A vehicle power battery fault diagnosis method based on hierarchical identification

Abstract

The present invention belongs to the field of power battery safety technology, and specifically relates to a vehicle power battery fault diagnosis method based on hierarchical identification, comprising: obtaining the vehicle's actual operating data; dividing the actual operating data into a number of data segments, and extracting one-dimensional or multi-dimensional feature quantities from the data segments; screening out abnormal data segments from all data segments based on the feature quantities of the data segments; for the abnormal data segments, calculating the longitudinal abnormal average value of the power battery cell voltage to identify the battery cells suspected of failure in the power battery, and analyzing the power battery failure type based on the performance of the suspected faulty battery cells. The present invention screens out abnormal segments by analyzing the vehicle's actual operating data, and determines whether the battery has a fault. It can also identify the battery cells suspected of failure in the power battery, so as to subsequently analyze the battery failure type and cause, and maintain the safety of the vehicle and personnel.

Description

Technical Field

[0001] The present invention belongs to the technical field of power battery safety, and in particular relates to a vehicle power battery fault diagnosis method based on hierarchical identification. Background Art

[0002] In the process of comprehensive electrification of transportation, power batteries are one of the core components of electric vehicles (including passenger cars, buses, trucks, etc.), and their performance directly affects the economy and safety of electric vehicles. However, due to the aging of the power batteries themselves, the failure of sensor acquisition or other external reasons, when the electric vehicle is stopped from charging or driving, various fault problems such as connection failure, sampling failure, insulation failure, self-discharge failure and sudden internal short circuit failure may occur. For example, in actual application, due to the loss of control of the internal mechanism of the power battery or external collision and compression during application, the heat accumulation rate of the battery is faster than the diffusion rate, which in turn causes a fire accident that burns rapidly and is difficult to extinguish, endangering personal safety.

[0003] In order to better identify power battery faults during operation and determine the fault type, it is necessary to conduct targeted analysis based on the vehicle's actual operating data. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of the present invention is to provide a fault diagnosis method that can solve the problem of power battery fault identification and diagnosis based on the actual operating data of electric vehicles powered by power batteries, so as to improve the safety and economic benefits of electric vehicles.

[0005] To achieve the above-mentioned objectives and other related objectives, the present invention provides a vehicle power battery fault diagnosis method based on hierarchical identification, including: obtaining actual operating data of the vehicle; dividing the actual operating data into a plurality of data segments, and extracting one-dimensional or multi-dimensional feature quantities from the data segments; screening out abnormal data segments from all data segments based on the feature quantities of the data segments; for abnormal data segments, calculating the longitudinal abnormal average value of the power battery cell voltage to identify the battery cells suspected of failure in the power battery, and analyzing the fault type of the power battery based on the performance of the battery cells suspected of failure.

[0006] According to a specific embodiment of the present invention, after obtaining the actual operation data of the vehicle, the method further includes: performing data cleaning on the actual operation data to remove failure values ​​in the original data.

[0007] According to a specific embodiment of the present invention, the steps of dividing the actual operating data into a plurality of data segments and extracting one-dimensional or multi-dimensional feature quantities from the data segments include: for any data segment, extracting the corresponding multi-dimensional feature quantities based on sampling parameters related to the power battery therein; wherein the sampling parameters of the power battery include: the total voltage of the power battery, the total current of the power battery, the state of charge of the power battery, the temperature extremes of the power battery, the insulation resistance of the power battery, and the single cell voltage of the power battery.

[0008] According to a specific embodiment of the present invention, for any data segment, the step of extracting corresponding multi-dimensional feature quantities based on the sampling parameters of the power battery therein includes: for the total voltage, total current, state of charge, or insulation resistance of the power battery, obtaining the sampling values ​​at all times, and calculating the average value, standard deviation, quartile, quartile, minimum value and maximum value of the sampling values ​​as one of the one-dimensional feature quantities; for the temperature extremes of the power battery, obtaining the sampling values ​​at all times, and calculating the minimum value, maximum value, and average value of the lowest sampling temperature, as well as the minimum value, maximum value, and average value of the highest sampling temperature, as one of the one-dimensional feature quantities; for the single cell voltage of the power battery, obtaining the sampling values ​​at all times, and calculating the maximum value and standard deviation of the highest single cell voltage, as well as the average value and standard deviation of the single cell voltage extreme difference, and the average value and maximum value of the voltage entropy, as one of the one-dimensional feature quantities.

