A production quality traceability method for new energy batteries with embedded RFID

Through the analysis and clustering of internal resistance and temperature data of lithium batteries, the problem of data correlation neglect in battery production quality traceability is solved, the traceability efficiency is improved, and the accurate identification and processing of lithium battery quality problems is achieved.

CN118863929BActive Publication Date: 2024-12-03LIAONING JUNENG XINCHI MATERIALS TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411345815.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-26
Publication Date
2024-12-03
Estimated Expiration
2044-09-26

AI Technical Summary

Technical Problem

The production quality traceability efficiency is not high due to ignoring the correlation between battery data during the battery production quality traceability process.

Method used

By obtaining the internal resistance data and temperature data of lithium batteries, using LOF outlier factor screening and clustering, the density center aggregation degree of clustering clusters is determined, the coefficient of variation between temperature correlation degree and density center aggregation degree is analyzed, and the abnormal consistency index is calculated. , cluster quality problem clusters and merge the differences between similar clusters to realize the traceability of production quality of embedded RFID lithium batteries.

Benefits of technology

Improve the efficiency of production quality traceability, distinguish between normal and abnormal internal resistance data, eliminate abnormal data affected by temperature, prevent redundancy of quality problem cluster division, and determine the final quality problem clustering results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118863929B_ABST
    Figure CN118863929B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of electrical performance testing, and proposes a method for tracing the production quality of new energy batteries with embedded RFID, including: obtaining the internal resistance data, temperature data and LOF outlier factor of the internal resistance data of all lithium batteries in the same batch as the lithium batteries with quality problems; screening and clustering the internal resistance data, obtaining the cluster clusters, obtaining the density center aggregation degree of the internal resistance data, determining the density concentration of the cluster clusters, determining the internal resistance data without quality problems and the suspected abnormal internal resistance data; obtaining the abnormal consistency index of the suspected abnormal internal resistance, and determining the first abnormal internal resistance data according to the abnormal consistency index; clustering the first abnormal internal resistance data, determining the quality problem cluster, obtaining the inter-cluster difference of the quality problem cluster, determining the final quality problem clustering result according to the inter-cluster difference, and completing the production quality tracing of lithium batteries with embedded RFID. The present invention aims to solve the problem of low efficiency of battery production quality tracing.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electrical performance testing, and in particular to a production quality tracing method for new energy batteries with embedded RFID. Background Art

[0002] RFID is a wireless communication technology used to wirelessly identify and track objects. By embedding RFID tags in batteries, radio frequency signals can be used to uniquely identify objects and transmit data, and important data in the production process can be recorded to facilitate battery quality traceability. Among them, important data in the production process includes production date, production process parameters, raw material information, test data, quality inspection results, factory batches, etc. When a problem is found in a battery, these data can be used to quickly and comprehensively trace batteries that may have the same problem.

[0003] At present, the quality monitoring of batteries is limited to testing the performance of a single battery. When quality problems are found in batteries, the batteries of the same batch need to be tested one by one in the quality tracing process to determine whether the quality problems are common to the same batch, and then the production links need to be traced to find and solve the problems in time. However, the workload of testing the batteries of the same batch one by one is large, which makes the efficiency of production quality tracing low. Summary of the invention

[0004] The present invention provides a method for tracing the production quality of new energy batteries with an embedded RFID to solve the problem of low efficiency of production quality tracing caused by ignoring the correlation between battery data during the battery production quality tracing process. The technical solution adopted is as follows:

[0005] An embodiment of the present invention provides a method for tracing the production quality of new energy batteries with embedded RFID, the method comprising the following steps:

[0006] According to the lithium batteries with quality problems detected during the sampling process, the internal resistance data, temperature data and LOF outlier factors of the internal resistance data of all lithium batteries in the same batch as the lithium batteries with quality problems are obtained;

[0007] The internal resistance data is screened and clustered according to the LOF outlier factor of the internal resistance data, the clustering cluster is obtained, the radius of the clustering cluster is determined, the density center aggregation degree of the internal resistance data is obtained according to the internal resistance data contained in the clustering cluster and the radius of the clustering cluster, the density concentration of the clustering cluster is determined, and the internal resistance data without quality problems is determined according to the density concentration, and the suspected abnormal internal resistance data is determined;

