Vehicle battery abnormity identification method and device, storage medium and electronic equipment

Through machine learning models, the fault identification of multi-frame data of vehicle batteries is solved by combining the characteristics of voltage, current, temperature, etc., and the problem that existing battery monitoring methods rely on manual inspection is solved, achieving efficient and accurate battery fault monitoring.

CN118938033APending Publication Date: 2024-11-12HEFEI GUOXUAN HIGH TECH POWER ENERGY
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411289694.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-13
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Existing battery monitoring methods rely on regular maintenance and manual inspection, which is time-consuming and labor-intensive and vulnerable to human factors, resulting in low warning accuracy.

Method used

By obtaining multi-frame data of vehicle batteries, the battery abnormality identification model based on machine learning is used to identify faults, and combining the derivative characteristics of voltage, current, temperature and residual state of charge, the fault category and fault occurrence probability are identified to achieve automated battery failures monitor.

Benefits of technology

It improves the accuracy and reliability of battery fault identification, reduces manual intervention, and improves battery monitoring efficiency and safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118938033A_ABST
    Figure CN118938033A_ABST
Patent Text Reader

Abstract

The invention discloses a vehicle battery abnormity identification method and device, a storage medium and electronic equipment. Relates to the technical field of batteries. The method comprises the following steps: acquiring multi-frame battery data of a vehicle battery correspondingly collected at multiple sampling moments in a current time period; based on the multiple frames of battery data, adopting a battery abnormity identification model to obtain initial fault identification results respectively corresponding to the multiple frames of battery data; based on the initial fault recognition results corresponding to the multiple frames of battery data, a battery fault recognition result of the vehicle battery is obtained, and the battery fault recognition result is at least used for indicating the fault type and the fault occurrence probability of the vehicle battery. According to the invention, the technical problem that the early warning accuracy is not high due to the fact that a battery monitoring method generally depends on regular maintenance and manual inspection and is time-consuming and labor-consuming and is easily influenced by human factors in the prior art is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of battery technology, and in particular to a vehicle battery abnormality identification method, device, storage medium and electronic equipment. Background Art

[0002] With the popularity of electric vehicles, the battery, as one of its core components, is of vital importance to vehicle performance and safety. However, due to the complexity of the battery itself and the variability of the use environment, battery failures are often difficult to detect and handle in a timely manner. Battery monitoring methods in related technologies usually rely on regular maintenance and manual inspections, which are not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in low accuracy of early warnings.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a vehicle battery abnormality identification method, device, storage medium and electronic device to at least solve the technical problem that the battery monitoring method in the related art usually relies on regular maintenance and manual inspection, which is not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in low warning accuracy.

[0005] According to one aspect of an embodiment of the present invention, a vehicle battery abnormality identification method is provided, comprising: obtaining multiple frames of battery data of a vehicle battery collected corresponding to multiple sampling moments within a current time period; based on the multiple frames of battery data, using a battery abnormality identification model to obtain initial fault identification results corresponding to the multiple frames of battery data, wherein the battery abnormality identification model is based on multiple groups of historical battery data and historical fault identification results corresponding to the multiple groups of historical battery data, obtained through machine learning; based on the initial fault identification results corresponding to the multiple frames of battery data, a battery fault identification result of the vehicle battery is obtained, wherein the battery fault identification result is at least used to indicate the fault category and fault occurrence probability of the vehicle battery.

[0006] Optionally, the method of obtaining multiple frames of battery data of the vehicle battery collected in the current time period includes: obtaining battery data corresponding to any sampling moment among the multiple sampling moments in the following manner; obtaining initial battery data of the vehicle battery at any sampling moment, wherein the initial battery data includes at least: voltages corresponding to multiple battery cells included in the vehicle battery, currents corresponding to the multiple battery cells, and a remaining state of charge of the vehicle battery; based on the initial battery data, determining derived features of the vehicle battery in multiple feature dimensions; based on the derived features, obtaining battery data corresponding to any sampling moment; and obtaining the multiple frames of battery data by adopting the method of obtaining battery data corresponding to any sampling moment.

[0007] Through the above methods, other key potential dimensional features can be derived, and the factors considered are more comprehensive, which can improve the richness and accuracy of the battery anomaly identification model input.

[0008] Optionally, based on the initial battery data, the derived characteristics of the vehicle battery in multiple characteristic dimensions are determined, including: based on the voltages corresponding to the multiple battery cells in the initial battery data, and the currents corresponding to the multiple battery cells, the statistical characteristics of the vehicle battery at any sampling moment are determined, wherein the statistical characteristics include voltage statistical characteristics and temperature statistical characteristics; the statistical characteristics at any sampling moment and the statistical characteristics at the previous sampling moment are differentially calculated to obtain the time differential characteristics of the vehicle battery at any sampling moment, wherein the previous sampling moment is the sampling moment before any sampling moment; based on the remaining state of charge of the vehicle battery in the initial battery data, the undervoltage risk degree of the vehicle battery at any sampling moment is determined; based on the statistical characteristics at any sampling moment, the time differential characteristics at any sampling moment, and the undervoltage risk degree at any sampling moment, the derived characteristics are obtained.