[0009] According to a specific embodiment of the present invention, the step of screening out abnormal data segments from all data segments based on the feature quantities of the data segments includes: for any two data segments, calculating the Euclidean distance between different data segments with respect to the same feature quantities, and adding one to the number of abnormalities of the two data segments when the Euclidean distance is abnormal; identifying abnormal data segments based on the cumulative number of abnormalities of each data segment, and integrating all abnormal data segments to diagnose whether the power battery is faulty.

[0010] According to a specific embodiment of the present invention, the step of judging whether the Euclidean distance is abnormal includes: for any type of feature quantity, calculating the Euclidean distance of any two different data fragments with respect to the feature quantity based on all data fragments, until all Euclidean distances with respect to the feature quantity are obtained; obtaining the corresponding quartile, quartile, and standard deviation based on all Euclidean distances with respect to the feature quantity, and calculating the first threshold and the second threshold based on the quartile, the quartile, and the standard deviation, which are used as reference standards for whether the Euclidean distance of the feature quantity is abnormal; wherein, when the Euclidean distance is lower than the corresponding first threshold or higher than the corresponding second threshold, the Euclidean distance is considered to be abnormal.

[0011] According to a specific embodiment of the present invention, the calculation formulas of the first threshold and the second threshold are as follows:

[0012] x1=Q1-1.5*Std,

[0013] x2=Q2+1.5*Std,

[0014] Wherein, x1 represents the first threshold, x2 represents the second threshold, Q1 represents the quartile, Q2 represents the quartile, and Std represents the standard deviation.

[0015] According to a specific embodiment of the present invention, for abnormal data segments, the longitudinal abnormal average value of the power battery cell voltage is calculated to identify the battery cell suspected of failure in the power battery, and the failure type of the power battery is analyzed according to the performance of the battery cell suspected of failure. The steps include: for each abnormal data segment, dividing it into a number of charging segments and discharging segments; calculating the longitudinal abnormal average value of each cell of the power battery within each of the charging segment / discharging segment time, and adding one to the abnormal number of the corresponding cell when the longitudinal abnormal average value is abnormal; identifying the battery cell suspected of failure according to the cumulative number of abnormalities of each cell of the power battery.

[0016] According to a specific embodiment of the present invention, the calculation formula of the longitudinal abnormal average value is as follows:

[0017]

[0018] Among them, LOA j Indicates the average longitudinal abnormality of the jth cell of the power battery, V ij Represents the cell voltage of the jth cell at the i-th moment, V imid It represents the median of all cell voltages at the i-th moment, and n represents the number of cells in the power battery.

[0019] According to a specific embodiment of the present invention, the step of determining whether the longitudinal abnormality average value is abnormal includes: for each charging segment / discharging segment, using the three sigma law to determine a corresponding standard interval; comparing the longitudinal abnormality average value calculated using the charging segment / discharging segment with the standard interval: if the longitudinal abnormality average value exceeds the standard interval, then the longitudinal abnormality average value is considered abnormal.

[0020] This invention provides a vehicle power battery fault diagnosis method based on hierarchical identification. It can quickly identify faults in faulty vehicles with only a few months of data. By analyzing actual operating data, it screens out abnormal segments and uses them to determine whether the vehicle's power battery is faulty. Furthermore, it can further identify suspected faulty cells within the power battery, facilitating subsequent analysis of the battery fault type and cause. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of a specific embodiment of a vehicle power battery fault diagnosis method based on hierarchical identification provided by the present invention;

[0022] Figure 2 This is a structural diagram of a specific embodiment of a vehicle power battery fault diagnosis system based on hierarchical identification provided by the present invention;

[0023] Figure 3 This is a structural block diagram of a specific embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0024] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0025] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0026] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, publicly known structures and devices are shown in block diagram form rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0027] Example 1

[0028] See Figure 1As shown, a vehicle power battery fault diagnosis method based on hierarchical identification includes:

[0029] Step S100: Acquire actual operating data of the vehicle.