[0008] Obtaining neighborhood data of the suspected abnormal internal resistance, obtaining an internal resistance distance sequence and a temperature distance sequence of the suspected abnormal internal resistance, obtaining a temperature correlation degree of the suspected abnormal internal resistance, obtaining a coefficient of variation of a density center aggregation degree of the suspected abnormal internal resistance, obtaining a local abnormal consistency index of the suspected abnormal internal resistance according to the temperature correlation degree of the suspected abnormal internal resistance and the coefficient of variation of a density center aggregation degree of the suspected abnormal internal resistance, obtaining an abnormal consistency index of the suspected abnormal internal resistance according to the local abnormal consistency indexes of all suspected abnormal internal resistances, and determining first abnormal internal resistance data according to the abnormal consistency index;

[0009] Cluster the first abnormal internal resistance data to determine the quality problem cluster, determine the adjacent quality problem cluster based on the quality problem cluster, determine the evaluation data set based on the first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem cluster, determine the degree of independence of the division of the quality problem cluster based on the first abnormal internal resistance data contained in the evaluation data set, obtain the inter-cluster difference of the quality problem cluster based on the degree of independence of the division of the quality problem cluster and the first abnormal internal resistance data contained in the quality problem cluster, determine the final quality problem clustering result based on the inter-cluster difference, and complete the production quality traceability of lithium batteries with embedded RFID.

[0010] Furthermore, the internal resistance data is screened and clustered according to the LOF outlier factor of the internal resistance data to obtain cluster clusters, including the specific method of:

[0011] The internal resistance data whose LOF outlier factor is greater than a first preset threshold value is eliminated, the variance of the remaining internal resistance data is used as the cutoff distance, and the remaining internal resistance data is clustered using a density peak clustering algorithm to obtain a cluster cluster.

[0012] Furthermore, the method of determining the radius of the cluster and obtaining the density center aggregation degree of the internal resistance data according to the internal resistance data contained in the cluster and the radius of the cluster includes the following specific methods:

[0013] The average value of the Euclidean distance from all the internal resistance data contained in the cluster to the cluster center is recorded as the radius of the cluster;

[0014] The product of the Euclidean distance between the internal resistance data and the cluster center of the cluster where the internal resistance data is located and the radius of the cluster where the internal resistance data is located is recorded as the first distance;

[0015] The ratio of the number of internal resistance data included in the cluster where the internal resistance data is located to the first distance is recorded as the density center aggregation degree of the internal resistance data.

[0016] Further, the method of determining the density concentration of the clusters, determining the internal resistance data without quality problems according to the density concentration, and determining the suspected abnormal internal resistance data includes the following specific methods:

[0017] The mean value of the density center aggregation degree of all the internal resistance data contained in the cluster is taken as the density concentration of the cluster;

[0018] The internal resistance data contained in the cluster with the largest density concentration is regarded as the internal resistance data without quality problems;

[0019] The internal resistance data that is not without quality problems is recorded as suspected abnormal internal resistance data.

[0020] Further, the specific method of obtaining the neighborhood data of suspected abnormal internal resistance includes:

[0021] The second preset threshold number of internal resistance data having the smallest Euclidean distance with the internal resistance data of the suspected abnormal internal resistance in the cluster where the suspected abnormal internal resistance is located is recorded as the neighborhood data of the suspected abnormal internal resistance.

[0022] Further, the method of obtaining the internal resistance distance sequence and the temperature distance sequence of the suspected abnormal internal resistance and obtaining the temperature correlation degree of the suspected abnormal internal resistance includes the following specific methods:

[0023] The internal resistance data of the suspected abnormal internal resistance and the neighborhood data of the suspected abnormal internal resistance are sorted in ascending order of the Euclidean distance from the internal resistance data of the suspected abnormal internal resistance to obtain an internal resistance distance sequence of the suspected abnormal internal resistance;

[0024] Arrange the temperature data corresponding to the internal resistance data contained in the internal resistance distance sequence to obtain a temperature distance sequence of suspected abnormal internal resistance;

[0025] The Pearson correlation coefficient between the internal resistance distance series and the temperature distance series of the suspected abnormal internal resistance is recorded as the temperature correlation degree of the suspected abnormal internal resistance.