[0009] Through the above methods, the information on battery voltage, current, temperature, remaining state of charge and other aspects as well as the changing trends can be comprehensively considered to obtain more comprehensive and accurate derived characteristics, thereby achieving more comprehensive monitoring and management of the battery status and improving battery performance and safety.

[0010] Optionally, when the initial fault identification result includes the fault category and fault occurrence probability of the corresponding frame battery data, the battery fault identification result of the vehicle battery is obtained based on the initial fault identification results corresponding to the multiple frames of battery data, including: determining the number of battery data frames corresponding to multiple fault categories based on the fault categories corresponding to the multiple frames of battery data; determining a target fault category with the largest number of battery data frames from the multiple fault categories based on the number of battery data frames corresponding to the multiple fault categories; determining the fault occurrence probability of the target fault category; and obtaining the battery fault identification result based on the target fault category and the fault occurrence probability of the target fault category.

[0011] The above method can help determine the most frequently occurring fault category in battery data and evaluate its probability of occurrence, thereby achieving accurate identification of battery faults.

[0012] Optionally, determining the failure probability of the target fault category includes: determining a plurality of frames of target battery data corresponding to the target fault category from the plurality of frames of battery data, wherein the number of the plurality of frames of target battery data is less than or equal to the number of the plurality of frames of battery data; performing mean calculation on the failure probabilities corresponding to the plurality of frames of target battery data respectively to obtain an average failure probability; and using the average failure probability as the failure probability of the target fault category.

[0013] Through the above methods, the occurrence probability of the target fault category can be evaluated more accurately, providing more reliable data support for fault identification and processing.

[0014] Optionally, the battery fault identification result is obtained based on the target fault category and the fault occurrence probability of the target fault category, including: determining whether the number of battery data frames corresponding to the target fault category is greater than a predetermined number of frames; when the number of battery data frames corresponding to the target fault category is greater than the predetermined number of frames, obtaining the battery fault identification result based on the target fault category and the fault occurrence probability of the target fault category.

[0015] Through the above method, the problem of inaccurate fault identification caused by single-frame data errors can be avoided, and the accuracy and reliability of vehicle battery fault identification can be improved.

[0016] Optionally, before obtaining initial fault identification results corresponding to the multiple frames of battery data respectively by using a battery abnormality identification model based on the multiple frames of battery data, the method also includes: dividing the multiple groups of historical battery data, and the historical fault identification results corresponding to the multiple groups of historical battery data respectively, into a training data set and a test data set, wherein the multiple groups of historical battery data include battery data of multiple vehicles during a battery failure period, and battery data of the multiple vehicles during a normal operating period, and the normal operating period is a period of a predetermined length of time after the battery failure of the corresponding vehicle is repaired; training multiple initial models based on the training data set to obtain multiple training models, wherein the multiple initial models are constructed based on different algorithms; testing the multiple training models based on the test data set to obtain test results corresponding to the multiple training models respectively; and determining the battery abnormality identification model from the multiple training models based on the test results corresponding to the multiple training models respectively.

[0017] Through the above methods, training and screening of multiple models based on the same data can not only improve the efficiency of model training, but also screen out the model with the best performance.

[0018] According to another aspect of an embodiment of the present invention, a vehicle battery abnormality identification device is also provided, including: a data acquisition module, used to obtain multiple frames of battery data of a vehicle battery collected corresponding to multiple sampling moments within a current time period; an initial fault identification module, used to obtain initial fault identification results corresponding to the multiple frames of battery data based on the multiple frames of battery data and using a battery abnormality identification model, wherein the battery abnormality identification model is based on multiple groups of historical battery data and historical fault identification results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; a battery fault identification module, used to obtain a battery fault identification result of the vehicle battery based on the initial fault identification results corresponding to the multiple frames of battery data, wherein the battery fault identification result is at least used to indicate the fault type and fault occurrence probability of the vehicle battery.

[0019] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided, wherein the non-volatile storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing any one of the vehicle battery abnormality identification methods.

[0020] According to another aspect of an embodiment of the present invention, an electronic device is also provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the vehicle battery abnormality identification methods.

