively improves the battery pack capacity while reducing resistance, thus achieving optimized battery pack parameter performance. The stable performance of the S-mPnS scheme in different clusters further demonstrates its superiority and reliability in tiered utilization. Furthermore, battery packs sorted by capacity exhibit higher capacity and better performance.

[0121] Step S5: Calculate the dispersion of capacity and internal resistance between individual units within the module under different combinations of topology and parameter distribution, construct a consistency evaluation index, analyze the suppression effect of each scheme on the "barrel effect", and provide a consistency basis for selecting the optimal reconstruction scheme.

[0122] The "barrel effect" refers to the phenomenon where poor consistency means the worst cell stops working early, leaving other cells with a lot of charge, causing the worst cell to drag down the overall performance. Conversely, when consistency is good, when the worst cell stops working, other cells have very little charge left, allowing each cell in the battery module to be used more efficiently.

[0123] The suppression of the "weakest link" effect essentially means that the consistency of individual cells within the module improves after sorting and reconfiguration. Two-stage sorting groups battery cells with similar performance, such as capacity and internal resistance, into the same category, initially improving the consistency of the cells that make up the module. Furthermore, for the different proposed reconfiguration schemes, the consistency within the group depends on the module's performance. For example, an mPnS battery pack is composed of n OmP modules connected in series, and its pack consistency depends on the consistency of those n OmP modules; the same applies to an nSmP battery pack. Therefore, different reconfiguration schemes have different levels of consistency, and the one with optimal consistency is considered to have the best effect on suppressing the "weakest link" effect.

[0124] By combining the coefficient of variation (CV) and the cumulative distribution function (CDF), a comprehensive evaluation method for the consistency of battery capacity and internal resistance is constructed.

[0125] Here, CV is the ratio of standard deviation to mean, with the mean measuring the dispersion of the data. A smaller CV indicates less data dispersion, thus indicating higher dataset consistency. The specific equation is as follows:

[0126] (17)

[0127] To further quantitatively assess the impact of different reconfiguration schemes on the consistency of retired battery packs, the coefficient of variation and cumulative distribution function were used. The cumulative distribution function describes the probability that a random variable is less than or equal to a specific value, and its definition is as follows:

[0128] (18)

[0129] in, x For a specific coefficient of variation value, P ( X ≤ x ) represents a random variable X Values ​​less than or equal to x The probability of.

[0130] This embodiment uses the CDF function to describe the distribution of the CV coefficients.

[0131] The capacity coefficient of variation (CV) is calculated from the capacity sequence of the modules. Based on the sequential distribution of the individual battery cell capacities constituting each series-parallel module, the CV calculation formulas for the four reconfiguration schemes can be obtained, as shown below:

[0132] First, the capacity CV of the battery pack using mPnS technology is calculated using the capacity sequence of the parallel modules. Furthermore, the capacity of the battery cells constituting each module is sequentially allocated in the H-mPnS recombination, resulting in:

[0133] (19)

[0134] in It is the first i The average capacity of the batteries in the parallel modules, and < < .

[0135] Furthermore, in the S-mPnS reconfiguration scheme, internal units are distributed according to capacity, making the capacity of each parallel module very similar. The CV calculation method is as follows:

[0136] (20)

[0137] in, It is the first i The average capacity of the batteries in the parallel modules. Due to the overlapping battery numbers in the S-shaped distribution, there is... Therefore, the coefficient of variation of the S-mPnS reconstruction scheme is very small, close to 0.

[0138] C VH-mPnS CV of the battery pack reconstructed using the H-mPnS scheme;

[0139] C VH-nSmP CV of the battery pack reconstructed using the H-nSmP scheme;

[0140] C VS-mPnS CV of the battery pack reconstructed using the S-mPnS scheme;

[0141] C VS-nSmP CV of the battery pack reconstructed using the S-nSmP scheme;

[0142] Based on the H-nSmP and S-nSmP reconstruction schemes, the capacity CV of the decommissioned LIB packet is obtained from the capacity sequence of each cascaded module:

[0143] (twenty one)

[0144] (twenty two)

[0145] As the cell index increases, the battery capacity also increases. Therefore, in the capacity sequence used to calculate the above CV values, the greater the difference in cell index, the wider the capacity distribution, resulting in a larger CV value, i.e., a worse consistency. In summary, based on formulas (19)-(22), the CV values ​​of various reconstruction schemes are ranked as follows: CV S-mPnS <CV H-nSmP <CV H-mPnS ≈ CV S-nSm The battery pack using the S-mPnS scheme has a very low capacity CV value. Furthermore, the resistance of each battery pack depends on multiple highly consistent cells. A smaller CV value indicates that the capacity / internal resistance values ​​of the modules within the pack are closer, meaning better consistency and better suppression of the "weakest link" effect.