[0026] Further, the method of obtaining the local abnormal consistency index of the suspected abnormal internal resistance according to the temperature correlation degree of the suspected abnormal internal resistance and the coefficient of variation of the density center aggregation degree of the suspected abnormal internal resistance, and obtaining the abnormal consistency index of the suspected abnormal internal resistance according to the local abnormal consistency index of all suspected abnormal internal resistances includes the following specific methods:

[0027] The ratio of the temperature correlation degree of the suspected abnormal internal resistance to the coefficient of variation of the density center aggregation degree of the suspected abnormal internal resistance is recorded as the local abnormal consistency index of the suspected abnormal internal resistance;

[0028] The ratio of the local abnormal consistency index of the suspected abnormal internal resistance to the maximum value of the local abnormal consistency indexes of all suspected abnormal internal resistances is recorded as the abnormal consistency index of the suspected abnormal internal resistance.

[0029] Further, the determining of the first abnormal internal resistance data according to the abnormal consistency index includes the following specific methods:

[0030] The lithium battery internal resistance data whose abnormal consistency index is less than the first abnormal threshold is recorded as the first abnormal internal resistance data.

[0031] Further, the method of determining the adjacent quality problem cluster according to the quality problem cluster and determining the evaluation data set according to the first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem cluster includes the following specific methods:

[0032] The quality problem cluster whose cluster center is closest to the cluster center of the quality problem cluster in Euclidean distance is recorded as the neighboring quality problem cluster;

[0033] The set consisting of all first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem clusters is recorded as an evaluation data set.

[0034] Furthermore, the method of determining the final quality problem clustering result according to the inter-cluster difference and completing the production quality traceability of lithium batteries with embedded RFID includes the following specific methods:

[0035] Merge the quality problem cluster whose inter-cluster difference is less than the first difference threshold with the neighboring quality problem cluster determined by the quality problem cluster to obtain the final quality problem clustering result;

[0036] The cluster center of the quality problem cluster with the largest degree of density center aggregation is used as the cluster center of the merged quality problem cluster, and the production process of the lithium battery corresponding to the cluster center of each quality problem cluster in the final quality problem clustering result is checked to determine the problematic link in the production process of the lithium battery;

[0037] All lithium batteries corresponding to the first abnormal internal resistance data included in the quality problem cluster have the same problem in the corresponding links. The lithium batteries with the same problem are recalled and processed centrally.

[0038] The beneficial effects of the present invention are:

[0039] The present invention clusters the internal resistance data of the battery through the different effects of different quality problems on the internal resistance data of the battery, obtains the degree of density center aggregation according to the size of the cluster cluster and the distance from the internal resistance data contained in the cluster cluster to the cluster center, and distinguishes the internal resistance data of the normal battery from the abnormal internal resistance data through the degree of density center aggregation; secondly, by analyzing the influence of the battery temperature on the change of the battery internal resistance data, obtains the abnormal consistency index according to the consistency between the density center aggregation degrees of the neighborhood data of the suspected abnormal internal resistance and the correlation with the battery temperature, and excludes the battery internal resistance data that is abnormal due to the influence of the temperature through the abnormal consistency index; then, the quality problem cluster is determined, the clustering results of the abnormal data are evaluated according to the differences between different quality problem clusters, and the quality problem clusters with smaller differences between clusters are merged to prevent the division redundancy of the quality problem clusters; finally, the final quality problem clustering result is determined according to the merged quality problem cluster, and the production quality traceability of the lithium battery with embedded RFID is completed, so as to solve the problem of low efficiency of production quality traceability caused by ignoring the correlation between battery data in the process of battery production quality traceability, and improve the efficiency of production quality traceability. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0041] Figure 1 A schematic diagram of a production quality tracing method for new energy batteries with embedded RFID provided by one embodiment of the present invention;

[0042] Figure 2 This is a flowchart for obtaining the first abnormal internal resistance data. DETAILED DESCRIPTION

[0043] 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.

[0044] See also Figure 1 , which shows a flow chart of a method for tracing the production quality of new energy batteries with embedded RFID provided by an embodiment of the present invention, the method comprising the following steps:

[0045] Step S001: According to lithium batteries with quality problems detected during the sampling process, the internal resistance data, temperature data and LOF outlier factors of the internal resistance data of all lithium batteries in the same batch as the lithium batteries with quality problems are obtained.

[0046] There are various types of new energy batteries, and this embodiment takes lithium batteries as an example for analysis.

[0047] During the lithium battery production process, an RFID tag is embedded in the lithium battery, and the internal resistance data and lithium battery temperature data of the lithium battery during the production process are recorded through the RFID tag.