[0021] In an embodiment of the present invention, multiple frames of battery data of a vehicle battery collected at multiple sampling moments within a current time period are obtained; based on the multiple frames of battery data, a battery abnormality recognition model is used to obtain initial fault recognition results corresponding to the multiple frames of battery data, wherein the battery abnormality recognition model is based on multiple groups of historical battery data and historical fault recognition results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; based on the initial fault recognition results corresponding to the multiple frames of battery data, a battery fault recognition result of the vehicle battery is obtained, wherein the battery fault recognition result is at least used to indicate the fault category and the probability of fault occurrence of the vehicle battery, thereby achieving the purpose of performing preliminary fault recognition based on multiple frames of battery data collected at multiple sampling moments within the current time period, using a battery abnormality recognition model, and comprehensively determining the fault category and the probability of fault occurrence of the vehicle battery based on the multiple initial fault recognition results obtained, thereby achieving the technical effect of improving the efficiency and accuracy of vehicle battery fault recognition, and further solving the technical problem that the battery monitoring method in the related art usually relies on regular maintenance and manual inspection, which is not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in low warning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0023] Figure 1 is a flow chart of a vehicle battery abnormality identification method according to an embodiment of the present invention;

[0024] Figure 2 is a flow chart of an optional vehicle battery abnormality identification method according to an embodiment of the present invention;

[0025] Figure 3 is a schematic diagram of a vehicle battery abnormality identification device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] First, to facilitate understanding of the embodiments of the present invention, some terms or nouns involved in the present invention are explained below:

[0029] AutoGluon, an open source automatic machine learning toolkit, aims to simplify the training and deployment process of machine learning models. It provides a simple and powerful way to train multiple types of machine learning models, including classification, regression, clustering, etc., and automatically performs hyperparameter tuning to improve model performance.

[0030] According to an embodiment of the present invention, a method embodiment for identifying vehicle battery abnormalities is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] Figure 1 is a flow chart of a vehicle battery abnormality identification method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0032] Step S102, obtaining multiple frames of battery data of the vehicle battery collected corresponding to multiple sampling moments in the current time period;

[0033] Optionally, the current period is a period of a predetermined length between current sampling moments, and the multiple sampling moments include the current sampling moment and a predetermined number of sampling moments between the current sampling moments. The multiple frames of battery data may be obtained based on initial battery data of the vehicle battery collected at the multiple sampling moments, wherein the initial battery data may include, but is not limited to, voltage data, current data, and remaining state of charge, etc.

[0034] In an optional embodiment, multiple frames of battery data of a vehicle battery collected during the current time period are obtained, including: obtaining battery data corresponding to any sampling moment among multiple sampling moments in the following manner; obtaining initial battery data of the vehicle battery at any sampling moment, wherein the initial battery data includes at least: voltages corresponding to multiple battery cells included in the vehicle battery, currents corresponding to the multiple battery cells, and a remaining state of charge of the vehicle battery; based on the initial battery data, determining derived features of the vehicle battery in multiple feature dimensions; based on the derived features, obtaining battery data corresponding to any sampling moment; and obtaining multiple frames of battery data by obtaining battery data corresponding to any sampling moment.

[0035] Optionally, by obtaining the initial battery data of the vehicle battery at any sampling moment, including information such as the voltage, current, and remaining state of charge of each cell in the battery, the state and performance of the vehicle battery at the corresponding moment can be accurately understood. Based on the initial battery data, derived features of the vehicle battery in multiple feature dimensions can be obtained through calculation, analysis, and processing. These derived features can more comprehensively describe the state and performance of the battery. The derived features are used as the battery data at the corresponding sampling moment. In the above way, other key potential dimensional features can be derived, and the factors considered are more comprehensive, thereby improving the richness and accuracy of the input of the battery anomaly recognition model.

[0036] Optionally, when determining the derived features of the vehicle battery in multiple feature dimensions based on the initial battery data, the initial battery data can be preprocessed, including but not limited to performing basic data preprocessing operations such as deduplication, removal of outliers, missing values, etc. on the original battery data to obtain processed data; based on the processed data, the derived features of the vehicle battery in multiple feature dimensions are obtained.

[0037] Optionally, the multiple feature dimensions may include but are not limited to statistical feature dimensions, time difference dimensions and undervoltage risk level dimensions, wherein the statistical feature dimensions may include but are not limited to the mean, maximum, minimum, median, variance, kurtosis and skewness of voltage and temperature; the features of the time difference dimension may be obtained by first-order difference calculation based on the statistical features.

[0038] In an optional embodiment, based on the initial battery data, derived features of the vehicle battery in multiple feature dimensions are determined, including: based on the voltages corresponding to the multiple battery cells in the initial battery data, and the currents corresponding to the multiple battery cells, the statistical features of the vehicle battery at any sampling moment are determined, wherein the statistical features include voltage statistical features and temperature statistical features; the statistical features at any sampling moment and the statistical features at the previous sampling moment are differentially calculated to obtain the time differential features of the vehicle battery at any sampling moment, wherein the previous sampling moment is the sampling moment before any sampling moment; based on the remaining state of charge of the vehicle battery in the initial battery data, the undervoltage risk degree of the vehicle battery at any sampling moment is determined; based on the statistical features at any sampling moment, the time differential features at any sampling moment, and the undervoltage risk degree at any sampling moment, the derived features are obtained.