[0030] It should be noted that the actual operating data mentioned in this embodiment refers to data collected / detected by relevant vehicle sensors or other devices when the electric vehicle is started or running using the power battery. For example, when the vehicle is driving on the road, the power battery's output current, output voltage, temperature, and other sampled parameters will be collected accordingly. As the heart of the electric vehicle, the power battery requires special attention to ensure its normal operation and the presence of any abnormalities or failures to ensure safe driving.

[0031] The fault diagnosis method provided in this embodiment can diagnose the fault status of the power battery online, and of course can also be used for offline diagnosis. For example, after the vehicle is started, the relevant vehicle sensors will collect the operating parameters of the power battery in real time and feed them back to the vehicle terminal. The vehicle terminal can then perform fault diagnosis based on the fed-back sampling parameters, thereby monitoring the operating status of the power battery in real time to ensure the safety of the vehicle and personnel; or, after the vehicle has been running for a period of time, the corresponding data can be integrated to perform fault diagnosis regularly to feedback the health status of the power battery.

[0032] It should also be added that the actual operating data mentioned in this embodiment is not just the operating data of the power battery at a certain moment, but the operating data of the power battery over a period of time. By specifically analyzing the sampling parameters of the power battery over a period of time, it is possible to effectively identify whether the power battery has a fault, the type of fault, etc.

[0033] Based on the above, after obtaining the actual operating data of the vehicle, the data can be pre-processed, such as denoising, filtering, etc., without excessive limitation. Those skilled in the art may modify and embellish the embodiments of the present invention without departing from the spirit of the present invention and still fall within the scope of the invention application of the present invention. In this embodiment, in order to simplify the data processing steps, it is preferred to perform only simple data cleaning, that is, to remove failure values ​​in the original data, thereby ensuring that the subsequent fault diagnosis of the power battery is not affected by some abnormal parameters.

[0034] Step S200: dividing the actual operation data into a plurality of data segments, and extracting one-dimensional or multi-dimensional feature quantities from the data segments.

[0035] After the data cleaning is completed, the actual operation data can be divided into several data segments. For example, it can be split according to the vehicle's mileage, time, or preset segment length. It can be understood here that since the actual operation data is the operating data of the vehicle's power battery over a period of time, it can be continuous or discontinuous and can be arranged in chronological order. If it is split according to the vehicle's mileage, then the data before and after the vehicle travels the preset mileage can be split into a data segment, thereby obtaining several data segments; if it is split according to time, then the data of the vehicle within half a day, a day or a few days can be split into a data segment, thereby obtaining several data segments; if it is split according to segment length, then all data can be split into several data segments of equal length, and it can also be split into data segments of unequal length, etc. There are no excessive restrictions on how to split the data and the number of data segments.

[0036] Secondly, corresponding feature quantities need to be extracted for each data segment. It should be noted that in this embodiment, to better illustrate the technical solution of the present invention, the sampling parameters of the power battery are taken as examples of total voltage, total current, state of charge (SOC), temperature extremes, insulation resistance, and cell voltage. That is, vehicle sensors or other devices need to obtain relevant parameters of the power battery during operation, such as total voltage, total current, SOC, temperature extremes, insulation resistance, and cell voltage, to monitor the operating status of the power battery and feed them back to the vehicle terminal for storage. Therefore, each data segment of actual operating data contains sampling parameters related to the total voltage, total current, SOC, temperature extremes, insulation resistance, and cell voltage of the power battery, and each data segment needs to extract corresponding feature quantities for different sampling parameters. That is, for any data segment, six-dimensional feature quantities need to be extracted. However, it should be understood that in actual applications, the sampling parameters of the power battery are not limited to the above. Sampling parameters can be added or reduced according to actual conditions and needs. Corresponding feature quantity extraction also needs to be adaptively adjusted to obtain corresponding one-dimensional or multi-dimensional feature quantities.