[0048] In the lithium battery manufacturing process, it is usually necessary to conduct random inspections on the manufactured lithium batteries and perform charge and discharge cycle tests on the sampled lithium batteries, that is, to test the internal resistance of the lithium batteries to evaluate whether the performance and life of the lithium batteries meet the requirements. When lithium batteries with quality problems are found during the sampling inspection process, the internal resistance data and temperature data of all lithium batteries in the same batch as the lithium batteries with quality problems are obtained through RFID tags.

[0049] When problems occur in the production process, some internal resistance data of lithium batteries in the same batch will be quite different, and the numerical distribution of the internal resistance of these lithium batteries will be chaotic and fluctuate greatly. For example, materials with different compositions have different internal resistance rates. If the quality of some raw materials is uneven or there are impurities, the material composition of the positive and negative electrodes in different lithium batteries in the same batch will be inconsistent, which will lead to large differences in internal resistance data.

[0050] Therefore, the stability of the internal resistance data can be changed by the production quality problem of lithium batteries, and the LOF anomaly detection algorithm can be used to perform anomaly detection on the internal resistance data of all lithium batteries in the same batch of lithium batteries with quality problems, and the LOF outlier factor of each internal resistance data can be obtained. Among them, the LOF outlier factor is an indicator calculated by the LOF algorithm, which is used to indicate the degree of abnormality of the data relative to the neighboring points. The larger the LOF outlier factor, the greater the degree of abnormality of the internal resistance data; the LOF anomaly detection algorithm is a well-known technology and will not be described in detail.

[0051] At this point, the internal resistance data, temperature data, and LOF outlier factors of the internal resistance data of all lithium batteries in the same batch as the lithium battery with quality problems are obtained.

[0052] Step S002: screen and cluster the internal resistance data according to the LOF outlier factor of the internal resistance data, obtain clusters, determine the radius of the clusters, obtain the density center aggregation degree of the internal resistance data according to the internal resistance data contained in the clusters and the radius of the clusters, determine the density concentration of the clusters, determine the internal resistance data without quality problems according to the density concentration, and determine the suspected abnormal internal resistance data.

[0053] In order to prevent the lithium batteries with quality problems that are bound to exist in the production process from affecting the traceability results of lithium battery production quality, the internal resistance data with LOF outlier factors greater than the first preset threshold are eliminated and not analyzed in subsequent steps. The empirical value of the first preset threshold is 0.8.

[0054] When a problem occurs in a certain production link of lithium batteries within a period of time, the internal resistance data of the lithium batteries produced during this period will usually fluctuate within a data distribution range. Within this data distribution range, the local density of the internal resistance data of lithium batteries is relatively high. Therefore, only using anomaly detection algorithms cannot screen out abnormal internal resistance data, and abnormal data with high local density can be easily misjudged as normal data.

[0055] The remaining internal resistance data are clustered using a density peak clustering algorithm, and the variance of the internal resistance data after elimination is used as the cutoff distance of the density peak clustering algorithm to obtain cluster clusters. The use of the density peak clustering algorithm to obtain cluster clusters is a well-known technique and will not be described in detail.

[0056] The average value of the Euclidean distance from all internal resistance data contained in the cluster to the cluster center is recorded as the radius of the cluster.

[0057] The density center aggregation degree of the internal resistance data is obtained according to the internal resistance data contained in the cluster and the radius of the cluster.

[0058]

[0059] in, For the Internal resistance data of lithium batteries The degree of concentration of density centers; Internal resistance data Cluster The amount of internal resistance data contained in; Internal resistance data and The cluster Cluster of Heart The Euclidean distance between Internal resistance data Cluster The radius of .

[0060] The more internal resistance data are contained in the cluster where the internal resistance data is located and the smaller the radius of the cluster is, the denser the cluster where the internal resistance data is located is, and the greater the degree of density center concentration of the internal resistance data is. At this time, the higher the credibility of the clustering result of the internal resistance data is, the greater the possibility that the lithium battery corresponding to the internal resistance data does not have quality problems.

[0061] At this point, the density center aggregation degree of all internal resistance data in all clusters is obtained.

[0062] In order to distinguish the internal resistance data of lithium batteries with no quality problems and those with quality problems, the mean of the density center aggregation degree of all the internal resistance data contained in the cluster is taken as the density concentration of the cluster.

[0063] The internal resistance data included in the cluster with the largest density concentration is regarded as the internal resistance data without quality problems, and the internal resistance data not having quality problems is recorded as suspected abnormal internal resistance data.