[0039] It can be understood that by statistically analyzing the voltages and currents corresponding to multiple cells in the initial battery data, voltage statistical features and temperature statistical features can be obtained; these statistical features can help understand the overall state and performance of the vehicle battery at that moment. By performing differential calculations on the statistical features at any sampling moment and the previous sampling moment, time differential features can be obtained; these time differential features reflect the changes in the vehicle battery at different moments, which helps to monitor the changing trend of battery performance. Based on the residual state of charge corresponding to multiple cells in the initial battery data, the undervoltage risk level of the vehicle battery at any sampling moment can be determined, which helps to evaluate the safety and stability of the battery. Based on statistical features, time differential features and undervoltage risk level, more comprehensive and comprehensive derived features can be obtained. These derived features can more accurately describe the state and performance of the battery, providing an important reference for subsequent abnormal identification of vehicle batteries. In the above manner, the information and change trends of the battery's voltage, current, temperature, residual state of charge, etc. can be comprehensively considered to obtain more comprehensive and accurate derived features, thereby achieving more comprehensive monitoring and management of the battery state and improving the performance and safety of the battery.

[0040] Optionally, since the entire battery pack of the vehicle is composed of multiple cells, each frame of voltage data and temperature data is in array form, that is, it includes the voltages corresponding to the multiple cells and the currents corresponding to the multiple cells. The mean, maximum, minimum, median, variance, kurtosis and skewness of the voltage and temperature can be calculated as statistical features. The statistical features obtained above are first-order differentiated according to the time dimension to obtain the time difference features used to measure the degree of change of the statistical features over time; further, the undervoltage risk degree of the vehicle battery is calculated based on the residual state of charge of the vehicle battery to measure the impact of the residual state of charge of the vehicle battery.

[0041] Optionally, determining the undervoltage risk level of the vehicle battery at any sampling time based on the remaining state of charge of the vehicle battery in the initial battery data includes: obtaining the undervoltage risk level of the vehicle battery at any sampling time by:

[0042]

[0043] Among them, R represents the undervoltage risk level at any sampling moment, U med Indicates the median of the voltages corresponding to multiple cells, U min Indicates the minimum value of the voltages corresponding to multiple cells, SOC max Indicates the preset maximum state of charge, SOC indicates the remaining state of charge of the vehicle battery. In order to avoid the situation where the denominator of the remaining state of charge of the vehicle battery is 0 when the denominator is full, you can set SOC max The value of is 101.

[0044] Step S104, based on the multiple frames of battery data, using a battery abnormality recognition model, obtain initial fault recognition results corresponding to the multiple frames of battery data, wherein the battery abnormality recognition model is based on multiple groups of historical battery data and historical fault recognition results corresponding to the multiple groups of historical battery data, and is obtained through machine learning;

[0045] Optionally, multiple frames of battery data are used as inputs of the battery anomaly recognition model, and the initial fault recognition results corresponding to the multiple frames of battery data can be quickly predicted, wherein the initial fault recognition results may include the fault category of the vehicle battery detected at the corresponding moment, and the probability of the fault occurring. The fault category may only be used to indicate whether the vehicle battery has a fault, or may indicate the type of fault of the vehicle battery, wherein the fault type may be, but is not limited to, undervoltage, pressure difference, high temperature, etc.

[0046] In an optional embodiment, when the initial fault identification result includes the fault category and fault occurrence probability of the corresponding frame battery data, a battery fault identification result of the vehicle battery is obtained based on the initial fault identification results corresponding to multiple frames of battery data, including: determining the number of battery data frames corresponding to multiple fault categories based on the fault categories corresponding to multiple frames of battery data; determining a target fault category with the largest number of battery data frames from multiple fault categories based on the number of battery data frames corresponding to multiple fault categories; determining the fault occurrence probability of the target fault category; and obtaining a battery fault identification result based on the target fault category and the fault occurrence probability of the target fault category.

[0047] Optionally, by performing statistical analysis on data of different fault categories in multiple frames of battery data, the number of times each fault category appears in the data, that is, the corresponding number of battery data frames, can be determined, which helps to understand the distribution of different fault categories in the data. Determining the fault category with the largest number of occurrences from multiple fault categories as the target fault category, that is, the fault category with the largest number of battery data frames, can help determine the fault type with the highest frequency in the data. Calculating the probability of occurrence of the target fault category based on the number of battery data frames where the target fault category appears helps to evaluate the importance of the target fault category in the overall data and the probability of its occurrence. Based on the target fault category and its probability of occurrence, the final result of battery fault identification can be obtained. The above operations can help determine the most frequently occurring fault category in the battery data and evaluate its probability of occurrence through statistical analysis and probability calculation, thereby achieving accurate identification of battery faults.

[0048] Optionally, taking the fault category including the presence of a fault and the absence of a fault as an example, a battery anomaly recognition model is used to classify and predict multiple frames of battery data, and each frame of data can obtain a prediction category (such as whether a fault exists) and a probability estimate (i.e., the probability of a fault occurring). Vehicles with a fault category of fault are screened, and the number of consecutive frames is calculated. If the number of consecutive frames is greater than the set number of fault frames (i.e., the predetermined number of frames), it means that the battery of the vehicle has shown a certain fault risk and needs to be reported as an early warning.

