WO2022143923A1

WO2022143923A1 - State of health evaluation method and system for vehicle battery, and electronic device and medium

Info

Publication number: WO2022143923A1
Application number: PCT/CN2021/143196
Authority: WO
Inventors: 郑立华
Original assignee: 奥动新能源汽车科技有限公司; 上海电巴新能源科技有限公司
Priority date: 2020-12-31
Filing date: 2021-12-30
Publication date: 2022-07-07
Also published as: CN114693043A

Abstract

A state of health evaluation method for a vehicle battery. The evaluation method comprises the following steps: S1, acquiring an SOH data set of a vehicle battery; S2, dividing the SOH data set into normal SOH values and abnormal SOH values; S3, repairing the abnormal SOH values, so as to obtain repaired SOH values corresponding to the abnormal SOH values; and S4, obtaining a state of health evaluation result of the vehicle battery according to the distribution of the normal SOH values and the repaired SOH values.

Description

车用电池的健康状况评估方法、***、电子设备和介质Health assessment method, system, electronic device and medium for vehicle battery

本申请要求申请日为2020/12/31的中国专利申请2020116187761的优先权。本申请引用上述中国专利申请的全文。This application claims the priority of Chinese patent application 2020116187761 with an application date of 2020/12/31. This application cites the full text of the above Chinese patent application.

技术领域technical field

本发明属于电池的健康状况的获取技术领域，尤其涉及一种车用电池的健康状况评估方法、***、电子设备和介质。The invention belongs to the technical field of acquiring the health status of batteries, and in particular relates to a method, system, electronic device and medium for evaluating the health status of a vehicle battery.

背景技术Background technique

在充换电领域，电池需要被反复充电以便反复投入使用，那么电池的健康情况就是运营过程中需要重点关注的方面。目前业界界定电池的健康情况，有不同的计算方式，最终都会得出一个小于1的数值，代表电池的健康情况，称为SOH(电池健康状态)。然而目前通常基于人工的方式对SOH值进行异常识别，但是电池的SOH值数据量巨大，通过人工识别效率低下。In the field of charging and replacing, the battery needs to be charged repeatedly in order to be put into use repeatedly, so the health of the battery is an aspect that needs to be paid attention to in the operation process. At present, the industry defines the health status of the battery, and there are different calculation methods. In the end, a value less than 1 is obtained, which represents the health status of the battery, which is called SOH (battery state of health). However, at present, the abnormal identification of SOH value is usually based on manual methods, but the amount of SOH value data of batteries is huge, and the efficiency of manual identification is low.

发明内容SUMMARY OF THE INVENTION

本发明要解决的技术问题是为了克服现有技术中电池的健康情况的评估的效率低的缺陷，提供一种车用电池的健康状况评估方法、***、电子设备和介质。The technical problem to be solved by the present invention is to provide a method, system, electronic device and medium for evaluating the health status of a vehicle battery in order to overcome the defect of low efficiency of battery health status assessment in the prior art.

本发明是通过下述技术方案来解决上述技术问题：The present invention solves the above-mentioned technical problems through the following technical solutions:

本发明提供一种车用电池的健康状况评估方法，包括以下步骤：The present invention provides a method for evaluating the health status of a vehicle battery, comprising the following steps:

获取车用电池的SOH数据集；Obtain the SOH dataset of vehicle batteries;

从SOH数据集中划分出正常SOH值和异常SOH值；Divide the normal SOH value and the abnormal SOH value from the SOH data set;

对异常SOH进行修复，得到异常SOH值相应的修复SOH值；Repair the abnormal SOH to obtain the repaired SOH value corresponding to the abnormal SOH value;

根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果。According to the distribution of the normal SOH value and the repaired SOH value, the health evaluation results of the vehicle battery are obtained.

在本技术方案中，自动从SOH数据集中划分出正常SOH值和异常SOH值，然后对异常SOH值进行修复，再基于修复后的数据的分布情况得到车用电池的健康状况评估结果，能够提高车用电池的健康状况评估的效率。而且对异常SOH值进行修复后再进行健康状况评估，还能提高车用电池的健康状况评估结果的准确性，对车用电池的后续操作提供了较佳的参考。In this technical solution, the normal SOH value and the abnormal SOH value are automatically divided from the SOH data set, and then the abnormal SOH value is repaired, and the health condition evaluation result of the vehicle battery is obtained based on the distribution of the repaired data, which can improve the Efficiency of vehicle battery health assessment. Moreover, after repairing the abnormal SOH value, the health status assessment can also improve the accuracy of the health status assessment result of the vehicle battery, and provide a better reference for the subsequent operation of the vehicle battery.

较佳地，从SOH数据集中划分出正常SOH值和异常SOH值之前，车用电池的健康状况评估方法还包括：Preferably, before dividing the normal SOH value and the abnormal SOH value from the SOH data set, the health condition assessment method of the vehicle battery further includes:

获取SOH数据集中各SOH值对应的降维分析数据，得到SOH数据集对应的分析数据集；Obtain the dimensionality reduction analysis data corresponding to each SOH value in the SOH data set, and obtain the analysis data set corresponding to the SOH data set;

确定分析数据集的聚类数量以及主成分；Determine the number of clusters and principal components of the analysis dataset;

根据聚类数量和主成分对分析数据集进行聚类。The analysis dataset is clustered according to the number of clusters and principal components.

在该技术方案中，通过获取各SOH值对应的降维分析数据，得到SOH数据集对应的分析数据集，基于分析数据集进行聚类可提高数据处理效率；而且选定分析数据集的聚类数量以及主成分，并根据聚类数量和主成分对分析数据集进行聚类，可以有效提高聚类的准确性，从而提高基于聚类结果划分正常SOH值和异常SOH值的准确性。In this technical solution, by obtaining the dimensionality reduction analysis data corresponding to each SOH value, the analysis data set corresponding to the SOH data set is obtained, and the clustering based on the analysis data set can improve the data processing efficiency; and the clustering of the analysis data set is selected. Quantity and principal components, and clustering the analysis data set according to the number of clusters and principal components can effectively improve the accuracy of clustering, thereby improving the accuracy of dividing normal SOH values and abnormal SOH values based on the clustering results.

较佳地，确定分析数据集的聚类数量以及主成分，包括：Preferably, determine the number of clusters and principal components of the analysis data set, including:

根据分析数据集中的降维分析数据，筛选候选成分；Screen candidate components according to the dimensionality reduction analysis data in the analysis data set;

根据候选成分的解释方差率，从候选成分中筛选主成分；Screen the principal components from the candidate components according to the explained variance ratio of the candidate components;

根据聚类数量和主成分对分析数据集进行聚类，包括：The analysis dataset is clustered according to the number of clusters and principal components, including:

选取聚类数量的且包括主成分的聚类中心点；Select the cluster center points of the number of clusters and including the principal components;

以聚类中心点为聚类中心进行聚类，在聚类后重新选取每类的聚类中心点且保持主成分为聚类中心点再继续聚类直至收敛。Clustering is performed with the cluster center point as the cluster center. After clustering, the cluster center point of each category is reselected and the principal component is kept as the cluster center point, and then the clustering is continued until convergence.

在该技术方案中，根据候选成分的解释方差率筛选主成分，且维持主成分为聚类中心点进行聚类，可以有效提高聚类的准确性。In this technical solution, the principal components are screened according to the explained variance rate of the candidate components, and the principal components are maintained as the cluster center points for clustering, which can effectively improve the accuracy of the clustering.

较佳地，从SOH数据集中划分出正常SOH值和异常SOH值，包括：Preferably, the normal SOH value and the abnormal SOH value are divided from the SOH data set, including:

获取一个降维分析数据与所对应的聚类中心之间的距离，并将对应的距离最大的多个降维分析数据为异常降维分析数据，以及将异常降维分析数据对应的SOH值识别为异常SOH值；Obtain the distance between a dimensionality reduction analysis data and the corresponding cluster center, and identify the multiple dimensionality reduction analysis data with the largest corresponding distance as abnormal dimensionality reduction analysis data, and identify the SOH value corresponding to the abnormal dimensionality reduction analysis data is abnormal SOH value;

或者，根据异常比例从每类聚类数据中筛选异常降维分析数据，并将异常降维分析数据对应的SOH值识别为异常SOH值。Alternatively, the abnormal dimensionality reduction analysis data is screened from each type of cluster data according to the abnormal proportion, and the SOH value corresponding to the abnormal dimensionality reduction analysis data is identified as the abnormal SOH value.

在该技术方案中，提供了基于聚类结果划分正常SOH值和异常SOH值的具体方式，将离散度最高的多个降维分析数据划分为异常降维分析数据或者将根据异常比例从每类聚类数据中筛选异常降维分析数据，再将异常降维分析数据对应的SOH值识别为异常SOH值，这样可以有效地从大量的SOH值中快速且准确地划分出异常SOH值。In this technical solution, a specific method for dividing the normal SOH value and the abnormal SOH value based on the clustering result is provided. The abnormal dimensionality reduction analysis data is screened from the cluster data, and then the SOH value corresponding to the abnormal dimensionality reduction analysis data is identified as the abnormal SOH value, which can effectively divide the abnormal SOH value from a large number of SOH values quickly and accurately.