[0064] At this point, the density center aggregation degree of all internal resistance data in all clusters, the internal resistance data without quality problems, and the suspected abnormal internal resistance data are obtained.

[0065] Step S003, obtain neighborhood data of suspected abnormal internal resistance, obtain internal resistance distance sequence and temperature distance sequence of suspected abnormal internal resistance, obtain temperature correlation degree of suspected abnormal internal resistance, obtain coefficient of variation of density center aggregation degree of suspected abnormal internal resistance, obtain local abnormal consistency index of suspected abnormal internal resistance according to temperature correlation degree of suspected abnormal internal resistance and coefficient of variation of density center aggregation degree of suspected abnormal internal resistance, obtain abnormal consistency index of suspected abnormal internal resistance according to local abnormal consistency index of all suspected abnormal internal resistances, and determine first abnormal internal resistance data according to abnormal consistency index.

[0066] After the lithium battery is manufactured, it is necessary to conduct quality inspection on the lithium battery. When conducting quality inspection on the lithium battery, in order to detect the internal resistance of the lithium battery, it is necessary to conduct charge and discharge cycle test, high rate charge and discharge test and high temperature test on the lithium battery. These tests all involve processes such as chemical reactions, physical changes and energy conversion. Therefore, the lithium battery that has undergone quality inspection will have different degrees of heat release, which will increase the temperature of the lithium battery itself, causing the internal resistance of the test to change, affecting the judgment of the abnormal degree of the internal resistance of the lithium battery, and thus affecting the quality traceability of the lithium battery.

[0067] Get the number of suspected abnormal internal resistance data.

[0068] There is a correlation between the resistance and temperature of a lithium battery. When the temperature of a lithium battery is low, the internal resistance of the lithium battery generally increases, and when the temperature of the lithium battery is high, the internal resistance of the lithium battery generally decreases.

[0069] Based on the correlation between resistance and temperature, the second preset threshold number of internal resistance data with the smallest Euclidean distance to the internal resistance data of the suspected abnormal internal resistance in the cluster where the suspected abnormal internal resistance is located is recorded as the neighborhood data of the suspected abnormal internal resistance. The empirical value of the second preset threshold is 5.

[0070] The internal resistance data of the suspected abnormal internal resistance and the neighborhood data of the suspected abnormal internal resistance are sorted in ascending order according to the Euclidean distance from the internal resistance data of the suspected abnormal internal resistance to obtain the internal resistance distance sequence of the suspected abnormal internal resistance. The temperature data of the suspected abnormal internal resistance and the temperature data corresponding to the neighborhood data of the suspected abnormal internal resistance are arranged in the same order to obtain the temperature distance sequence of the suspected abnormal internal resistance.

[0071] The Pearson correlation coefficient between the internal resistance distance series and the temperature distance series of the suspected abnormal internal resistance is recorded as the temperature correlation degree of the suspected abnormal internal resistance.

[0072] When the temperature correlation degree of the suspected abnormal internal resistance is greater, the correlation between the internal resistance data corresponding to the neighborhood data of the suspected abnormal internal resistance and the lithium battery temperature is higher, and the possibility that the suspected abnormal internal resistance is the lithium battery internal resistance data affected by temperature is greater.

[0073] The density center aggregation degree of lithium battery internal resistance data with the same problem should be relatively consistent. Therefore, when the density center aggregation degree of the neighborhood data of suspected abnormal internal resistance is closer, these suspected abnormal internal resistances are more likely to be lithium battery internal resistance data with the same quality problem.

[0074] According to the density center aggregation degree of the suspected abnormal internal resistance and the neighborhood data of the suspected abnormal internal resistance, the coefficient of variation of the density center aggregation degree of the suspected abnormal internal resistance is obtained. The calculation method of the coefficient of variation is a well-known technology and will not be repeated here.

[0075] When the coefficient of variation of the density center aggregation degree of the suspected abnormal internal resistance is larger, the numerical distribution of the density center aggregation degree of the suspected abnormal internal resistance and its neighborhood data is more discrete, and the possibility that the abnormal internal resistance of the lithium battery caused by the same problem as the suspected abnormal internal resistance corresponding to its neighborhood data is smaller.

[0076] The local abnormal consistency index of the suspected abnormal internal resistance is obtained according to the coefficient of variation of the temperature correlation degree of the suspected abnormal internal resistance and the density center aggregation degree of the suspected abnormal internal resistance, and the abnormal consistency index of the suspected abnormal internal resistance is obtained according to the local abnormal consistency index of all suspected abnormal internal resistances.