[0049] In an optional embodiment, determining the failure probability of a target failure category includes: determining a plurality of frames of target battery data corresponding to the target failure category from the plurality of frames of battery data, wherein the number of the plurality of frames of target battery data is less than or equal to the number of the plurality of frames of battery data; calculating the mean of the failure probabilities corresponding to the plurality of frames of target battery data to obtain an average failure probability; and using the average failure probability as the failure probability of the target failure category.

[0050] Optionally, by screening out data frames related to the target fault category from multiple frames of battery data, the mean of the fault occurrence probability corresponding to each frame of data in the multiple frames of target battery data is calculated to obtain an average fault occurrence probability, and the calculated average fault occurrence probability is used as the fault occurrence probability of the target fault category, thereby obtaining a more comprehensive and stable fault occurrence probability. The above method can more accurately evaluate the probability of occurrence of the target fault category and provide more reliable data support for fault identification and processing.

[0051] In an optional embodiment, a battery fault identification result is obtained based on a target fault category and a fault occurrence probability of the target fault category, including: determining whether the number of battery data frames corresponding to the target fault category is greater than a predetermined number of frames; if the number of battery data frames corresponding to the target fault category is greater than a predetermined number of frames, a battery fault identification result is obtained based on the target fault category and a fault occurrence probability of the target fault category.

[0052] By comparing the number of battery data frames corresponding to the target fault category with the predetermined number of frames, it is possible to determine whether the number of data for the target fault category meets the predetermined requirements; if the number of battery data frames corresponding to the target fault category exceeds the predetermined number of frames, the final battery fault identification result can be obtained based on the target fault category and the probability of the fault occurring. The above operation can determine whether the amount of data meets the requirements and perform fault identification analysis when the number of data frames is sufficient to obtain the final battery fault identification result. In the above manner, the problem of inaccurate fault identification caused by single-frame data errors can be avoided, thereby improving the accuracy and reliability of vehicle battery fault identification.

[0053] Step S106, obtaining a battery fault identification result of the vehicle battery based on the initial fault identification results corresponding to the multiple frames of battery data, wherein the battery fault identification result is at least used to indicate the fault type and fault occurrence probability of the vehicle battery.

[0054] Comprehensively determining the fault category of the vehicle battery based on the multiple initial fault identification results obtained can avoid the problem of inaccurate fault identification caused by single-frame data errors and improve the accuracy and reliability of vehicle battery fault identification.

[0055] In an optional embodiment, before obtaining initial fault identification results corresponding to multiple frames of battery data by using a battery abnormality identification model based on multiple frames of battery data, the method also includes: dividing multiple groups of historical battery data, and historical fault identification results corresponding to the multiple groups of historical battery data into a training data set and a test data set, wherein the multiple groups of historical battery data include battery data of multiple vehicles during a battery failure period, and battery data of multiple vehicles during a normal operating period, and the normal operating period is a period of a predetermined length of time after the battery failure of the corresponding vehicle is repaired; training multiple initial models based on the training data set to obtain multiple training models, wherein the multiple initial models are constructed based on different algorithms; testing the multiple training models based on the test data set to obtain test results corresponding to the multiple training models; and determining a battery abnormality identification model from the multiple training models based on the test results corresponding to the multiple training models.

[0056] Optionally, you can use the AutoGluon automated machine learning model based on the training data set and the test data set to train and test multiple initial models at the same time, and select the best model as the battery anomaly recognition model based on the test results. By training and screening multiple models based on the same data, you can not only improve the model training efficiency, but also screen out the model with the best performance.

[0057] Optionally, multiple initial models are constructed based on different algorithm models, wherein the algorithm may include but is not limited to a logistic regression model, a support vector machine (SVM) model, a decision tree model, a random forest model, a K-nearest neighbors (KNN), a naive Bayes algorithm, a gradient boosting tree (GBM) algorithm, a neural network algorithm, and an ensemble learning method (Ensemble Learning), etc.

[0058] Through the above steps S102 to S106, it is possible to achieve the purpose of performing preliminary fault identification based on multiple frames of battery data collected at multiple sampling moments in the current time period, using a battery abnormality identification model, and comprehensively determining the fault type of the vehicle battery and the probability of the fault occurring based on the multiple initial fault identification results obtained, thereby achieving the technical effect of improving the efficiency and accuracy of vehicle battery fault identification, and further solving the technical problem that the battery monitoring method in the related art usually relies on regular maintenance and manual inspection, which is not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in low warning accuracy.

[0059] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation mode: Figure 2 is a flow chart of an optional vehicle battery abnormality identification method according to an embodiment of the present invention, such as Figure 2 As shown, the method includes:

[0060] S1: Obtain battery data and label it, including:

[0061] The big data cloud platform refers to a cloud platform built using big data technology to process and store vehicle power battery information and after-sales processing. The list of batteries of faulty vehicles that have been repaired is obtained from the work order information table of the cloud platform, as well as the corresponding order creation time, repair completion time, fault category and other related information. Fault categories include undervoltage, pressure difference, high temperature, etc.