较佳地，对异常SOH值进行修复，得到异常SOH值相应的修复SOH值，包括：Preferably, the abnormal SOH value is repaired to obtain the repaired SOH value corresponding to the abnormal SOH value, including:

根据异常SOH值相邻的正常SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值；Repair the abnormal SOH value according to the normal SOH value adjacent to the abnormal SOH value, and obtain the repaired SOH value corresponding to the abnormal SOH value;

或者，获取异常SOH值对应的至少一个备用SOH值，并在备用SOH值为正常SOH值时，根据备用SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值；异常SOH值和对应的备用SOH值根据不同的计算方式得到。Or, obtain at least one backup SOH value corresponding to the abnormal SOH value, and when the backup SOH value is a normal SOH value, repair the abnormal SOH value according to the backup SOH value, and obtain a repaired SOH value corresponding to the abnormal SOH value; the abnormal SOH value and The corresponding backup SOH value is obtained according to different calculation methods.

在该技术方案中，根据异常SOH值相邻的正常SOH值对异常SOH值进行修复，或者基于备用SOH值对异常SOH值进行修复，可以提高数据修复的可信度以及准确性，进而保证车用电池的健康状况评估的准确性。In this technical solution, the abnormal SOH value is repaired according to the normal SOH value adjacent to the abnormal SOH value, or the abnormal SOH value is repaired based on the standby SOH value, which can improve the reliability and accuracy of data repair, thereby ensuring the vehicle The accuracy of the battery health assessment.

较佳地，获取异常SOH值对应的至少一个备用SOH值，并在备用SOH值为正常SOH值时，根据备用SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值，包括：Preferably, at least one backup SOH value corresponding to the abnormal SOH value is obtained, and when the backup SOH value is a normal SOH value, the abnormal SOH value is repaired according to the backup SOH value, and the repaired SOH value corresponding to the abnormal SOH value is obtained, including:

获取异常SOH值对应的备用SOH值；Obtain the backup SOH value corresponding to the abnormal SOH value;

在备用SOH值为两个且均为正常SOH值时，获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；When there are two backup SOH values and both are normal SOH values, obtain the variation range between the two backup SOH values and the adjacent normal SOH values;

选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值。The smaller of the two changes is selected as the corresponding repaired SOH value for the abnormal SOH value.

在该技术方案中，提供了基于异常SOH值的备用SOH值对异常SOH值进行修复的具体方式，在备用SOH值为两个且均为正常SOH值时，获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值，可以提高数据修复的准确性，进而保证车用电池的健康状况评估的准确性。In this technical solution, a specific method for repairing an abnormal SOH value based on the backup SOH value of the abnormal SOH value is provided. When the backup SOH values are two and both are normal SOH values, the two backup SOH values are obtained respectively with the corresponding SOH values. The change range between the adjacent normal SOH values; the smaller of the two change ranges is selected as the repaired SOH value corresponding to the abnormal SOH value, which can improve the accuracy of data repair and ensure the health assessment of the vehicle battery. accuracy.

较佳地，根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果之后，还包括：Preferably, according to the distribution of the normal SOH value and the repaired SOH value, after obtaining the health condition assessment result of the vehicle battery, the method further includes:

对车用电池执行与健康状况评估结果相匹配的目标操作。Perform targeted actions on the vehicle battery that match the results of the health assessment.

在该技术方案中，基于车用电池的健康状况评估结果对车用电池执行与健康状况评估结果相匹配的目标操作，既能够及时处理异常避免安全问题发生，又能够物尽其用避免资源浪费。In this technical solution, based on the health status assessment result of the vehicle battery, a target operation that matches the health status assessment result is performed on the vehicle battery, which can not only deal with the abnormality in time to avoid the occurrence of safety problems, but also make the best use of everything to avoid waste of resources. .

本发明还提供一种车用电池的健康状况评估***，包括获取单元、划分单元、修复单元、评估单元；The present invention also provides a system for evaluating the health status of a vehicle battery, including an acquisition unit, a division unit, a repair unit, and an evaluation unit;

获取单元用于获取车用电池的SOH数据集；The acquisition unit is used to acquire the SOH data set of the vehicle battery;

划分单元用于从SOH数据集中划分出正常SOH值和异常SOH值；The division unit is used to divide the normal SOH value and the abnormal SOH value from the SOH data set;

修复单元用于对异常SOH进行修复，得到异常SOH值相应的修复SOH值；The repair unit is used to repair the abnormal SOH to obtain the repaired SOH value corresponding to the abnormal SOH value;

评估单元用于根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果。The evaluation unit is used to obtain the health condition evaluation result of the vehicle battery according to the distribution of the normal SOH value and the repaired SOH value.

较佳地，获取单元还用于获取SOH数据集中各SOH值对应的降维分析数据，得到SOH数据集对应的分析数据集；Preferably, the obtaining unit is further configured to obtain the dimensionality reduction analysis data corresponding to each SOH value in the SOH data set, and obtain the analysis data set corresponding to the SOH data set;

获取单元还用于确定分析数据集的聚类数量以及主成分；The acquisition unit is also used to determine the number of clusters and principal components of the analysis data set;

获取单元还用于根据聚类数量和主成分对分析数据集进行聚类。The acquisition unit is also used to cluster the analysis dataset according to the number of clusters and principal components.

较佳地，获取单元还用于根据分析数据集中的降维分析数据，筛选候选成分；Preferably, the acquisition unit is also used for screening candidate components according to the dimensionality reduction analysis data in the analysis data set;

获取单元还用于根据候选成分的解释方差率，从候选成分中筛选主成分；The acquisition unit is also used to filter principal components from the candidate components according to the explained variance rate of the candidate components;

获取单元还用于根据聚类数量和主成分对分析数据集进行聚类，包括：The acquisition unit is also used to cluster analysis datasets according to the number of clusters and principal components, including:

获取单元还用于选取聚类数量的且包括主成分的聚类中心点；The acquisition unit is also used to select the cluster center points of the number of clusters and including the principal components;

获取单元还用于以聚类中心点为聚类中心进行聚类，在聚类后重新选取每类的聚类中心点且保持主成分为聚类中心点再继续聚类直至收敛。The acquisition unit is also used for clustering with the cluster center point as the cluster center, and after clustering, reselects the cluster center point of each type and keeps the principal component as the cluster center point, and then continues the clustering until convergence.

较佳地，划分单元还用于获取一个降维分析数据与所对应的聚类中心之间的距离，并将对应的距离最大的多个降维分析数据为异常降维分析数据，以及将异常降维分析数据对应的SOH值识别为异常SOH值；Preferably, the dividing unit is also used to obtain the distance between one dimension reduction analysis data and the corresponding cluster center, and classify the multiple dimension reduction analysis data with the largest corresponding distances as abnormal dimension reduction analysis data, and classify the abnormal The SOH value corresponding to the dimensionality reduction analysis data is identified as an abnormal SOH value;

或者，划分单元还用于根据异常比例从每类聚类数据中筛选异常降维分析数据，并将异常降维分析数据对应的SOH值识别为异常SOH值。Alternatively, the dividing unit is further configured to filter the abnormal dimension reduction analysis data from each type of cluster data according to the abnormal proportion, and identify the SOH value corresponding to the abnormal dimension reduction analysis data as the abnormal SOH value.

较佳地，修复单元还用于根据异常SOH值相邻的正常SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值；Preferably, the repairing unit is also used to repair the abnormal SOH value according to the adjacent normal SOH value of the abnormal SOH value, and obtain the corresponding repaired SOH value of the abnormal SOH value;

或者，修复单元还用于获取异常SOH值对应的至少一个备用SOH值，并在备用SOH值为正常SOH值时，根据备用SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值；异常SOH值和对应的备用SOH值根据不同的计算方式得到。Alternatively, the repair unit is further configured to obtain at least one backup SOH value corresponding to the abnormal SOH value, and when the backup SOH value is a normal SOH value, repair the abnormal SOH value according to the backup SOH value, and obtain a repaired SOH value corresponding to the abnormal SOH value ; Abnormal SOH value and corresponding backup SOH value are obtained according to different calculation methods.

较佳地，修复单元还用于获取异常SOH值对应的备用SOH值；Preferably, the repair unit is also used to obtain the standby SOH value corresponding to the abnormal SOH value;

在备用SOH值为两个且均为正常SOH值时，修复单元还用于获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；When there are two backup SOH values and both are normal SOH values, the repair unit is further configured to obtain the variation range between the two backup SOH values and the adjacent normal SOH values;

修复单元还用于选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值。The repair unit is also used to select the smaller of the two variation ranges and use it as the repaired SOH value corresponding to the abnormal SOH value.

较佳地，评估单元还用于对车用电池执行与健康状况评估结果相匹配的目标操作。Preferably, the evaluation unit is further configured to perform a target operation matching the health condition evaluation result on the vehicle battery.