[0146] This embodiment comprehensively evaluates the battery pack's capacity / internal resistance values ​​and capacity / internal resistance consistency under different reconstruction schemes, and the evaluation results are accurate.

[0147] The consistency analysis when sorted by resistance is similar to the analysis above. Under the S-mPnS reconfiguration scheme, the capacity and resistance (CV) distribution range of all clusters is minimized, achieving optimal consistency. Therefore, S-mPnS exhibits a significant advantage in improving battery pack consistency, outperforming other reconfiguration schemes. It can effectively reduce performance differences within reconfigured lithium-ion battery packs, thereby improving their reliability and stability.

[0148] In summary, for all the original data generated by the two-dimensional Weibull distribution, after the first step of sorting, and the second step of sorting (i.e., clustering to divide the remaining battery samples into several classes, where the capacity and internal resistance of the batteries in each class are relatively similar), the next step is reconstruction. The idea is to randomly select m×n batteries from each of the clustered classes to reconstruct them into a group. Then, the classes with g batteries can form g / (m×n) modules. Each module is reconstructed according to four reconstruction schemes, and then the superior reconstruction scheme is determined by the magnitude and consistency (CV coefficient) of the capacity and internal resistance distribution of the g / (m×n) modules (to avoid the randomness of a single module).

[0149] In summary, this invention achieves efficient and safe module reconstruction of retired batteries through a complete technical chain encompassing data augmentation, clustering sorting, reconstruction optimization, and multi-index evaluation. First, addressing the limited sample size of retired batteries, a two-dimensional Weibull distribution of capacity and internal resistance is introduced to generate simulated data, effectively expanding the sample space and enhancing the method's generalization ability. Subsequently, a two-stage sorting strategy is employed: the first stage removes abnormal and faulty batteries, while the second stage precisely sorts the remaining normal batteries, overcoming the sensitivity of traditional clustering to initial centers and significantly improving battery cell consistency. Building upon this, the connection method and parameter distribution pattern within the battery pack are incorporated into module reconstruction optimization for the first time. Four reconstruction schemes are proposed, and parameter indices for module capacity and internal resistance under different connection topologies are established to quantitatively analyze module performance. Finally, a multi-index evaluation system covering module capacity, internal resistance, and consistency is constructed to comprehensively assess the reconstruction effect.

[0150] Example 2

[0151] As a typical application example, such as Figure 1 As shown, the classification method for the comprehensive utilization of retired batteries in this embodiment includes the following:

[0152] Sixty actual retired lithium-ion battery cells (e.g., NCM lithium batteries with a rated capacity of 2.5 Ah) were selected as the basic sample. Standardized charge-discharge tests were performed on each battery under a constant temperature environment of 25°C, and their static capacity and DC internal resistance were collected as basic characteristic parameters to form the original dataset.

[0153] To overcome the limitation of the limited number of actual retired battery samples, a two-dimensional Weibull distribution is used to fit the joint distribution of capacity and internal resistance. First, the scale and shape parameters of capacity and internal resistance in the original data are estimated to establish a two-dimensional Weibull distribution model. Then, 50,000 enhanced battery samples that conform to the distribution characteristics of the original data are randomly sampled from this distribution model through Monte Carlo simulation. This enhanced dataset fully preserves the statistical distribution characteristics of the original batteries, effectively expands the sample space, provides a more complete data foundation for subsequent sorting models, and significantly improves the generalization ability and robustness of the method.

[0154] Two-stage sorting was performed on the enhanced 50,000 battery samples:

[0155] The first stage employs a density-based outlier detection method to identify and eliminate faulty batteries exhibiting abnormal characteristics such as micro-short circuits, lithium plating, or capacity jumps. (Parameters...) Eps and MinptsThe values ​​were set to 0.25 and 6 respectively. By calculating the local density of each sample in the capacity-internal resistance two-dimensional space, samples with significantly lower density than their neighbors were identified as abnormal cells and removed. A total of about 67 abnormal cells were identified and removed, effectively preventing them from interfering with subsequent clustering.