[0062] The battery data of each faulty vehicle in the list from the time the order was created to the time the repair was completed is obtained from the vehicle battery information table as a negative sample, and the category label of each frame of data is annotated with the fault category of the vehicle; the battery data of each faulty vehicle from the time the repair was completed to 8 hours later is taken as a positive sample, and the category label of each frame of data is annotated as normal.

[0063] S2: Data processing and model feature derivation, including:

[0064] The training sample data obtained in step S1 is subjected to basic data preprocessing operations such as deduplication, removal of outliers, and missing values. Based on the original feature data (i.e., original vehicle data), three dimensions of potential features are derived: statistical features: Since the whole pack of batteries is composed of multiple cells, each frame of voltage data and temperature data is in array form, that is, it includes the voltages corresponding to multiple cells and the currents corresponding to multiple cells. The mean, maximum, minimum, median, variance, kurtosis, and skewness of the voltage and temperature can be calculated; time difference features: the above statistical derived features are first-order differentiated to measure the degree of change before and after time; undervoltage risk features: comprehensive SOC change impact, according to the calculation formula: (median voltage - minimum voltage) / (101-SOC), calculate the voltage risk level corresponding to each frame of battery data, and measure the degree of undervoltage risk.

[0065] S3: AutoGluon automates machine learning model training, including:

[0066] Import the data obtained in step S2 into AutoGluon as training samples. AutoGluon will select more than a dozen better algorithm models for training at the same time, automatically divide the training set and the validation set, automatically adjust the hyperparameters to select the optimal algorithm model, and automatically save the optimal model obtained through training as the battery anomaly recognition model locally.

[0067] S4: Real-time detection and abnormal warning of vehicle batteries, including:

[0068] Run the program at a certain frequency, first obtain the battery data of the vehicle running online some time before the current moment, perform the same data preprocessing and feature derivation as step S2, and obtain multiple frames of battery data. Use the saved battery anomaly recognition model to classify and predict the multiple frames of battery data, so that each frame of data for each vehicle can get a prediction category (such as whether there is a fault) and probability estimate. Filter the vehicles predicted to be classified as fault categories, and calculate the number of consecutive frames. If the number of consecutive frames is greater than the set number of fault frames (i.e., the predetermined number of frames), it means that the battery of the vehicle has shown a certain fault risk and needs to be reported as an early warning.

[0069] According to the method of step S1 to step S4, vehicle maintenance and battery maintenance can be achieved during the maintenance of the whole vehicle, and battery failure signs can be predicted in advance to reduce the number of market failure reports.

[0070] In this embodiment, a vehicle battery abnormality identification device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and the descriptions that have been made will not be repeated. As used below, the terms "module" and "device" can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0071] According to an embodiment of the present invention, there is also provided an embodiment of a device for implementing the above-mentioned vehicle battery abnormality identification method. Figure 3 is a schematic diagram of the structure of a vehicle battery abnormality identification device according to an embodiment of the present invention. Figure 3 As shown, the above-mentioned vehicle battery abnormality identification device includes: a data acquisition module 300, an initial fault identification module 302, and a battery fault identification module 304, wherein:

[0072] The data acquisition module 300 is used to acquire multiple frames of battery data of the vehicle battery collected corresponding to multiple sampling moments in the current period;

[0073] The initial fault identification module 302 is connected to the data acquisition module 300 and is used to obtain initial fault identification results corresponding to the multiple frames of battery data by using a battery abnormality identification model based on multiple frames of battery data, wherein the battery abnormality identification model is based on multiple groups of historical battery data and historical fault identification results corresponding to the multiple groups of historical battery data, and is obtained through machine learning;

[0074] The battery fault identification module 304 is connected to the initial fault identification module 302, and is used to obtain a battery fault identification result of the vehicle battery based on the initial fault identification results corresponding to multiple frames of battery data, wherein the battery fault identification result is at least used to indicate the fault type and fault occurrence probability of the vehicle battery.

[0075] In an embodiment of the present invention, a data acquisition module 300 is provided to acquire multiple frames of battery data of a vehicle battery collected at multiple sampling moments within a current period; an initial fault identification module 302 is connected to the data acquisition module 300 and is used to obtain initial fault identification results corresponding to the multiple frames of battery data by adopting a battery abnormality identification model based on the multiple frames of battery data, wherein the battery abnormality identification model is based on multiple groups of historical battery data and historical fault identification results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; a battery fault identification module 304 is connected to the initial fault identification module 302 and is used to obtain initial fault identification results corresponding to the multiple frames of battery data based on the initial fault identification results corresponding to the multiple frames of battery data. The battery fault identification result of the vehicle battery is obtained, wherein the battery fault identification result is at least used to indicate the fault type and fault probability of the vehicle battery, and achieves the purpose of preliminary fault identification based on multiple frames of battery data collected at multiple sampling moments in the current time period, using a battery abnormality identification model, and comprehensively determining the fault type and fault probability of the vehicle battery based on the multiple initial fault identification results obtained, thereby achieving the technical effect of improving the efficiency and accuracy of vehicle battery fault identification, and further solving the technical problem that the battery monitoring method in the related art usually relies on regular maintenance and manual inspection, which is not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in low warning accuracy.