[0037] In a specific embodiment, for the total voltage, total current, SOC and insulation resistance of the power battery, the average value, standard deviation, quartile, quartile, minimum value and maximum value of the corresponding sampling values ​​in the segment can be extracted, totaling 6 feature quantities, that is, a data segment will contain multiple sampling moments. By obtaining the sampling values ​​of the total voltage, total current, SOC and insulation resistance at all sampling moments, and calculating the average value, standard deviation, quartile, quartile, minimum value and maximum value of all the sampling values ​​of the total voltage, the average value, standard deviation, quartile, quartile, minimum value and maximum value of all the sampling values ​​of the total current, the average value, standard deviation, quartile, quartile, minimum value and maximum value of all the sampling values ​​of the SOC, and the average value, standard deviation, quartile, quartile, minimum value and maximum value of all the sampling values ​​of the insulation resistance, the total voltage, total current, SOC and insulation resistance distribution of the power battery in the segment are constructed.

[0038] For the extreme temperature values ​​of the power battery, the same approach is used to obtain sample values ​​at all sampling moments. The minimum, maximum, and average values ​​of the lowest sampled temperature, as well as the minimum, maximum, and average values ​​of the highest sampled temperature, are calculated, totaling six characteristic quantities. This constitutes the temperature extreme value distribution of the power battery for that segment. It should be noted that the temperature sample can be a corresponding distribution change—that is, the temperature change of the power battery over a certain period of time—as a sample value. For example, the temperature sample at a sampling moment can be the temperature distribution change between the previous moment and the current moment. A distribution change inevitably has a maximum and a minimum value, corresponding to the highest and lowest sampled temperatures, respectively. Therefore, the highest and lowest sampled temperatures can be collected from all sampled values, and the minimum, maximum, and average values ​​of all lowest sampled temperatures, as well as the minimum, maximum, and average values ​​of all highest sampled temperatures, can be calculated.

[0039] For the power battery cell voltage, we also obtain the sampled values ​​at all sampling moments and then calculate the voltage distribution changes of the power battery cells over a certain period of time. We can then calculate the maximum value and standard deviation of the highest cell voltage, as well as the average value and standard deviation of the cell voltage range. At the same time, we also calculate the voltage entropy at each sampling moment within the segment, and the average and maximum value of the voltage entropy at all sampling moments. A total of six characteristic quantities are calculated, thus forming the power battery cell voltage distribution for that segment.

[0040] Based on the above, corresponding 6-dimensional feature quantities can be extracted from each data segment to represent the operating status of the power battery. At the same time, since each dimension contains 6 feature quantities and there are 6 dimensions in total, a 6*6 feature matrix can be obtained in the end, which fully reflects the status performance of the power battery in this segment.

[0041] It should be noted that there are no excessive restrictions on the feature quantities of each dimension. Feature quantities can be added or reduced according to actual conditions and needs, and are not limited to the above 6 types. There are no excessive restrictions on this. Those skilled in the art can make modifications and improvements to the embodiments of the present invention without departing from the spirit of the present invention. They still fall within the scope of the invention patent application of the present invention.

[0042] Step S300 : Screening out abnormal data segments from all data segments based on the feature quantities of the data segments.

[0043] Furthermore, by measuring the similarity between two data fragments, abnormal data fragments can be screened out. Specifically, by calculating the Euclidean distance of each feature quantity between the two data fragments, the similarity of the sampling value distribution between each independent fragment is determined. The larger the distance, the lower the similarity. It should be noted here that the calculation of the Euclidean distance between each feature quantity specifically refers to the calculation between similar feature quantities, that is, the Euclidean distance between the feature quantity of the total voltage of the power battery of one data fragment and the feature quantity of the total voltage of the power battery of another data fragment, and all feature quantities need to be calculated one by one. Therefore, when there are multiple data fragments, any two data fragments can be matched by permutations and combinations, and the Euclidean distance of each feature quantity between the two data fragments can be calculated.