本发明还提供一种电子设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，处理器执行计算机程序时实现本发明的车用电池的健康状况评估方法。The present invention also provides an electronic device, comprising a memory, a processor and a computer program stored in the memory and running on the processor. The processor implements the method for evaluating the health condition of a vehicle battery of the present invention when the processor executes the computer program.

本发明还提供一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现本发明的车用电池的健康状况评估方法的步骤。The present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the steps of the method for evaluating the health condition of a vehicle battery of the present invention.

本发明的积极进步效果在于：自动从SOH数据集中划分出正常SOH值和异常SOH值，然后对异常SOH值进行修复，再基于修复后的数据的分布情况得到车用电池的健康状况评估结果，能够提高车用电池的健康状况评估的效率。而且对异常SOH值进行修复后再进行健康状况评估，还能提高车用电池的健康状况评估结果的准确性，对车用电池的后续操作提供了较佳的参考。The positive improvement effect of the present invention is that: the normal SOH value and the abnormal SOH value are automatically divided from the SOH data set, and then the abnormal SOH value is repaired, and then the health condition evaluation result of the vehicle battery is obtained based on the distribution of the repaired data, It can improve the efficiency of the health assessment of the vehicle battery. Moreover, after repairing the abnormal SOH value, the health status assessment can also improve the accuracy of the health status assessment result of the vehicle battery, and provide a better reference for the subsequent operation of the vehicle battery.

附图说明Description of drawings

图1为本发明的实施例1的车用电池的健康状况评估方法的流程图。FIG. 1 is a flowchart of a method for evaluating the state of health of a vehicle battery according to Embodiment 1 of the present invention.

图2为本发明的实施例1的车用电池的健康状况评估方法的SOH数据集的各个SOH值的分布情况的示意图。FIG. 2 is a schematic diagram of the distribution of each SOH value in the SOH data set of the method for evaluating the state of health of a vehicle battery according to Embodiment 1 of the present invention.

图3为本发明的实施例1的车用电池的健康状况评估方法的elbow曲线的示意图。3 is a schematic diagram of an elbow curve of the method for evaluating the state of health of a vehicle battery according to Embodiment 1 of the present invention.

图4为本发明的实施例1的车用电池的健康状况评估方法的SOH数据集的累积解释方差和独立的解释方差的示意图。4 is a schematic diagram of the cumulative explained variance and the independent explained variance of the SOH data set of the vehicle battery health condition assessment method according to Embodiment 1 of the present invention.

图5为本发明的实施例1的车用电池的健康状况评估方法的划分的正常值和异常值的示意图。FIG. 5 is a schematic diagram of divided normal values and abnormal values of the method for evaluating the state of health of a vehicle battery according to Embodiment 1 of the present invention.

图6为本发明的实施例1的车用电池的健康状况评估方法的修复后的数据的分布状况的示意图。FIG. 6 is a schematic diagram of the distribution state of the restored data of the method for evaluating the state of health of a vehicle battery according to Embodiment 1 of the present invention.

图7为本发明的实施例3的电子设备的结构示意图。FIG. 7 is a schematic structural diagram of an electronic device according to Embodiment 3 of the present invention.

图8为本发明的实施例5的车用电池的健康状况评估***的结构示意图。FIG. 8 is a schematic structural diagram of a vehicle battery health evaluation system according to Embodiment 5 of the present invention.

具体实施方式Detailed ways

下面通过实施例的方式进一步说明本发明，但并不因此将本发明限制在所述的实施例范围之中。The present invention is further described below by way of examples, but the present invention is not limited to the scope of the described examples.

实施例1Example 1

本实施例提供一种车用电池的健康状况评估方法。参照图1，该车用电池的健康状况评估方法包括以下步骤：This embodiment provides a method for evaluating the health status of a vehicle battery. Referring to FIG. 1 , the method for evaluating the health status of the vehicle battery includes the following steps:

步骤S1、获取车用电池的SOH数据集。Step S1, acquiring the SOH data set of the vehicle battery.

步骤S2、从SOH数据集中划分出正常SOH值和异常SOH值。Step S2, dividing the normal SOH value and the abnormal SOH value from the SOH data set.

步骤S3、对异常SOH进行修复，得到异常SOH值相应的修复SOH值。Step S3, repairing the abnormal SOH to obtain a repaired SOH value corresponding to the abnormal SOH value.

步骤S4、根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果。Step S4, according to the distribution of the normal SOH value and the repaired SOH value, obtain the health condition evaluation result of the vehicle battery.

具体实施时，首先，在步骤S1中，获取车用电池的SOH数据集。During specific implementation, first, in step S1, the SOH data set of the vehicle battery is acquired.

在一块电池的SOH数据集中，正常情况下，是由大部分正常的SOH值和部分异常的SOH值组成。正常的SOH值具备相似的特征，异常的SOH值的特征跟正常的SOH值的特征相差甚大。图2示出了一个SOH数据集的各个SOH值的分布情况，其中，横轴表征SOH值，纵轴表征对应的SOH值的数据的数量。该SOH数据集大量的SOH值处于0.8-1.0之间，符合正常的SOH值的特点；其中少量数据分布与0-0.8之间。可以判断，这部分少量数据，其中存在一些异常值。In the SOH data set of a battery, under normal circumstances, it consists of most of the normal SOH values and some abnormal SOH values. Normal SOH values have similar characteristics, and the characteristics of abnormal SOH values are quite different from those of normal SOH values. FIG. 2 shows the distribution of each SOH value of an SOH data set, wherein the horizontal axis represents the SOH value, and the vertical axis represents the number of data corresponding to the SOH value. A large number of SOH values in this SOH data set are between 0.8-1.0, which is in line with the characteristics of normal SOH values; a small amount of data is distributed between 0-0.8. It can be judged that there are some outliers in this small amount of data.

为了识别异常值，首先，获取SOH数据集中各SOH值对应的降维分析数据，得到SOH数据集对应的分析数据集。In order to identify outliers, first, the dimensionality reduction analysis data corresponding to each SOH value in the SOH data set is obtained, and the analysis data set corresponding to the SOH data set is obtained.

然后，确定分析数据集的聚类数量以及主成分。Then, determine the number of clusters and principal components of the analysis dataset.

可实施地，可通过任意能够确定数据集的聚类数量以及主成分的算法来确定分析数据集的聚类数量以及主成分，本实施例并不具体限定，该算法可根据实际需求进行相应的选择及调整。Implementably, the number of clusters and the principal components of the analysis data set can be determined by any algorithm that can determine the number of clusters and the principal components of the data set. Select and adjust.

例如可基于elbow(肘系数)方法来确定聚类数量。图3示出了一elbow曲线，当聚类数量趋于6时，曲线开始趋于收敛，此时可以取聚类数量为6。The number of clusters can be determined based on the elbow method, for example. Figure 3 shows an elbow curve, when the number of clusters tends to 6, the curve begins to converge, and the number of clusters can be taken as 6 at this time.

然后，基于主成分分析法分析数据集中的主成分及每个主成分的解释方差率。图4示出了SOH数据集的累积解释方差和独立的解释方差，其中，横轴表征主成分，纵轴表征解释方差率，图中折线对应累积解释方差，矩形色块对应独立的解释方差。纵轴表征解释方差率。根据图4所示，数据集中共有两个主成分，每个主成分大概解释了一半的方差，故两个主成分均需要保留。Then, the principal components in the data set and the explained variance ratio of each principal component are analyzed based on the principal component analysis method. Figure 4 shows the cumulative explained variance and the independent explained variance of the SOH dataset, where the horizontal axis represents the principal components, the vertical axis represents the explained variance rate, the broken line in the figure corresponds to the cumulative explained variance, and the rectangular color block corresponds to the independent explained variance. The vertical axis represents the explained variance rate. As shown in Figure 4, there are two principal components in the data set, and each principal component explains about half of the variance, so both principal components need to be retained.

确定了聚类数量，确定了主成分个数，接下来则进行数据集的标准化处理并进行KMeans聚类。聚类的过程包括：选取聚类数量的且包括主成分的聚类中心点；以聚类中心点为聚类中心进行聚类，在聚类后重新选取每类的聚类中心点且保持主成分为聚类中心点再继续聚类直至收敛。The number of clusters is determined, the number of principal components is determined, and then the data set is standardized and KMeans clustering is performed. The process of clustering includes: selecting the number of cluster center points and including the principal components; clustering with the cluster center points as the cluster centers, and re-selecting the cluster center points of each type after clustering and keeping the main points. The components are cluster center points and continue to cluster until convergence.