[0076] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0077] It should be noted that the data acquisition module 300, the initial fault identification module 302, and the battery fault identification module 304 correspond to steps S102 to S106 in the embodiment, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments. It should be noted that the modules as part of the device can be run in a computer terminal.

[0078] It should be noted that the optional or preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0079] The above-mentioned vehicle battery abnormality identification device may also include a processor and a memory. The above-mentioned data acquisition module 300, initial fault identification module 302, battery fault identification module 304, etc. are all stored in the memory as program modules, and the processor executes the above-mentioned program modules stored in the memory to realize corresponding functions.

[0080] The processor includes a kernel, which retrieves the corresponding program module from the memory. The kernel may be one or more. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory includes at least one memory chip.

[0081] According to an embodiment of the present application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute any of the above-mentioned vehicle battery abnormality identification methods.

[0082] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group, and the non-volatile storage medium includes a stored program.

[0083] Optionally, when the program is running, the device where the non-volatile storage medium is located is controlled to perform the following functions: obtaining multiple frames of battery data of the vehicle battery collected corresponding to multiple sampling moments within the current time period; based on the multiple frames of battery data, using a battery abnormality recognition model to obtain initial fault recognition results corresponding to the multiple frames of battery data, wherein the battery abnormality recognition model is based on multiple groups of historical battery data, and historical fault recognition results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; based on the initial fault recognition results corresponding to the multiple frames of battery data, a battery fault recognition result of the vehicle battery is obtained, wherein the battery fault recognition result is at least used to indicate the fault type and fault probability of the vehicle battery.

[0084] According to an embodiment of the present application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein when the program is run, any one of the above-mentioned vehicle battery abnormality identification methods is executed.

[0085] According to an embodiment of the present application, an embodiment of a computer program product is also provided, which, when executed on a data processing device, is suitable for executing a program that initializes any of the above-mentioned vehicle battery abnormality identification method steps.

[0086] Optionally, the above-mentioned computer program product, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: obtaining multiple frames of battery data of the vehicle battery collected corresponding to multiple sampling moments within the current time period; based on the multiple frames of battery data, using a battery abnormality recognition model to obtain initial fault recognition results corresponding to the multiple frames of battery data, wherein the battery abnormality recognition model is based on multiple groups of historical battery data and historical fault recognition results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; based on the initial fault recognition results corresponding to the multiple frames of battery data, a battery fault recognition result of the vehicle battery is obtained, wherein the battery fault recognition result is at least used to indicate the fault type and fault occurrence probability of the vehicle battery.

[0087] An embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining multiple frames of battery data of a vehicle battery collected corresponding to multiple sampling moments in a current time period; based on the multiple frames of battery data, using a battery abnormality recognition model, obtaining initial fault recognition results corresponding to the multiple frames of battery data, wherein the battery abnormality recognition model is based on multiple groups of historical battery data and historical fault recognition results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; based on the initial fault recognition results corresponding to the multiple frames of battery data, obtaining a battery fault recognition result of the vehicle battery, wherein the battery fault recognition result is at least used to indicate the fault category and fault occurrence probability of the vehicle battery.

[0088] The above sequence of the embodiments of the present invention is for description only and does not represent the superiority or inferiority of the embodiments.

[0089] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0090] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the above modules can be a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, modules or indirect coupling or communication connection of modules, which can be electrical or other forms.

[0091] The modules described above as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0092] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of software functional modules.

[0093] If the above-mentioned integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present invention. The aforementioned non-volatile storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0094] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for identifying abnormality of a vehicle battery, characterized in that: include: Obtain multiple frames of battery data of the vehicle battery collected corresponding to multiple sampling moments in the current time period; Based on the multiple frames of battery data, a battery abnormality recognition model is used to obtain initial fault recognition results corresponding to the multiple frames of battery data, wherein the battery abnormality recognition model is based on multiple groups of historical battery data and historical fault recognition results corresponding to the multiple groups of historical battery data, and is obtained through machine learning; Based on the initial fault identification results respectively corresponding to the multiple frames of battery data, a battery fault identification result of the vehicle battery is obtained, wherein the battery fault identification result is at least used to indicate the fault type and fault occurrence probability of the vehicle battery.