[0156] The second stage employs an initial center optimization clustering algorithm to finely sort the remaining approximately 49,933 healthy batteries based on multidimensional characteristics of capacity and internal resistance. This algorithm optimizes the selection strategy of initial cluster centers, ensuring that initial center points are distributed as widely as possible within high-density areas, overcoming the sensitivity of traditional clustering to initial centers. It automatically divides the batteries into five highly consistent clusters, named High Capacity Low Resistance 1 (HC-LR1), High Capacity Low Resistance 2 (HC-LR2), High Capacity High Resistance (HC-HR), Low Capacity Low Resistance (LC-LR), and Low Capacity High Resistance (LC-HR). All cells in each cluster are grouped into sets of 40 (m=5, n=8) cells, forming battery packs of 260, 400, 243, 197, and 146 cells respectively. Testing showed that after sorting, the average coefficient of variation (CV) of batteries within the same cluster decreased to 52% of its pre-sorting value, and the consistency of battery cells nearly doubled, providing a high-quality battery cell foundation for subsequent reconstruction.

[0157] For batteries within the same cluster after sorting, this study, for the first time, incorporates electrical connection methods into the optimization scope, proposing several optional intra-cluster topologies, including parallel-then-series (mPnS), series-then-parallel (nSmP), and hybrid connection schemes. Furthermore, considering H-type and S-type cell distribution patterns, four cell reconfiguration schemes are proposed: H-mPnS, H-nSmP, S-mPnS, and S-nSmP, such as... Figure 2 As shown, analytical formulas for calculating the total module capacity, equivalent internal resistance, and consistency are established for each topology. Through these analytical models, the impact of different connection methods on the module's electrical performance is quantified, providing a topological basis for subsequent parameter distribution analysis.

[0158] Based on the topology designed in step S3 and the battery clusters sorted in step S2, the system analyzes the performance variation of module capacity and internal resistance under different arrangement modes. Four parameter distribution schemes are proposed: H-type and S-type. Combining two topologies—parallel-to-series and series-to-parallel—the total capacity and equivalent internal resistance of modules under different reconfiguration modes (H-mPnS, H-nSmP, S-mPnS, S-nSmP) are calculated.

[0159] In particular, a special arrangement with an S-shaped increasing capacity was simulated, and the performance differences of each scheme were compared through numerical simulation. The results show that, under the same battery cluster, a topology combining parallel and then series connections with an S-shaped increasing capacity distribution maximizes the total module capacity while maintaining a low equivalent internal resistance. The influence of parameter distribution on the module's output capability is quantitatively revealed. For different combinations of topologies and parameter distributions, the coefficients of variation and other dispersion indices of capacity and internal resistance among individual cells within the module were calculated, constructing a comprehensive consistency evaluation coefficient covering multiple parameters. By comparing the dispersion of individual cell parameters under each scheme, the effectiveness of different reconfiguration methods in suppressing the "weakest link" effect is analyzed in depth.

[0160] Four reconfiguration schemes were set up: H-nSmP, H-mPnS, S-nSmP, and S-mPnS. The average capacity, average internal resistance, and consistency coefficient of the module under each scheme were calculated. Experimental results show that:

[0161] 1) The S-mPnS scheme has the highest average capacity, which is 5.084% higher than the traditional random sorting schemes (R-nSmP and R-mPnS), and the capacity utilization rate is significantly improved.

[0162] 2) The capacity variation coefficient of the S-mPnS scheme is only 0.85%, which is much lower than the 6% of the traditional scheme. The consistency is improved by more than 5 times, which fully verifies the significant effect of the method of the present invention in suppressing the "barrel effect".

[0163] 3) All schemes using the two-stage sorting method of this invention are superior to traditional random sorting schemes in terms of capacity and consistency, proving the necessity and effectiveness of the sorting step.

[0164] Example 3

[0165] An efficient classification and evaluation system for the comprehensive utilization of retired batteries, such as Figure 3 As shown, it includes:

[0166] Battery data testing module: Used to conduct charge and discharge tests on batteries to obtain data such as voltage and current, record the battery's discharge capacity, calculate the battery's internal resistance based on known data, and generate a large number of retired batteries based on the battery charge and discharge test data and the two-dimensional Weibull distribution algorithm.

[0167] Two-stage sorting module: Performs two-stage sorting on the enhanced dataset: The first stage uses a density-based outlier detection algorithm to identify and remove faulty batteries with abnormal features such as micro-short circuits, lithium plating, or capacity jumps; The second stage applies an initial center-optimized clustering algorithm to automatically divide healthy batteries with highly consistent performance into several clusters based on the multidimensional features of capacity and internal resistance, providing highly consistent battery cells for subsequent reconstruction.