通过获取各SOH值对应的降维分析数据，得到SOH数据集对应的分析数据集，基于分析数据集进行聚类可提高数据处理效率；而且选定分析数据集的聚类数量以及主成分，并根据聚类数量和主成分对分析数据集进行聚类，可以有效提高聚类的准确性，从而提高基于聚类结果划分正常SOH值和异常SOH值的准确性。By obtaining the dimensionality reduction analysis data corresponding to each SOH value, the analysis data set corresponding to the SOH data set is obtained, and clustering based on the analysis data set can improve the data processing efficiency; Clustering the analysis data set according to the number of clusters and principal components can effectively improve the accuracy of clustering, thereby improving the accuracy of dividing normal SOH values and abnormal SOH values based on the clustering results.

根据候选成分的解释方差率筛选主成分，且维持主成分为聚类中心点进行聚类，可以有效提高聚类的准确性。Screening the principal components according to the explained variance rate of the candidate components, and maintaining the principal components as the cluster center points for clustering, can effectively improve the accuracy of the clustering.

聚类完成后，从SOH数据集中划分出正常SOH值和异常SOH值。After the clustering is completed, normal SOH values and abnormal SOH values are divided from the SOH dataset.

在第一种可选的实施方式中，获取一个降维分析数据与所对应的聚类中心之间的距离，并将对应的距离最大的多个降维分析数据为异常降维分析数据，以及将异常降维分析数据对应的SOH值识别为异常SOH值。In a first optional implementation manner, the distance between a piece of dimensionality reduction analysis data and the corresponding cluster center is obtained, and a plurality of dimensionality reduction analysis data with the largest corresponding distances are regarded as abnormal dimensionality reduction analysis data, and The SOH value corresponding to the abnormal dimensionality reduction analysis data is identified as the abnormal SOH value.

在第二种可选的实施方式中，根据异常比例从每类聚类数据中筛选异常降维分析数据，并将异常降维分析数据对应的SOH值识别为异常SOH值。例如，在一种情形下，设定一个异常值的比例outliers_fraction为1％，这样设置是因为在标准正太分布的情况下(N(0,1))，一般认定3个标准差以外的数据为异常值。3个标准差以内的数据包含了数据集中99％以上的数据，所以剩下的1％的数据可以视为异常值。In a second optional embodiment, the abnormal dimensionality reduction analysis data is screened from each type of cluster data according to the abnormality ratio, and the SOH value corresponding to the abnormal dimensionality reduction analysis data is identified as the abnormal SOH value. For example, in one case, the ratio of an outliers_fraction is set to 1%. This setting is because in the case of a standard normal distribution (N(0,1)), it is generally considered that data beyond 3 standard deviations are outliers. Data within 3 standard deviations contains more than 99% of the data in the dataset, so the remaining 1% of data can be considered outliers.

本实施例提供了基于聚类结果划分正常SOH值和异常SOH值的具体方式，将离散度最高的多个降维分析数据划分为异常降维分析数据或者将根据异常比例从每类聚类数据中筛选异常降维分析数据，再将异常降维分析数据对应的SOH值识别为异常SOH值，这样可以有效地从大量的SOH值中快速且准确地划分出异常SOH值。This embodiment provides a specific method for dividing the normal SOH value and the abnormal SOH value based on the clustering result, dividing the multiple dimensionality reduction analysis data with the highest dispersion into abnormal dimensionality reduction analysis data, or dividing the data from each type of clustering data according to the abnormal ratio The abnormal dimensionality reduction analysis data is screened, and then the SOH value corresponding to the abnormal dimensionality reduction analysis data is identified as the abnormal SOH value, which can effectively divide the abnormal SOH value from a large number of SOH values quickly and accurately.

图5以图示的方式示出了划分的正常值和异常值，其中，横轴表征时间，纵轴表征数据对应的SOH值，其中圆点用于表征该数值属于异常SOH值。FIG. 5 graphically shows the divided normal values and abnormal values, wherein the horizontal axis represents time, the vertical axis represents the SOH value corresponding to the data, and the dots are used to represent that the value belongs to the abnormal SOH value.

然后，对异常SOH值进行修复，得到异常SOH值相应的修复SOH值。Then, repair the abnormal SOH value to obtain the repaired SOH value corresponding to the abnormal SOH value.

在一种可选的实施方式中，根据异常SOH值相邻的正常SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值。电池正常的SOH值分布在0.8～1之间，且随使用过程的变化趋势是一个缓慢下降的趋势。据此，在一种可选的实施方式中，可以选择该异常SOH值所在点的距离最近的一个正常SOH值作为该异常SOH值的修复值。In an optional embodiment, the abnormal SOH value is repaired according to the normal SOH value adjacent to the abnormal SOH value, and the repaired SOH value corresponding to the abnormal SOH value is obtained. The normal SOH value of the battery is distributed between 0.8 and 1, and the trend of change with the use process is a slow downward trend. Accordingly, in an optional implementation manner, a normal SOH value closest to the point where the abnormal SOH value is located may be selected as the repair value of the abnormal SOH value.

在该实施方式中，根据异常SOH值相邻的正常SOH值对异常SOH值进行修复，或者基于备用SOH值对异常SOH值进行修复，可以提高数据修复的可信度以及准确性，进而保证车用电池的健康状况评估的准确性。In this embodiment, the abnormal SOH value is repaired according to the normal SOH value adjacent to the abnormal SOH value, or the abnormal SOH value is repaired based on the backup SOH value, which can improve the reliability and accuracy of the data repair, thereby ensuring that the vehicle The accuracy of the battery health assessment.

在对异常SOH值进行修复之后，根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果。图6示出了修复后的数据的分布状况，其中，横轴表征时间，纵轴表征数据对应的SOH值。After repairing the abnormal SOH value, according to the distribution of the normal SOH value and the repaired SOH value, the health evaluation result of the vehicle battery is obtained. FIG. 6 shows the distribution of the repaired data, wherein the horizontal axis represents time, and the vertical axis represents the SOH value corresponding to the data.

进一步地，根据车用电池的健康状况评估结果对车用电池执行与健康状况评估结果相匹配的目标操作。修复了异常SOH值，才能还原最接近电池健康状态的情况，才能比较准确的判断电池是否可以继续使用。Further, a target operation matching the health evaluation result of the vehicle battery is performed on the vehicle battery according to the health state evaluation result of the vehicle battery. The abnormal SOH value has been repaired to restore the situation closest to the battery's health status, and to more accurately determine whether the battery can continue to be used.

比如，假如修复后的SOH分布，都分布在0.8～1之间，那么电池健康状态就正常，可以继续使用。如果出现大量低于0.8的SOH值，则该电池明显已经老化，可以弃用(报废)或者用作其他地方的电池(转用储能电池)，不再作为汽车电池使用。基于车用电池的健康状况评估结果对车用电池执行与健康状况评估结果相匹配的目标操作，既能够及时处理异常避免安全问题发生，又能够物尽其用避免资源浪费。For example, if the SOH distribution after repair is between 0.8 and 1, then the battery health status is normal and can be used continuously. If there is a large amount of SOH value below 0.8, the battery is obviously aged and can be discarded (scrapped) or used as a battery in other places (converted to an energy storage battery), and is no longer used as a car battery. Based on the health status evaluation result of the vehicle battery, the target operation that matches the health status evaluation result of the vehicle battery can be performed, which can not only deal with the abnormality in time to avoid the occurrence of safety problems, but also make the best use of it to avoid waste of resources.

本实施例的车用电池的健康状况评估方法自动从SOH数据集中划分出正常SOH值和异常SOH值，然后对异常SOH值进行修复，再基于修复后的数据的分布情况得到车用电池的健康状况评估结果，能够提高车用电池的健康状况评估的效率。而且对异常SOH 值进行修复后再进行健康状况评估，还能提高车用电池的健康状况评估结果的准确性，对车用电池的后续操作提供了较佳的参考。The method for evaluating the health status of a vehicle battery in this embodiment automatically divides the normal SOH value and the abnormal SOH value from the SOH data set, then repairs the abnormal SOH value, and then obtains the health status of the vehicle battery based on the distribution of the repaired data. The condition evaluation result can improve the efficiency of the health condition evaluation of the vehicle battery. Moreover, after repairing the abnormal SOH value and then performing the health evaluation, it can also improve the accuracy of the health evaluation result of the vehicle battery, and provide a better reference for the follow-up operation of the vehicle battery.

实施例2Example 2

本实施例提供一种车用电池的健康状况评估方法。该车用电池的健康状况评估方法与实施例1的车用电池的健康状况评估方法大致相同，区别在于对异常SOH值进行修复的步骤。This embodiment provides a method for evaluating the health status of a vehicle battery. The method for evaluating the health condition of the vehicle battery is substantially the same as the method for evaluating the health condition of the vehicle battery in Example 1, and the difference lies in the step of repairing the abnormal SOH value.