[0077]

[0078]

[0079] in, Suspected abnormal internal resistance The local anomaly consistency index; Suspected abnormal internal resistance The coefficient of variation of the degree of density center aggregation; Suspected abnormal internal resistance The degree of temperature correlation; Suspected abnormal internal resistance The abnormal consistency index; It is the maximum value of the local abnormal consistency index of all suspected abnormal internal resistances.

[0080] The greater the temperature correlation of the suspected abnormal internal resistance and the smaller the coefficient of variation of the density center aggregation of the neighborhood data of the suspected abnormal internal resistance, the greater the local anomaly consistency index of the suspected abnormal internal resistance, and the more likely the suspected abnormal internal resistance is normal data that is affected by temperature and classified as abnormal data.

[0081] When the local abnormal consistency index of the suspected abnormal internal resistance is larger than the maximum value of the local abnormal consistency index of all suspected abnormal internal resistances, the abnormal consistency index of the suspected abnormal internal resistance is larger, that is, the suspected abnormal internal resistance is more likely to be abnormal data obtained after being affected by temperature, and the lithium battery corresponding to the suspected abnormal internal resistance is a normal lithium battery with no production quality problems.

[0082] Set a first abnormal threshold, and record the lithium battery internal resistance data with an abnormal consistency index less than the first abnormal threshold as the first abnormal internal resistance data. The empirical value of the first abnormal threshold is 0.7. The first abnormal internal resistance data acquisition flow chart is as follows: Figure 2 shown.

[0083] At this point, the first abnormal internal resistance data is acquired.

[0084] Step S004, clustering the first abnormal internal resistance data to determine the quality problem cluster, determining the adjacent quality problem cluster based on the quality problem cluster, determining the evaluation data set based on the first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem cluster, determining the degree of independence of the division of the quality problem cluster based on the first abnormal internal resistance data contained in the evaluation data set, obtaining the inter-cluster difference of the quality problem cluster based on the degree of independence of the division of the quality problem cluster and the first abnormal internal resistance data contained in the quality problem cluster, determining the final quality problem clustering result based on the inter-cluster difference, and completing the production quality traceability of lithium batteries with embedded RFID.

[0085] When different abnormal problems occur in different production links during the manufacturing process of lithium batteries, lithium batteries will have quality problems of different degrees and characteristics, resulting in different fluctuation degrees and directions of abnormal value fluctuations in the lithium battery internal resistance data. Therefore, when clustering the first abnormal internal resistance data, the first abnormal internal resistance data corresponding to lithium batteries with different quality problems will be divided into different clusters.

[0086] The first abnormal internal resistance data is clustered using a density peak clustering algorithm, and the variance of the first abnormal internal resistance data is used as the cutoff distance of the density peak clustering algorithm to obtain clusters. Each cluster containing data of the first abnormal internal resistance data is recorded as a quality problem cluster. Among them, clustering using the density peak clustering algorithm is a well-known technology and will not be repeated here.

[0087] In order to prevent the redundancy of the division of quality problem clusters and make one quality problem correspond to multiple quality problem clusters, the Rand index is used to evaluate the clustering results.

[0088] First, cluster the core and the quality problem cluster The quality problem cluster with the closest Euclidean distance to the cluster center is recorded as the neighboring quality problem cluster When there are multiple cluster centers and quality problem clusters When the cluster center has the closest quality problem cluster in terms of Euclidean distance, only the quality problem cluster with the smallest cluster radius is recorded as the neighboring quality problem cluster. .

[0089] Clustering quality issues and The set consisting of all the first abnormal internal resistance data contained in is recorded as an evaluation data set.

[0090] Quality Problem Cluster and The cluster center is the initial cluster center of clustering. All the first abnormal internal resistance data contained in the evaluation data set are clustered using the K-means clustering algorithm. The first abnormal internal resistance data are clustered into two categories to obtain the reference cluster , . Computational Quality Problem Clusters and Clustering results with reference and The Rand index between two independent clustering results is calculated as a well-known technique and will not be described in detail.

[0091] The Rand index can reflect the similarity between two sets of clustering results. The larger the Rand index is, the more similar the data divisions of the two sets of clustering results are.

[0092] Clustering quality issues and Clustering results with reference and The Rand index between them is the quality problem cluster The degree of independence of the division .

[0093] At this point, the degree of independence of the division of quality problem clusters is obtained.