[0044] Furthermore, it is necessary to determine whether the calculated Euclidean distance meets the standard. First, based on any type of characteristic quantity, such as the total voltage, total current, SOC, temperature extremes, insulation resistance, or cell voltage of the power battery, the Euclidean distance of any two data segments is calculated based on all permutations and combinations of all data segments, resulting in several calculated values. Second, based on all calculated values, the corresponding quantiles, quartiles, and standard deviations are calculated, denoted as Q1, Q2, and Std, respectively. Based on Q1, Q2, and Std, a first threshold x1 and a second threshold x2 are calculated: x1 = Q1 - 1.5 * Std, and x2 = Q2 + 1.5 * Std, respectively. These serve as the criteria for determining whether the Euclidean distance for this type of characteristic quantity is abnormal. Finally, when the Euclidean distance between any two data segments for this type of characteristic quantity is calculated, if the calculated Euclidean distance is lower than the first threshold x1 = Q1 - 1.5 * Std or higher than the second threshold x2 = Q2 + 1.5 * Std, it is considered abnormal, and the number of abnormalities for each data segment is recorded once. Similarly, the Euclidean distance of the feature value of this type can be calculated for all data segments after permutation and combination for identification.

[0045] Similarly, other types of feature quantities also use the above method to calculate the Euclidean distance and identify whether they are abnormal. Finally, the cumulative number of abnormalities in all data segments is counted to filter out abnormal data segments. In practical applications, for example, a data segment can be considered abnormal when the cumulative number of abnormalities exceeds one-third of the total number of data segments. There is no need to limit the standard for determining abnormalities, and it can be adjusted according to actual conditions and needs.

[0046] Based on the above, a preliminary diagnosis of whether a power battery failure has occurred can be made based on the abnormal data segments. It should be noted that the cumulative number of abnormalities in the data segments can be further categorized, such as which type of feature value has the most abnormal Euclidean distances in a particular data segment, or the distribution of abnormal Euclidean distances for all types of feature values. This allows for statistical analysis of the number of abnormal Euclidean distances for different types of feature values, enabling a more accurate analysis of the power battery failure type and cause.

[0047] In one specific embodiment, for example, whether a power battery has experienced insulation failure can be determined based on all abnormal data segments. If the ratio of the number of abnormal Euclidean distances associated with insulation resistance to the number of abnormal Euclidean distances associated with total voltage does not exceed 0.6, the cause of the abnormal Euclidean distance for insulation resistance, a characteristic variable of this type, is considered to be a sampling anomaly. Otherwise, the fluctuation in insulation resistance of the power battery is considered to be related to its total voltage, indicating insulation failure. Temperature anomalies in a power battery can also be determined based on abnormal data segments. If the ratio of the number of abnormal Euclidean distances associated with temperature extremes to the number of abnormal Euclidean distances associated with total current does not exceed 0.6, the cause of the abnormal Euclidean distance for temperature extremes, a characteristic variable of this type, is considered to be a sampling anomaly, meaning it is unrelated to the power battery's operating conditions. Otherwise, the cause is related to the operating conditions and not a sampling anomaly, but rather the power battery's operating state does not meet normal standards.

[0048] It can be seen that abnormal data fragments can be used to preliminarily diagnose whether the power battery is operating normally, so that early warning can be given to avoid endangering the safety of vehicles and personnel.

[0049] In step S400 , for the abnormal data segments, the vertical abnormal average value of the power battery cell voltage is calculated to identify the battery cell suspected of failure in the power battery, and the failure type of the power battery is analyzed based on the performance of the battery cell suspected of failure.