在一种可选的实施方式中，在进行修复时，获取异常SOH值对应的至少一个备用SOH值，并在备用SOH值为正常SOH值时，根据备用SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值；异常SOH值和对应的备用SOH值根据不同的计算方式得到。具体实施时，获取异常SOH值对应的备用SOH值；在备用SOH值为两个且均为正常SOH值时，获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值。In an optional embodiment, when repairing, obtain at least one backup SOH value corresponding to the abnormal SOH value, and when the backup SOH value is a normal SOH value, repair the abnormal SOH value according to the backup SOH value, and obtain The abnormal SOH value corresponds to the repaired SOH value; the abnormal SOH value and the corresponding backup SOH value are obtained according to different calculation methods. During specific implementation, obtain the backup SOH values corresponding to the abnormal SOH values; when there are two backup SOH values and both are normal SOH values, obtain the variation range between the two backup SOH values and the adjacent normal SOH values respectively; The smaller of the two changes is used as the corresponding repaired SOH value for the abnormal SOH value.

在一种可选的实施方式中，基于SOH1(即积分电量/额定电量*百分比)形成的SOH数据集对车用电池的健康状况进行评估，则以SOH2(即积分容量/额定容量*百分比)、SOH3(即积分容量/可用容量*百分比)作为备用SOH值。如果该异常SOH值对应的SOH2值和SOH3值均属于正常SOH值，获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值。具体实施时，计算该异常SOH值对应的SOH2值与其前一个正常的SOH2的变化幅度，记为diff2；同时计算该异常SOH值对应的SOH3与其前一个正常的SOH3的变化幅度，记为diff3。选取diff2、diff3中较小者对应的备用SOH值作为修复值。如果该异常SOH值对应的SOH2值和SOH3值中只有一个属于正常SOH值，则以该备用SOH值作为该异常SOH值相应的修复SOH值。如果该异常SOH值对应的SOH2值和SOH3值均属于异常SOH值，则不进行修复。In an optional embodiment, the health status of the vehicle battery is evaluated based on the SOH data set formed by SOH1 (that is, the integral capacity/rated capacity*percentage), and then the health status of the vehicle battery is evaluated by SOH2 (that is, the integral capacity/rated capacity*percentage) , SOH3 (ie integral capacity / available capacity * percentage) as the spare SOH value. If the SOH2 value and SOH3 value corresponding to the abnormal SOH value belong to the normal SOH value, obtain the variation range between the two standby SOH values and the adjacent normal SOH value; select the smaller of the two variation ranges and use Repair the SOH value corresponding to the abnormal SOH value. During specific implementation, calculate the variation range of the SOH2 value corresponding to the abnormal SOH value and its previous normal SOH2, denoted as diff2; at the same time, calculate the variation range of the SOH3 corresponding to the abnormal SOH value and its previous normal SOH3, denoted as diff3. The spare SOH value corresponding to the smaller of diff2 and diff3 is selected as the repair value. If only one of the SOH2 value and the SOH3 value corresponding to the abnormal SOH value belongs to the normal SOH value, the backup SOH value is used as the repaired SOH value corresponding to the abnormal SOH value. If both the SOH2 value and the SOH3 value corresponding to the abnormal SOH value belong to the abnormal SOH value, the repair will not be performed.

类似地，在其他可选的实施方式中，基于SOH2形成的SOH数据集对车用电池的健康状况进行评估，则以SOH1、SOH3作为备用SOH值；基于SOH3形成的SOH数据集对车用电池的健康状况进行评估，则以SOH2、SOH1作为备用SOH值。异常SOH值的修复方式参照以上说明，不再赘述。Similarly, in other optional embodiments, the health status of the vehicle battery is evaluated based on the SOH data set formed by SOH2, then SOH1 and SOH3 are used as the backup SOH values; the SOH data set formed based on SOH3 is used to evaluate the vehicle battery. For evaluation of the health status, SOH2 and SOH1 are used as the spare SOH values. Refer to the above description for the repairing method of the abnormal SOH value, which is not repeated here.

基于备用SOH值对异常SOH值进行修复，可以提高数据修复的准确性，进而保证车用电池的健康状况评估的准确性。Repairing the abnormal SOH value based on the spare SOH value can improve the accuracy of data repair, thereby ensuring the accuracy of the health status assessment of the vehicle battery.

在本实施例中，提供了基于异常SOH值的备用SOH值对异常SOH值进行修复的具体方式，在备用SOH值为两个且均为正常SOH值时，获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值，可以提高数据修复的准确性，进而保证车用电池的健康状况评估的准确性。In this embodiment, a specific method for repairing the abnormal SOH value based on the backup SOH value of the abnormal SOH value is provided. When the backup SOH values are two and both are normal SOH values, the two backup SOH values are obtained respectively with the corresponding SOH value. The change range between the adjacent normal SOH values; the smaller of the two change ranges is selected as the repaired SOH value corresponding to the abnormal SOH value, which can improve the accuracy of data repair and ensure the health assessment of the vehicle battery. accuracy.

实施例3Example 3

图7为本实施例提供的一种电子设备的结构示意图。所述电子设备包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述程序时实现实施例1或实施例2的车用电池的健康状况评估方法。图7显示的电子设备30仅仅是一个示例，不应对本发明实施例的功能和使用范围带来任何限制。FIG. 7 is a schematic structural diagram of an electronic device provided in this embodiment. The electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, when the processor executes the program, the vehicle battery health assessment of Embodiment 1 or Embodiment 2 is implemented method. The electronic device 30 shown in FIG. 7 is only an example, and should not impose any limitation on the function and scope of use of the embodiments of the present invention.

电子设备30可以以通用计算设备的形式表现，例如其可以为服务器设备。电子设备30的组件可以包括但不限于：上述至少一个处理器31、上述至少一个存储器32、连接不同***组件(包括存储器32和处理器31)的总线33。The electronic device 30 may take the form of a general-purpose computing device, which may be, for example, a server device. Components of the electronic device 30 may include, but are not limited to, the above-mentioned at least one processor 31 , the above-mentioned at least one memory 32 , and a bus 33 connecting different system components (including the memory 32 and the processor 31 ).

总线33包括数据总线、地址总线和控制总线。The bus 33 includes a data bus, an address bus and a control bus.

存储器32可以包括易失性存储器，例如随机存取存储器(RAM)321和/或高速缓存存储器322，还可以进一步包括只读存储器(ROM)323。 Memory 32 may include volatile memory, such as random access memory (RAM) 321 and/or cache memory 322 , and may further include read only memory (ROM) 323 .

存储器32还可以包括具有一组(至少一个)程序模块324的程序/实用工具325，这样的程序模块324包括但不限于：操作***、一个或者多个应用程序、其它程序模块以及程序数据，这些示例中的每一个或某种组合中可能包括网络环境的实现。The memory 32 may also include a program/utility 325 having a set (at least one) of program modules 324 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, which An implementation of a network environment may be included in each or some combination of the examples.

处理器31通过运行存储在存储器32中的计算机程序，从而执行各种功能应用以及数据处理，例如本发明实施例1或实施例2的车用电池的健康状况评估方法。The processor 31 executes various functional applications and data processing by running the computer program stored in the memory 32 , such as the method for evaluating the health condition of a vehicle battery according to Embodiment 1 or Embodiment 2 of the present invention.

电子设备30也可以与一个或多个外部设备34(例如键盘、指向设备等)通信。这种通信可以通过输入/输出(I/O)接口35进行。并且，模型生成的设备30还可以通过网络适配器36与一个或者多个网络(例如局域网(LAN)，广域网(WAN)和/或公共网络，例如因特网)通信。如图所示，网络适配器36通过总线33与模型生成的设备30的其它模块通信。应当明白，尽管图中未示出，可以结合模型生成的设备30使用其它硬件和/或软件模块，包括但不限于：微代码、设备驱动器、冗余处理器、外部磁盘驱动阵列、RAID(磁盘阵列)***、磁带驱动器以及数据备份存储***等。The electronic device 30 may also communicate with one or more external devices 34 (eg, keyboards, pointing devices, etc.). Such communication may take place through input/output (I/O) interface 35 . Also, the model-generating device 30 may also communicate with one or more networks (eg, a local area network (LAN), a wide area network (WAN), and/or a public network, such as the Internet) through a network adapter 36 . As shown, the network adapter 36 communicates with other modules of the model generation device 30 via the bus 33 . It should be understood that, although not shown, other hardware and/or software modules may be used in conjunction with the model-generated device 30, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk) array) systems, tape drives, and data backup storage systems.

应当注意，尽管在上文详细描述中提及了电子设备的若干单元/模块或子单元/模块，但是这种划分仅仅是示例性的并非强制性的。实际上，根据本发明的实施方式，上文描述的两个或更多单元/模块的特征和功能可以在一个单元/模块中具体化。反之，上文描述的一个单元/模块的特征和功能可以进一步划分为由多个单元/模块来具体化。It should be noted that although several units/modules or sub-units/modules of the electronic device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. Indeed, the features and functions of two or more units/modules described above may be embodied in one unit/module according to embodiments of the present invention. Conversely, the features and functions of one unit/module described above may be further subdivided to be embodied by multiple units/modules.

实施例4Example 4

本实施例提供了一种计算机可读存储介质，其上存储有计算机程序，所述程序被处理器执行时实现实施例1或实施例2的车用电池的健康状况评估方法的步骤。This embodiment provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, implements the steps of the method for evaluating the health status of a vehicle battery in Embodiment 1 or Embodiment 2.