2. The method according to claim 1, characterized in that The step of obtaining multiple frames of battery data of the vehicle battery collected during the current period includes: Obtaining battery data corresponding to any sampling time among the multiple sampling times in the following manner; Acquire initial battery data of the vehicle battery at any sampling time, wherein the initial battery data at least includes: voltages corresponding to a plurality of battery cells included in the vehicle battery, currents corresponding to the plurality of battery cells, and a remaining state of charge of the vehicle battery; Determining derived features of the vehicle battery in a plurality of feature dimensions based on the initial battery data; Based on the derived features, obtaining battery data corresponding to any sampling moment; The multiple frames of battery data are obtained by obtaining the battery data corresponding to any sampling moment.

3. The method according to claim 2, characterized in that The step of determining, based on the initial battery data, derived features of the vehicle battery in multiple feature dimensions includes: Determine statistical characteristics of the vehicle battery at any sampling time based on the voltages corresponding to the multiple battery cells and the currents corresponding to the multiple battery cells in the initial battery data, wherein the statistical characteristics include voltage statistical characteristics and temperature statistical characteristics; Performing differential calculation on the statistical features at any sampling moment and the statistical features at a previous sampling moment to obtain the time differential features of the vehicle battery at any sampling moment, wherein the previous sampling moment is a sampling moment before any sampling moment; Determining the undervoltage risk level of the vehicle battery at any sampling time based on the remaining state of charge of the vehicle battery in the initial battery data; The derived feature is obtained based on the statistical feature of any sampling moment, the time difference feature of any sampling moment, and the undervoltage risk degree of any sampling moment.

4. The method according to claim 1, characterized in that: In the case where the initial fault identification result includes the fault category and the fault occurrence probability of the corresponding frame battery data, the battery fault identification result of the vehicle battery is obtained based on the initial fault identification results respectively corresponding to the multiple frames of battery data, including: Based on the fault categories respectively corresponding to the multiple frames of battery data, determine the number of battery data frames corresponding to the multiple fault categories respectively; Based on the numbers of battery data frames respectively corresponding to the multiple fault categories, determining a target fault category with the largest number of battery data frames from the multiple fault categories; Determining a failure occurrence probability of the target failure category; The battery fault identification result is obtained based on the target fault category and the fault occurrence probability of the target fault category.

5. The method according to claim 4, characterized in that The determining the failure occurrence probability of the target failure category includes: Determining a plurality of frames of target battery data corresponding to the target fault category from the plurality of frames of battery data, wherein the number of the plurality of frames of target battery data is less than or equal to the number of the plurality of frames of battery data; Calculating the mean of the fault occurrence probabilities corresponding to the multiple frames of target battery data to obtain an average fault occurrence probability; The average failure occurrence probability is used as the failure occurrence probability of the target failure category.

6. The method according to claim 4, characterized in that The obtaining of the battery fault identification result based on the target fault category and the fault occurrence probability of the target fault category includes: Determining whether the number of battery data frames corresponding to the target fault category is greater than a predetermined number of frames; When the number of battery data frames corresponding to the target fault category is greater than the predetermined number of frames, the battery fault identification result is obtained based on the target fault category and the fault occurrence probability of the target fault category.

7. The method according to any one of claims 1 to 6, characterized in that: Before obtaining initial fault identification results respectively corresponding to the multiple frames of battery data by using a battery abnormality identification model based on the multiple frames of battery data, the method further includes: Dividing the multiple sets of historical battery data and the historical fault identification results respectively corresponding to the multiple sets of historical battery data into a training data set and a test data set, wherein the multiple sets of historical battery data include battery data of multiple vehicles during a battery fault period and battery data of the multiple vehicles during a normal operation period, wherein the normal operation period is a period of a predetermined length of time corresponding to the vehicle after the battery fault is repaired; Training multiple initial models based on the training data set to obtain multiple training models, wherein the multiple initial models are constructed based on different algorithms; Testing the multiple training models based on the test data set to obtain test results corresponding to the multiple training models respectively; Based on the test results respectively corresponding to the multiple training models, the battery abnormality recognition model is determined from the multiple training models.

8. A vehicle battery abnormality identification device, characterized in that: include: A data acquisition module, used to acquire multiple frames of battery data of the vehicle battery collected corresponding to multiple sampling moments in the current period; An initial fault identification module, used to obtain initial fault identification results corresponding to the multiple frames of battery data respectively using a battery abnormality identification model based on the multiple frames of battery data, wherein the battery abnormality identification model is based on multiple groups of historical battery data and historical fault identification results corresponding to the multiple groups of historical battery data respectively, and is obtained through machine learning; A battery fault identification module is used to obtain a battery fault identification result of the vehicle battery based on the initial fault identification results corresponding to the multiple frames of battery data, wherein the battery fault identification result is at least used to indicate the fault type and fault occurrence probability of the vehicle battery.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the vehicle battery abnormality identification method described in any one of claims 1 to 7.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle battery abnormality identification method described in any one of claims 1 to 7.

Citation Information

Cited By

  • Fault identification method and device for battery cluster and nonvolatile storage medium

    CN121208691A