[0050] Based on the abnormal data segments obtained above, fault identification can be further performed to obtain the specific faulty cell, that is, the battery cell (cell) in the power battery that may have a fault. Specifically, first, the abnormal data segments need to be divided into several charging segments and discharging segments. Secondly, the specific faulty cell is identified by calculating the Longitudinal Outlier Average (LOA) of the cell voltage of the power battery. Specifically, for all charging segments and discharging segments, the Longitudinal Outlier Average (LOA) of each battery cell of the power battery within each segment is calculated. The calculation formula is as follows:

[0051]

[0052] Among them, LOA j Indicates the average longitudinal abnormality of the jth cell of the power battery, V ij Represents the cell voltage of the jth cell at the i-th moment, V imid It represents the median of all cell voltages at the i-th moment, and n represents the number of cells in the power battery.

[0053] Furthermore, according to the three-sigma principle, the reliable interval of the LOA value of each segment, that is, the standard interval, is obtained to measure whether the LOA value is normal. It should be noted here that the LOA value of the battery cell calculated in any charging segment / discharging segment can only be compared with the reliable interval determined by the current segment, and when a battery cell's LOA value is abnormal, that is, the LOA value exceeds the corresponding reliable interval, the number of abnormalities of the battery cell is recorded once, and finally the cumulative number of abnormalities of all battery cells can be statistically obtained. The more times, the higher the degree of suspected fault of the battery cell. The specific evaluation criteria can be determined according to actual conditions and needs, and there are no excessive restrictions on this.

[0054] Finally, after determining the specific fault unit, the corresponding fault characteristics can be extracted for diagnosis based on the occurrence principles of different fault types, including typical fault types such as abnormal self-discharge and internal short circuit. Without making too many restrictions on this, modifications and improvements made to the embodiments of the present invention by those skilled in the art without departing from the spirit of the present invention will still fall within the scope of the invention patent application of the present invention.

[0055] In a specific embodiment, a connection fault of a power battery can be determined by the voltage difference between adjacent cells. The specific fault feature extraction method and type identification method are as follows:

[0056] Adjacent Cell Voltage Difference Extraction: Based on all the aforementioned charging and discharging segments, analyze the voltage inconsistency of the suspected faulty cell. After confirming the cell number of the suspected faulty cell, select the voltage data of the two adjacent battery cells before and after the suspected faulty cell as the center to compare the fault manifestations. The voltage data of the four adjacent battery cells are averaged and the difference between the average voltage and the voltage of the suspected faulty cell is calculated. This difference can be used to quantify the voltage inconsistency.

[0057] Fault type identification: Based on the voltage difference characteristics of adjacent cells, 5mV is used as the difference threshold to determine the fault time point. Analyze the difference performance of the charging segment and the discharging segment after the fault time point respectively. According to the formation mechanism of the connection fault, when a battery cell has a connection fault, the voltage of the faulty battery cell is lower than that of the other battery cells during the discharge behavior, and the voltage of the faulty battery cell is higher than that of the other battery cells during the charging behavior. Based on the voltage difference characteristics of adjacent cells, check whether the charging and discharging segments after the fault time point are greater than or less than 0. Voltage spike check: Compare the voltage of the suspected faulty cell with the total current of the power battery to check whether the voltage spike of the suspected faulty cell appears at the same frequency as the total current spike, so as to confirm whether the voltage mutation of the suspected faulty battery cell is caused by high-current charging due to kinetic energy recovery during vehicle driving.

[0058] For abnormal self-discharge, the battery SOC and the voltage change of the faulty cell are analyzed together to determine whether it is a self-discharge abnormality. The specific steps are as follows:

[0059] The voltage data of two adjacent battery cells, centered around the faulty cell, are selected for fault manifestation comparison. The voltage data of the four adjacent battery cells are averaged and subtracted from the voltage of the suspected faulty cell. This difference allows for quantification of voltage inconsistency. The voltage inconsistency differences obtained from the individual cells are combined with the overall power battery SOC to determine the changing trends of these differences when the vehicle is at high and low SOC, and calibration is performed. Based on the characteristics of the adjacent cell voltage differences, a 5mV difference threshold is used to determine the time of the fault. Analysis of SOC and voltage trends after the fault time point reveals that, based on the mechanism of self-discharge failure, high SOC increases cell inconsistency when a battery cell self-discharge occurs. In the high SOC operating range (SOC>60%), the absolute value of the adjacent voltage difference is between 15mV and 25mV; in the low SOC operating range (SOC<60%), the absolute value of the adjacent voltage difference is between 0mV and 15mV.