其中，可读存储介质可以采用的更具体可以包括但不限于：便携式盘、硬盘、随机存取存储器、只读存储器、可擦拭可编程只读存储器、光存储器件、磁存储器件或上述的任意合适的组合。Wherein, the readable storage media may include, but are not limited to, portable disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, optical storage devices, magnetic storage devices, or any of the above suitable combination.

在可能的实施方式中，本发明还可以实现为一种程序产品的形式，其包括程序代码，当所述程序产品在终端设备上运行时，所述程序代码用于使所述终端设备执行实现实施例1或实施例2的车用电池的健康状况评估方法的步骤。In a possible implementation manner, the present invention can also be implemented in the form of a program product, which includes program codes, when the program product runs on a terminal device, the program code is used to cause the terminal device to execute the implementation Steps of the method for evaluating the state of health of a vehicle battery according to Embodiment 1 or Embodiment 2.

其中，可以以一种或多种程序设计语言的任意组合来编写用于执行本发明的程序代码，所述程序代码可以完全地在用户设备上执行、部分地在用户设备上执行、作为一个独立的软件包执行、部分在用户设备上部分在远程设备上执行或完全在远程设备上执行。Wherein, the program code for executing the present invention can be written in any combination of one or more programming languages, and the program code can be completely executed on the user equipment, partially executed on the user equipment, as an independent The software package executes on the user's device, partly on the user's device, partly on the remote device, or entirely on the remote device.

实施例5Example 5

本实施例提供一种车用电池的健康状况评估***。参照图8，该车用电池的健康状况评估***包括获取单元201、划分单元202、修复单元203、评估单元204。This embodiment provides a system for evaluating the health status of a vehicle battery. Referring to FIG. 8 , the system for evaluating the state of health of a vehicle battery includes an acquiring unit 201 , a dividing unit 202 , a repairing unit 203 , and an evaluating unit 204 .

获取单元201用于获取车用电池的SOH数据集；划分单元202用于从SOH数据集中划分出正常SOH值和异常SOH值；修复单元203用于对异常SOH进行修复，得到异常SOH值相应的修复SOH值；评估单元204用于根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果。The obtaining unit 201 is used to obtain the SOH data set of the vehicle battery; the dividing unit 202 is used to divide the normal SOH value and the abnormal SOH value from the SOH data set; Repairing the SOH value; the evaluation unit 204 is configured to obtain the health condition evaluation result of the vehicle battery according to the distribution of the normal SOH value and the repaired SOH value.

具体实施时，首先，获取单元201获取车用电池的SOH数据集。在一块电池的SOH数据集中，正常情况下，是由大部分正常的SOH值和部分异常的SOH值组成。正常的SOH值具备相似的特征，异常的SOH值的特征跟正常的SOH值的特征相差甚大。图2示出了一个SOH数据集的各个SOH值的分布情况，其中，横轴表征SOH值，纵轴表征对应的SOH值的数据的数量。该SOH数据集大量的SOH值处于0.8-1.0之间，符合正常的SOH值的特点；其中少量数据分布与0-0.8之间。可以判断，这部分少量数据，其中存在一些异常值。During specific implementation, first, the acquiring unit 201 acquires the SOH data set of the vehicle battery. In the SOH data set of a battery, under normal circumstances, it consists of most of the normal SOH values and some abnormal SOH values. Normal SOH values have similar characteristics, and the characteristics of abnormal SOH values are quite different from those of normal SOH values. FIG. 2 shows the distribution of each SOH value of an SOH data set, wherein the horizontal axis represents the SOH value, and the vertical axis represents the number of data corresponding to the SOH value. A large number of SOH values in this SOH data set are between 0.8-1.0, which is in line with the characteristics of normal SOH values; a small amount of data is distributed between 0-0.8. It can be judged that there are some outliers in this small amount of data.

为了识别异常值，首先，划分单元202获取SOH数据集中各SOH值对应的降维分析数据，得到SOH数据集对应的分析数据集。In order to identify outliers, first, the dividing unit 202 obtains the dimensionality reduction analysis data corresponding to each SOH value in the SOH data set, and obtains the analysis data set corresponding to the SOH data set.

然后，划分单元202确定分析数据集的聚类数量以及主成分。Then, the dividing unit 202 determines the number of clusters and principal components of the analysis data set.

可实施地，可通过任意能够确定数据集的聚类数量以及主成分的算法来确定分析数据集的聚类数量以及主成分，本实施例并不具体限定，该算法可根据实际需求进行相应的选择及调整。Implementably, the number of clusters and the principal components of the analysis data set can be determined by any algorithm that can determine the number of clusters and the principal components of the data set. This embodiment is not specifically limited, and the algorithm can be adapted according to actual needs. Select and adjust.

例如划分单元202可基于elbow(肘系数)方法来确定聚类数量。图3示出了一elbow曲线，当聚类数量趋于6时，曲线开始趋于收敛，此时可以取聚类数量为6。For example, the dividing unit 202 may determine the number of clusters based on an elbow (elbow coefficient) method. Figure 3 shows an elbow curve, when the number of clusters tends to 6, the curve begins to converge, and the number of clusters can be taken as 6 at this time.

然后，划分单元202基于主成分分析法分析数据集中的主成分及每个主成分的解释方差率。图4示出了SOH数据集的累积解释方差和独立的解释方差，其中，横轴表征主成分，纵轴表征解释方差率。根据图4所示，数据集中共有两个主成分，每个主成分大概解释了一半的方差，故两个主成分均需要保留。Then, the dividing unit 202 analyzes the principal components in the data set and the explained variance ratio of each principal component based on the principal component analysis method. Figure 4 shows the cumulative explained variance and the independent explained variance for the SOH dataset, where the horizontal axis represents the principal components and the vertical axis represents the explained variance rate. As shown in Figure 4, there are two principal components in the data set, and each principal component explains about half of the variance, so both principal components need to be retained.

确定了聚类数量，确定了主成分个数，接下来则划分单元202进行数据集的标准化处理并进行KMeans聚类。聚类的过程包括：选取聚类数量的且包括主成分的聚类中心点；以聚类中心点为聚类中心进行聚类，在聚类后重新选取每类的聚类中心点且保持主成分为聚类中心点再继续聚类直至收敛。After the number of clusters is determined, the number of principal components is determined, and then the dividing unit 202 performs normalization processing of the data set and KMeans clustering. The process of clustering includes: selecting the number of cluster center points and including the principal components; clustering with the cluster center points as the cluster centers, and re-selecting the cluster center points of each type after clustering and keeping the main points. The components are cluster center points and continue to cluster until convergence.

聚类完成后，划分单元202从SOH数据集中划分出正常SOH值和异常SOH值。After the clustering is completed, the dividing unit 202 divides the normal SOH value and the abnormal SOH value from the SOH data set.

在第一种可选的实施方式中，划分单元202获取一个降维分析数据与所对应的聚类中心之间的距离，并将对应的距离最大的多个降维分析数据为异常降维分析数据，以及将异常降维分析数据对应的SOH值识别为异常SOH值。In a first optional embodiment, the dividing unit 202 obtains the distance between one dimension reduction analysis data and the corresponding cluster center, and regards a plurality of dimension reduction analysis data with the largest corresponding distances as abnormal dimension reduction analysis data, and identify the SOH value corresponding to the abnormal dimensionality reduction analysis data as the abnormal SOH value.

在第二种可选的实施方式中，划分单元202根据异常比例从每类聚类数据中筛选异常降维分析数据，并将异常降维分析数据对应的SOH值识别为异常SOH值。例如，在一种情形下，设定一个异常值的比例outliers_fraction为1％，这样设置是因为在标准正太分布的情况下(N(0,1))，一般认定3个标准差以外的数据为异常值。3个标准差以内的数据包含了数据集中99％以上的数据，所以剩下的1％的数据可以视为异常值。In a second optional embodiment, the dividing unit 202 screens the abnormal dimensionality reduction analysis data from each type of cluster data according to the abnormality ratio, and identifies the SOH value corresponding to the abnormal dimensionality reduction analysis data as the abnormal SOH value. For example, in one case, the proportion of outliers_fraction of an outlier is set to 1%. This setting is because in the case of a standard normal distribution (N(0,1)), it is generally considered that data beyond 3 standard deviations are outliers. Data within 3 standard deviations contains more than 99% of the data in the dataset, so the remaining 1% of data can be considered outliers.

图5以图示的方式示出了划分的正常值和异常值，其中，横轴“Data Time Integer”表征时间，纵轴“SOH1”表征数据对应的SOH值。FIG. 5 graphically shows the divided normal values and abnormal values, wherein the horizontal axis "Data Time Integer" represents time, and the vertical axis "SOH1" represents the SOH value corresponding to the data.

然后，修复单元203对异常SOH值进行修复，得到异常SOH值相应的修复SOH值。Then, the repairing unit 203 repairs the abnormal SOH value to obtain a repaired SOH value corresponding to the abnormal SOH value.