[0094] The inter-cluster difference of the quality problem clusters is obtained according to the degree of independence of the division of the quality problem clusters and the first abnormal internal resistance data contained in the quality problem clusters.

[0095]

[0096] in, Quality problem cluster The inter-cluster difference of Quality problem cluster the degree of independence of the division; Quality problem cluster The information entropy of the first abnormal internal resistance data contained therein; is an exponential function with a natural constant as its base.

[0097] The larger the information entropy of the first abnormal internal resistance data contained in the quality problem cluster, the more chaotic the first abnormal internal resistance data contained in the quality problem cluster is, the smaller the inter-cluster difference between the quality problem cluster and the adjacent quality problem clusters, the lower the degree of clustering within the quality problem cluster and the adjacent quality problem clusters, the lower the difference between different quality problem clusters, and the less credible the clustering result is.

[0098] The quality problem clusters whose inter-cluster differences are less than the first difference threshold are merged with the neighboring quality problem clusters determined by the quality problem clusters, and the cluster center of the quality problem cluster with the largest density center concentration is used as the cluster center of the merged quality problem cluster to obtain the final quality problem clustering result. Among them, the empirical value of the first difference threshold is 0.5.

[0099] The production process of lithium batteries corresponding to the cluster center of each quality problem cluster in the final quality problem clustering result is checked to determine the problematic link in the production process of lithium batteries. All lithium batteries corresponding to the first abnormal internal resistance data contained in the quality problem cluster have the same problem in the corresponding link, and these lithium batteries are recalled and processed centrally.

[0100] At this point, the production quality traceability of lithium batteries with embedded RFID is completed.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for tracing the production quality of new energy batteries with embedded RFID, characterized in that: The method comprises the following steps: According to the lithium batteries with quality problems detected during the sampling process, the internal resistance data and temperature data of all lithium batteries in the same batch as the lithium batteries with quality problems are obtained, and the LOF algorithm is used to perform anomaly detection on the internal resistance data to obtain the local outlier factor LOF of the internal resistance data; The internal resistance data is screened and clustered according to the local outlier factor LOF of the internal resistance data, the clustering cluster is obtained, the radius of the clustering cluster is determined, the density center aggregation degree of the internal resistance data is obtained according to the internal resistance data contained in the clustering cluster and the radius of the clustering cluster, the density concentration of the clustering cluster is determined, and the internal resistance data without quality problems is determined according to the density concentration, and the suspected abnormal internal resistance data is determined; Obtaining neighborhood data of the suspected abnormal internal resistance, obtaining an internal resistance distance sequence and a temperature distance sequence of the suspected abnormal internal resistance, obtaining a temperature correlation degree of the suspected abnormal internal resistance, obtaining a coefficient of variation of a density center aggregation degree of the suspected abnormal internal resistance, obtaining a local abnormal consistency index of the suspected abnormal internal resistance according to the temperature correlation degree of the suspected abnormal internal resistance and the coefficient of variation of a density center aggregation degree of the suspected abnormal internal resistance, obtaining an abnormal consistency index of the suspected abnormal internal resistance according to the local abnormal consistency indexes of all suspected abnormal internal resistances, and determining first abnormal internal resistance data according to the abnormal consistency index; Cluster the first abnormal internal resistance data to determine the quality problem cluster, determine the adjacent quality problem cluster based on the quality problem cluster, determine the evaluation data set based on the first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem cluster, determine the degree of independence of the division of the quality problem cluster based on the first abnormal internal resistance data contained in the evaluation data set, obtain the inter-cluster difference of the quality problem cluster based on the degree of independence of the division of the quality problem cluster and the first abnormal internal resistance data contained in the quality problem cluster, determine the final quality problem clustering result based on the inter-cluster difference, and complete the production quality traceability of lithium batteries with embedded RFID.

2. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The method of screening and clustering the internal resistance data according to the local outlier factor LOF of the internal resistance data to obtain the cluster clusters includes the following specific methods: The internal resistance data whose local outlier factor LOF is greater than a first preset threshold value is eliminated, the variance of the remaining internal resistance data is used as the cutoff distance, and the remaining internal resistance data is clustered using a density peak clustering algorithm to obtain a cluster cluster.

3. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The method of determining the radius of the cluster and obtaining the density center aggregation degree of the internal resistance data according to the internal resistance data contained in the cluster and the radius of the cluster includes the following specific methods: The average value of the Euclidean distance from all the internal resistance data contained in the cluster to the cluster center is recorded as the radius of the cluster; The product of the Euclidean distance between the internal resistance data and the cluster center of the cluster where the internal resistance data is located and the radius of the cluster where the internal resistance data is located is recorded as the first distance; The ratio of the number of internal resistance data included in the cluster where the internal resistance data is located to the first distance is recorded as the density center aggregation degree of the internal resistance data.

4. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The method of determining the density concentration of the clusters, determining the internal resistance data without quality problems according to the density concentration, and determining the suspected abnormal internal resistance data includes the following specific methods: The mean value of the density center aggregation degree of all the internal resistance data contained in the cluster is taken as the density concentration of the cluster; The internal resistance data contained in the cluster with the largest density concentration is regarded as the internal resistance data without quality problems; The internal resistance data that is not without quality problems is recorded as suspected abnormal internal resistance data.

5. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The specific method of obtaining the neighborhood data of suspected abnormal internal resistance includes: The second preset threshold number of internal resistance data having the smallest Euclidean distance with the internal resistance data of the suspected abnormal internal resistance in the cluster where the suspected abnormal internal resistance is located is recorded as the neighborhood data of the suspected abnormal internal resistance.

6. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The specific method of obtaining the internal resistance distance sequence and the temperature distance sequence of the suspected abnormal internal resistance and obtaining the temperature correlation degree of the suspected abnormal internal resistance includes: The internal resistance data of the suspected abnormal internal resistance and the neighborhood data of the suspected abnormal internal resistance are sorted in ascending order of the Euclidean distance from the internal resistance data of the suspected abnormal internal resistance to obtain an internal resistance distance sequence of the suspected abnormal internal resistance; Arrange the temperature data corresponding to the internal resistance data contained in the internal resistance distance sequence to obtain a temperature distance sequence of suspected abnormal internal resistance; The Pearson correlation coefficient between the internal resistance distance series and the temperature distance series of the suspected abnormal internal resistance is recorded as the temperature correlation degree of the suspected abnormal internal resistance.

7. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The method of obtaining the local abnormal consistency index of the suspected abnormal internal resistance according to the temperature correlation degree of the suspected abnormal internal resistance and the coefficient of variation of the density center aggregation degree of the suspected abnormal internal resistance, and obtaining the abnormal consistency index of the suspected abnormal internal resistance according to the local abnormal consistency index of all suspected abnormal internal resistances includes the following specific methods: The ratio of the temperature correlation degree of the suspected abnormal internal resistance to the coefficient of variation of the density center aggregation degree of the suspected abnormal internal resistance is recorded as the local abnormal consistency index of the suspected abnormal internal resistance; The ratio of the local abnormal consistency index of the suspected abnormal internal resistance to the maximum value of the local abnormal consistency indexes of all suspected abnormal internal resistances is recorded as the abnormal consistency index of the suspected abnormal internal resistance.

8. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1 is characterized in that: The specific method of determining the first abnormal internal resistance data according to the abnormal consistency index includes: The lithium battery internal resistance data whose abnormal consistency index is less than the first abnormal threshold is recorded as the first abnormal internal resistance data.

9. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1, characterized in that: The method of determining the adjacent quality problem cluster according to the quality problem cluster and determining the evaluation data set according to the first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem cluster includes the following specific methods: The quality problem cluster whose cluster center is closest to the cluster center of the quality problem cluster in Euclidean distance is recorded as the neighboring quality problem cluster; The set consisting of all first abnormal internal resistance data contained in the quality problem cluster and the adjacent quality problem clusters is recorded as an evaluation data set.

10. The method for tracing the production quality of new energy batteries with embedded RFID according to claim 1, characterized in that: The specific method of determining the final quality problem clustering result according to the inter-cluster difference and completing the production quality traceability of lithium batteries with embedded RFID includes: Merge the quality problem cluster whose inter-cluster difference is less than the first difference threshold with the neighboring quality problem cluster determined by the quality problem cluster to obtain the final quality problem clustering result; The cluster center of the quality problem cluster with the largest degree of density center aggregation is used as the cluster center of the merged quality problem cluster, and the production process of the lithium battery corresponding to the cluster center of each quality problem cluster in the final quality problem clustering result is checked to determine the problematic link in the production process of the lithium battery; All lithium batteries corresponding to the first abnormal internal resistance data included in the quality problem cluster have the same problem in the corresponding links. The lithium batteries with the same problem are recalled and processed centrally.