[0168] The comprehensive performance evaluation system module optimizes the electrical connection method for batteries within the same cluster after sorting. It proposes various topologies, including parallel-to-series, series-to-parallel, and hybrid connections, and establishes analytical formulas for calculating the total module capacity and equivalent internal resistance under each topology. Through numerical simulation, it compares the performance differences under different combinations of topologies and parameter distributions, selecting the optimal arrangement that maximizes module capacity and minimizes internal resistance. Simultaneously, it calculates the coefficient of variation and other dispersion indicators for capacity and internal resistance under each scheme, constructs a comprehensive consistency evaluation coefficient, and compares the suppression effect of different reconfiguration methods on the "weakest link" effect, providing a decision-making basis for selecting the optimal reconfiguration scheme.

[0169] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of one or more computer-usable storage media (including, but not limited to, disk storage, etc.) containing computer-usable program code. CD - ROM It takes the form of a computer program product implemented on (such as optical memory, etc.).

[0170] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0171] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0172] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0173] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for comprehensive battery utilization based on multi-parameter collaborative optimization, characterized in that, Includes the following steps: Obtain the battery's capacity and internal resistance parameters, enhance them, and generate an enhanced dataset; The augmented dataset is analyzed, and abnormal faulty batteries are removed based on the data distribution density. Healthy batteries are divided into highly consistent clusters by optimizing the clustering of the initial centers. Based on the connection method and parameter distribution pattern of the individual cells within the battery pack, a battery module reconfiguration scheme is determined. The specific process of the battery module reconfiguration scheme includes: converting the basic components from... n A module consisting of 10 units connected in series is defined as OnS, and its basic components are... m A module consisting of units connected in parallel is defined as OmP; Setting the battery pack n × m The batteries are arranged in ascending order of capacity or resistance. As the battery index increases, its capacity or resistance increases accordingly. The distribution pattern of the individual cells is divided into two categories based on the direction of increase of battery parameters. The first category is that the battery parameters increase horizontally from left to right in the inter-group OmP and OnS modules, showing an H distribution. The second category is that the battery parameters increase in an S-shaped pattern in the inter-group OmP and OnS modules, showing an S distribution. Based on the two distribution patterns and the connection patterns of mPnS and nSmP, four battery reconfiguration schemes are obtained, which are denoted as H-mPnS, H-nSmP, S-mPnS, and S-nSmP, respectively. The capacity and resistance of the battery pack are calculated as follows: The capacity and resistance of the OmP module are as follows: ； in, C OmP,i and R OmP,i of i Representing the first i The capacity and resistance of each OmP module, C i,j and R i,j These represent the module's first... j The capacity and resistance of each unit; The resistance and capacitance of an OnS module are expressed as follows: ； in, C OnS,j and R OnS,j Representing the first j The capacity and resistance of each OnS module; mPnS battery pack is composed of n It consists of several OmP modules connected in series, and its capacity and resistance are denoted as follows: C mPnS and R mPnS : ； Depend on m The capacity and resistance of an nSmP battery pack composed of several OnS modules connected in parallel are expressed as follows: C nSmP and R nSmP : ； The capacity and resistance of a battery pack's connection mode can be obtained from the parameters of its battery cells; Based on the relationships obtained from the analysis and the divided battery clusters, analytical formulas for module capacity and internal resistance are established to quantitatively evaluate the performance of each battery module reconfiguration scheme in terms of total module capacity and equivalent internal resistance. The analytical formula process includes: For the H-mPnS reconfiguration scheme, the battery pack capacity is expressed as: ； In the formula, Q1, Q2…Q nm These are the capacities of the corresponding labeled battery cells in a module with n×m battery cells; Q H-mPnS The capacity of the battery pack reconstructed using the H-mPnS scheme; Q H-nSmP The capacity of the battery pack reconstructed using the H-nSmP scheme; Q S-mPnS The capacity of the battery pack reconstructed using the S-mPnS scheme; Q S-nSm P represents the capacity of the battery pack reconstructed using the S-nSmP scheme. For the H-nSmP reconfiguration scheme, the battery pack capacity is expressed as: ； For the S-mPnS reconfiguration scheme, the battery pack capacity is expressed as: ； For the S-nSmP reconfiguration scheme, the battery pack capacity is expressed as: ； Assuming that battery capacity increases with battery index, the capacity ranking of various reconfiguration schemes is as follows: Q S-mPnS > Q S-nSmP > Q H-mPnS = Q H-nSmP ; The dispersion of capacity and internal resistance among individual cells in a battery module under different connection methods and parameter distribution combinations is calculated. A consistency evaluation index is constructed, the suppression effect of each battery module reconfiguration scheme on the barrel effect is analyzed, and the optimal battery module reconfiguration scheme is determined.