在一种可选的实施方式中，修复单元203根据异常SOH值相邻的正常SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值。电池正常的SOH值分布在0.8～1之间，且随使用过程的变化趋势是一个缓慢下降的趋势。据此，在一种可选的实施方式中，可以选择该异常SOH值所在点的距离最近的一个正常SOH值作为该异常SOH值的修复值。In an optional embodiment, the repairing unit 203 repairs the abnormal SOH value according to the normal SOH value adjacent to the abnormal SOH value, to obtain a repaired SOH value corresponding to the abnormal SOH value. The normal SOH value of the battery is between 0.8 and 1, and the trend of change with the use process is a slow downward trend. Accordingly, in an optional implementation manner, a normal SOH value that is closest to the point where the abnormal SOH value is located may be selected as the repair value of the abnormal SOH value.

在对异常SOH值进行修复之后，评估单元204根据正常SOH值和修复SOH值的分布，得到车用电池的健康状况评估结果。图6示出了修复后的数据的分布状况，其中，横轴表征时间，纵轴表征数据对应的SOH值。After repairing the abnormal SOH value, the evaluation unit 204 obtains the health condition evaluation result of the vehicle battery according to the distribution of the normal SOH value and the repaired SOH value. FIG. 6 shows the distribution of the repaired data, wherein the horizontal axis represents time, and the vertical axis represents the SOH value corresponding to the data.

进一步地，评估单元204根据车用电池的健康状况评估结果对车用电池执行与健康状况评估结果相匹配的目标操作。修复了异常SOH值，才能还原最接近电池健康状态的情况，才能比较准确的判断电池是否可以继续使用。比如，假如修复后的SOH分布，都分布在0.8～1之间，那么电池健康状态就正常，可以继续使用。如果出现大量低于0.8的SOH值，则该电池明显已经老化，可以弃用或者用作其他地方的电池，不再作为汽车电池使用。基于车用电池的健康状况评估结果对车用电池执行与健康状况评估结果相匹配的目标操作，既能够及时处理异常避免安全问题发生，又能够物尽其用避免资源浪费。Further, the evaluation unit 204 performs a target operation on the vehicle battery that matches the state of health evaluation result of the vehicle battery according to the evaluation result of the state of health of the vehicle battery. The abnormal SOH value has been repaired to restore the situation closest to the battery's health status, and to more accurately determine whether the battery can continue to be used. For example, if the SOH distribution after repair is between 0.8 and 1, then the battery health status is normal and can be used continuously. If there is a large number of SOH values below 0.8, the battery is obviously aged and can be discarded or used as a battery in other places, and is no longer used as a car battery. Based on the health status evaluation result of the vehicle battery, the target operation that matches the health status evaluation result of the vehicle battery can be performed, which can not only deal with the abnormality in time to avoid the occurrence of safety problems, but also make the best use of it to avoid waste of resources.

本实施例的车用电池的健康状况评估***自动从SOH数据集中划分出正常SOH值和异常SOH值，然后对异常SOH值进行修复，再基于修复后的数据的分布情况得到车用电池的健康状况评估结果，能够提高车用电池的健康状况评估的效率。而且对异常SOH值进行修复后再进行健康状况评估，还能提高车用电池的健康状况评估结果的准确性，对车用电池的后续操作提供了较佳的参考。The vehicle battery health evaluation system of this embodiment automatically divides the normal SOH value and the abnormal SOH value from the SOH data set, then repairs the abnormal SOH value, and then obtains the health status of the vehicle battery based on the distribution of the repaired data. The condition evaluation result can improve the efficiency of the health condition evaluation of the vehicle battery. Moreover, after repairing the abnormal SOH value, the health status assessment can also improve the accuracy of the health status assessment result of the vehicle battery, and provide a better reference for the subsequent operation of the vehicle battery.

实施例6Example 6

本实施例提供一种车用电池的健康状况评估***。该车用电池的健康状况评估***与实施例5的车用电池的健康状况评估***大致相同，区别在于修复单元203对异常SOH值进行修复的过程。This embodiment provides a system for evaluating the health status of a vehicle battery. The system for evaluating the health condition of a vehicle battery is substantially the same as the system for evaluating the health condition of the vehicle battery in Embodiment 5, and the difference lies in the process of repairing the abnormal SOH value by the repairing unit 203 .

在一种可选的实施方式中，在进行修复时，修复单元203获取异常SOH值对应的至少一个备用SOH值，并在备用SOH值为正常SOH值时，根据备用SOH值对异常SOH值进行修复，得到异常SOH值相应的修复SOH值；异常SOH值和对应的备用SOH值根据不同的计算方式得到。具体实施时，修复单元203获取异常SOH值对应的备用SOH值；在备用SOH值为两个且均为正常SOH值时，修复单元203获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；修复单元203选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值。In an optional implementation manner, during repairing, the repairing unit 203 obtains at least one backup SOH value corresponding to the abnormal SOH value, and when the backup SOH value is a normal SOH value, performs the abnormal SOH value according to the backup SOH value. Repair, and obtain the repaired SOH value corresponding to the abnormal SOH value; the abnormal SOH value and the corresponding spare SOH value are obtained according to different calculation methods. During specific implementation, the repair unit 203 obtains the backup SOH values corresponding to the abnormal SOH values; when the backup SOH values are two and both are normal SOH values, the repair unit 203 obtains the difference between the two backup SOH values and the adjacent normal SOH values respectively The change range between the two; the repair unit 203 selects the smaller of the two change ranges as the corresponding repair SOH value of the abnormal SOH value.

在一种可选的实施方式中，基于SOH1(即积分电量/额定电量*百分比)形成的SOH数据集对车用电池的健康状况进行评估，则以SOH2(即积分容量/额定容量*百分比)、SOH3(即积分容量/可用容量*百分比)作为备用SOH值。如果该异常SOH值对应的SOH2值和SOH3值均属于正常SOH值，修复单元203获取两个备用SOH值分别与相邻的正常SOH值之间的变化幅度；修复单元203选取两个变化幅度中的较小者，用作异常SOH值相应的修复SOH值。具体实施时，修复单元203计算该异常SOH值对应的SOH2值与其前一个正常的SOH2的变化幅度，记为diff2；同时，修复单元203计算该异常SOH值对应的SOH3与其前一个正常的SOH3的变化幅度，记为diff3。修复单元203选取diff2、diff3中较小者对应的备用SOH值作为修复值。如果该异常SOH值对应的SOH2值和SOH3值中只有一个属于正常SOH值，则修复单元203以该备用SOH值作为该异常SOH值相应的修复SOH值。如果该异常SOH值对应的SOH2值和SOH3值均属于异常SOH值，则修复单元203不进行修复。In an optional embodiment, the health status of the vehicle battery is evaluated based on the SOH data set formed by SOH1 (that is, the integral capacity/rated capacity*percentage), and then the health status of the vehicle battery is evaluated by SOH2 (that is, the integral capacity/rated capacity*percentage) , SOH3 (ie integral capacity / available capacity * percentage) as the spare SOH value. If both the SOH2 value and the SOH3 value corresponding to the abnormal SOH value belong to the normal SOH value, the repair unit 203 obtains the variation range between the two standby SOH values and the adjacent normal SOH value respectively; the restoration unit 203 selects one of the two variation ranges The smaller one is used as the corresponding repaired SOH value for the abnormal SOH value. During specific implementation, the repairing unit 203 calculates the variation range between the SOH2 value corresponding to the abnormal SOH value and its previous normal SOH2, which is denoted as diff2; at the same time, the repairing unit 203 calculates the difference between the SOH3 corresponding to the abnormal SOH value and its previous normal SOH3 The magnitude of change, denoted as diff3. The repair unit 203 selects the spare SOH value corresponding to the smaller of diff2 and diff3 as the repair value. If only one of the SOH2 value and the SOH3 value corresponding to the abnormal SOH value belongs to a normal SOH value, the repair unit 203 uses the spare SOH value as the repaired SOH value corresponding to the abnormal SOH value. If both the SOH2 value and the SOH3 value corresponding to the abnormal SOH value belong to the abnormal SOH value, the repairing unit 203 does not repair.

类似地，在其他可选的实施方式中，基于SOH2形成的SOH数据集对车用电池的健康状况进行评估，则以SOH1、SOH3作为备用SOH值；基于SOH3形成的SOH数据集对车用电池的健康状况进行评估，则以SOH2、SOH1作为备用SOH值。异常SOH值的修复方式参照以上说明，不再赘述。Similarly, in other optional embodiments, the health status of the vehicle battery is evaluated based on the SOH data set formed by SOH2, then SOH1 and SOH3 are used as the backup SOH values; the SOH data set formed based on SOH3 is used to evaluate the vehicle battery. For evaluation of the health status, SOH2 and SOH1 are used as the spare SOH values. Refer to the above description for the repairing method of the abnormal SOH value, which will not be repeated here.