2. The battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, The process of obtaining battery capacity and internal resistance parameters and generating augmented data using a two-dimensional Weibull distribution includes: obtaining the static capacity and DC internal resistance of retired lithium-ion batteries as basic feature parameters, introducing a two-dimensional Weibull distribution model to fit the joint distribution of battery capacity and internal resistance, and generating an augmented dataset that conforms to the distribution characteristics of the original data through Monte Carlo simulation.

3. The battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, Based on the data distribution density, the process of eliminating abnormal faulty batteries includes: using a density-based outlier detection algorithm to identify and eliminate faulty batteries with abnormal characteristics such as micro-short circuits, lithium plating, or capacity jumps.

4. The battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, The process of dividing healthy batteries into highly consistent clusters by optimizing the initial center clustering includes: applying a clustering algorithm optimized with the initial center, and dividing healthy batteries with highly consistent performance into several clusters based on the multidimensional characteristics of battery capacity and internal resistance. Specifically, this involves: randomly selecting a point as the cluster center; calculating the shortest distance between each sample point in the dataset and the current cluster center; calculating the probability of each sample being selected as the next cluster center; using a roulette wheel method to select the next cluster center until a predetermined number K cluster centers are selected; and calculating the distance from each sample point to the current cluster center. K The distance between each cluster center is calculated, and the nearest cluster is assigned to each cluster. The average value of all points within each cluster is then recalculated as the new cluster center, and this process continues until the cluster center is no longer changed.

5. The battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, Based on the connection method and parameter distribution pattern of the individual cells in the battery pack, the battery module reconfiguration scheme is determined. The connection method of the individual cells in the battery pack includes two methods: parallel connection followed by series connection and series connection followed by parallel connection. The parameter distribution pattern includes various arrangements based on different sorting methods according to capacity or internal resistance.

6. The battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, The process of calculating the dispersion of capacity and internal resistance among individual cells within a battery module under different connection methods and parameter distribution combinations, and constructing a consistency evaluation index, includes: combining the dispersion coefficient and the cumulative distribution function to conduct a comprehensive evaluation of the consistency of battery capacity and internal resistance; for each reconfiguration scheme, calculating the capacity variation coefficient based on the capacity sequence of the battery module; calculating the dispersion coefficient values ​​of various reconfiguration schemes based on the sequential distribution of the capacity of the individual cells constituting each series and parallel module; and sorting the dispersion coefficient values.

7. The battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, The process of analyzing the effect of each battery module reconfiguration scheme on suppressing the barrel effect and determining the optimal battery module reconfiguration scheme includes: screening the preferred arrangement that maximizes module capacity and minimizes internal resistance, and comprehensively considering the ranking of the discrete coefficient values, selecting the arrangement with discrete coefficient values ​​less than the set value.

8. A battery comprehensive utilization system based on multi-parameter collaborative optimization, wherein the system is used in the battery comprehensive utilization method based on multi-parameter collaborative optimization as described in claim 1, characterized in that, include: The battery data testing module is configured to acquire and enhance the battery's capacity and internal resistance parameters, generating an enhanced dataset. The two-stage sorting module is configured to analyze the augmented dataset, remove abnormal faulty batteries based on data distribution density, and divide healthy batteries into highly consistent clusters by optimizing the clustering of the initial centers. The reconfiguration scheme design module is configured to determine the battery module reconfiguration scheme based on the connection method and parameter distribution pattern of the individual cells in the battery pack. The performance quantification analysis module is configured to establish analytical formulas for module capacity and internal resistance based on the relationships obtained from the analysis and the divided battery clusters, so as to quantitatively evaluate the performance of each battery module reconfiguration scheme in terms of total module capacity and equivalent internal resistance. The multi-index comprehensive evaluation module is configured to calculate the dispersion of capacity and internal resistance among individual cells in the battery module under different connection methods and parameter distribution combinations, construct consistency evaluation index, analyze the suppression effect of each battery module reconstruction scheme on the barrel effect, and determine the optimal battery module reconstruction scheme.

Images