虽然以上描述了本发明的具体实施方式，但是本领域的技术人员应当理解，这仅是举例说明，本发明的保护范围是由所附权利要求书限定的。本领域的技术人员在不背离本发明的原理和实质的前提下，可以对这些实施方式做出多种变更或修改，但这些变更和修改均落入本发明的保护范围。Although the specific embodiments of the present invention are described above, those skilled in the art should understand that this is only an illustration, and the protection scope of the present invention is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principle and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.

Claims

一种车用电池的健康状况评估方法，其特征在于，包括以下步骤：A method for evaluating the health status of a vehicle battery, comprising the following steps:

获取车用电池的SOH数据集；Obtain the SOH dataset of vehicle batteries;

从所述SOH数据集中划分出正常SOH值和异常SOH值；dividing normal SOH values and abnormal SOH values from the SOH data set;

对所述异常SOH进行修复，得到所述异常SOH值相应的修复SOH值；Repairing the abnormal SOH to obtain a repaired SOH value corresponding to the abnormal SOH value;

根据所述正常SOH值和所述修复SOH值的分布，得到所述车用电池的健康状况评估结果。According to the distribution of the normal SOH value and the repaired SOH value, a health condition evaluation result of the vehicle battery is obtained.
如权利要求1所述的车用电池的健康状况评估方法，其特征在于，所述从所述SOH数据集中划分出正常SOH值和异常SOH值之前，所述车用电池的健康状况评估方法还包括：The method for evaluating the state of health of a vehicle battery according to claim 1, wherein before dividing the normal SOH value and the abnormal SOH value from the SOH data set, the method for evaluating the state of health of the vehicle battery further comprises: include:

获取所述SOH数据集中各SOH值对应的降维分析数据，得到所述SOH数据集对应的分析数据集；Obtain the dimensionality reduction analysis data corresponding to each SOH value in the SOH data set, and obtain the analysis data set corresponding to the SOH data set;

确定所述分析数据集的聚类数量以及主成分；determining the number of clusters and principal components of the analysis data set;

根据所述聚类数量和所述主成分对所述分析数据集进行聚类。The analysis data set is clustered according to the number of clusters and the principal components.
如权利要求2所述的车用电池的健康状况评估方法，其特征在于，所述确定所述分析数据集的聚类数量以及主成分，包括：The method for evaluating the health condition of a vehicle battery according to claim 2, wherein the determining the number of clusters and principal components of the analysis data set includes:

根据所述分析数据集中的降维分析数据，筛选候选成分；Screening candidate components according to the dimensionality reduction analysis data in the analysis data set;

根据所述候选成分的解释方差率，从所述候选成分中筛选主成分；screening principal components from the candidate components according to the explained variance ratio of the candidate components;

所述根据所述聚类数量和所述主成分对所述分析数据集进行聚类，包括：The clustering of the analysis data set according to the number of clusters and the principal components includes:

选取所述聚类数量的且包括所述主成分的聚类中心点；Selecting the cluster center point of the number of clusters and including the principal component;

以所述聚类中心点为聚类中心进行聚类，在聚类后重新选取每类的聚类中心点且保持所述主成分为聚类中心点再继续聚类直至收敛。Clustering is performed with the cluster center point as the cluster center, and after clustering, the cluster center point of each type is reselected and the principal component is kept as the cluster center point, and then the clustering is continued until convergence.
如权利要求2-3中至少一项所述的车用电池的健康状况评估方法，其特征在于，所述从所述SOH数据集中划分出正常SOH值和异常SOH值，包括：The method for evaluating the health status of a vehicle battery according to at least one of claims 2-3, wherein the dividing the normal SOH value and the abnormal SOH value from the SOH data set includes:

获取一个所述降维分析数据与所对应的聚类中心之间的距离，并将对应的距离最大的多个所述降维分析数据为异常降维分析数据，以及将所述异常降维分析数据对应的SOH值识别为异常SOH值；Obtain the distance between one of the dimensionality reduction analysis data and the corresponding cluster center, and use a plurality of the dimensionality reduction analysis data with the largest corresponding distances as abnormal dimensionality reduction analysis data, and analyze the abnormal dimensionality reduction analysis. The SOH value corresponding to the data is identified as an abnormal SOH value;

或者，or,

根据异常比例从每类聚类数据中筛选异常降维分析数据，并将所述异常降维分析数据对应的SOH值识别为异常SOH值。The abnormal dimensionality reduction analysis data is screened from each type of cluster data according to the abnormality ratio, and the SOH value corresponding to the abnormal dimensionality reduction analysis data is identified as the abnormal SOH value.
如权利要求1-4中至少一项所述的车用电池的健康状况评估方法，其特征在于，所述对所述异常SOH值进行修复，得到所述异常SOH值相应的修复SOH值，包括：The method for evaluating the health status of a vehicle battery according to at least one of claims 1 to 4, wherein the repairing the abnormal SOH value to obtain the repaired SOH value corresponding to the abnormal SOH value, comprising: :

根据所述异常SOH值相邻的正常SOH值对所述异常SOH值进行修复，得到所述异常SOH值相应的修复SOH值；Repair the abnormal SOH value according to the normal SOH value adjacent to the abnormal SOH value, and obtain the repaired SOH value corresponding to the abnormal SOH value;

或者，or,

获取所述异常SOH值对应的至少一个备用SOH值，并在所述备用SOH值为正常SOH值时，根据所述备用SOH值对所述异常SOH值进行修复，得到所述异常SOH值相应的修复SOH值；所述异常SOH值和对应的备用SOH值根据不同的计算方式得到。Obtain at least one backup SOH value corresponding to the abnormal SOH value, and when the backup SOH value is a normal SOH value, repair the abnormal SOH value according to the backup SOH value, and obtain a corresponding value of the abnormal SOH value. Repair the SOH value; the abnormal SOH value and the corresponding backup SOH value are obtained according to different calculation methods.
如权利要求5所述的车用电池的健康状况评估方法，其特征在于，所述获取所述异常SOH值对应的至少一个备用SOH值，并在所述备用SOH值为正常SOH值时，根据所述备用SOH值对所述异常SOH值进行修复，得到所述异常SOH值相应的修复SOH值，包括：The method for evaluating the health status of a vehicle battery according to claim 5, wherein the acquiring at least one backup SOH value corresponding to the abnormal SOH value, and when the backup SOH value is a normal SOH value, according to The standby SOH value repairs the abnormal SOH value, and obtains a repaired SOH value corresponding to the abnormal SOH value, including:

获取所述异常SOH值对应的备用SOH值；Obtain the standby SOH value corresponding to the abnormal SOH value;

在所述备用SOH值为两个且均为正常SOH值时，获取两个所述备用SOH值分别与相邻的正常SOH值之间的变化幅度；When the backup SOH values are two and both are normal SOH values, obtain the variation range between the two backup SOH values and the adjacent normal SOH values respectively;

选取两个所述变化幅度中的较小者，用作所述异常SOH值相应的修复SOH值。The smaller of the two variation ranges is selected to be used as the repaired SOH value corresponding to the abnormal SOH value.
如权利要求1-6中至少一项所述的车用电池的健康状况评估方法，其特征在于，所述根据所述正常SOH值和所述修复SOH值的分布，得到所述车用电池的健康状况评估结果之后，还包括：The method for evaluating the health condition of a vehicle battery according to at least one of claims 1 to 6, wherein, according to the distribution of the normal SOH value and the repaired SOH value, obtaining the health status of the vehicle battery Following the results of the health assessment, include:

对所述车用电池执行与所述健康状况评估结果相匹配的目标操作。A target operation matching the health condition evaluation result is performed on the vehicle battery.
一种车用电池的健康状况评估***，其特征在于，包括获取单元、划分单元、修复单元、评估单元；A system for evaluating the health condition of a vehicle battery, characterized in that it includes an acquisition unit, a division unit, a repair unit, and an evaluation unit;

所述获取单元用于获取车用电池的SOH数据集；The obtaining unit is used to obtain the SOH data set of the vehicle battery;

所述划分单元用于从所述SOH数据集中划分出正常SOH值和异常SOH值；The dividing unit is used to divide the normal SOH value and the abnormal SOH value from the SOH data set;

所述修复单元用于对所述异常SOH进行修复，得到所述异常SOH值相应的修复SOH值；The repairing unit is used to repair the abnormal SOH to obtain a repaired SOH value corresponding to the abnormal SOH value;

所述评估单元用于根据所述正常SOH值和所述修复SOH值的分布，得到所述车用电池的健康状况评估结果。The evaluation unit is configured to obtain a health condition evaluation result of the vehicle battery according to the distribution of the normal SOH value and the repaired SOH value.
一种电子设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1-7中至少一项所述的车用电池的健康状况评估方法。An electronic device, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, when the processor executes the computer program, at least one of claims 1-7 is implemented The described method for evaluating the health status of a vehicle battery.
一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1-7中至少一项所述的车用电池的健康状况评估方法的步骤。A computer-readable storage medium on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the method for evaluating the health status of a vehicle battery according to at least one of claims 1-7 is implemented. step.