[0060] For sudden internal short circuit faults, by calculating the voltage difference between adjacent cells and the first-order difference variance of the cell voltage, combined with the battery temperature change trend, it is determined whether it is a sudden internal short circuit fault. The voltage difference between adjacent cells can be calculated in the same way as the extraction method of the connection fault feature, which will not be described in detail. Regarding the first-order difference variance feature of the cell voltage: first, calculate the first-order difference of each cell voltage to characterize the rate of change of the battery cell; calculate the cell voltage difference variance at all sampling moments in each abnormal segment. The larger the difference variance, the more asynchronous the cell voltage change in the power battery. Combine the above two feature change trends with the temperature change trend for comparison. If the temperature mutation is greater than a certain threshold, it is determined that a sudden internal short circuit fault has occurred.

[0061] Thus, the fault diagnosis method provided by this application can be used to diagnose whether a vehicle's power battery is operating normally or has a fault condition, and can implement a power battery early warning function, thereby maintaining the safety of the vehicle and personnel. Furthermore, it can further analyze the type of power battery fault and the cause of the fault to facilitate battery maintenance.

[0062] It should be noted that the step division of the various methods above is only for clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they contain the same logical relationship, they are all within the scope of protection of this application; adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the scope of protection of this application.

[0063] Example 2

[0064] See Figure 2 As shown, the embodiment of the present application also provides a vehicle power battery fault diagnosis system based on hierarchical identification, including:

[0065] The data collection module 10 is used to obtain the actual operation data of the vehicle.

[0066] The feature extraction module 20 is configured to divide the actual operation data into a plurality of data segments and extract one-dimensional or multi-dimensional feature quantities from the data segments.

[0067] The abnormality judgment module 30 is used to filter out abnormal data segments from all data segments according to the feature quantities of the data segments.

[0068] The fault identification module 40 calculates the longitudinal abnormal average value of the power battery cell voltage for the abnormal data segment to identify the battery cell suspected of failure in the power battery, and analyzes the fault type of the power battery based on the performance of the battery cell suspected of failure.

[0069] It should be noted that the vehicle power battery fault diagnosis system based on hierarchical identification provided in the above embodiment and the vehicle power battery fault diagnosis method based on hierarchical identification provided in the above embodiment 1 are based on the same concept, and the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here. In actual applications, the vehicle power battery fault diagnosis method based on hierarchical identification provided in the above embodiment 1 can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0070] Example 3

[0071] See Figure 3 As shown, an embodiment of the present application further provides an electronic device, comprising a memory 2, a processor 1, and a program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above methods when executing the program.

[0072] Among them, the memory includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory can be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory can also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the memory can also include both an internal storage unit of the electronic device and an external storage device. The memory can be used not only to store application software and various types of data installed in the electronic device, but also to temporarily store data that has been output or is to be output.

[0073] In some embodiments, the processor may be composed of an integrated circuit, such as a single packaged integrated circuit or a combination of multiple packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor is the control core (Control Unit) of the electronic device, connecting the various components of the entire electronic device using various interfaces and circuits. It executes or runs programs or modules stored in the memory and calls data stored in the memory to perform various functions of the electronic device and process data.

[0074] The processor executes the operating system of the electronic device and various installed application programs. The processor executes the application programs to implement the steps in the above method embodiment.

[0075] Exemplarily, the program may be divided into one or more modules, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules may be a series of program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the program in the electronic device.

[0076] The above-mentioned integrated unit implemented in the form of a software functional module can be stored in a computer-readable storage medium. The above-mentioned software functional module is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, computer equipment, or network equipment, etc.) or a processor to perform part of the functions of various embodiments of the present invention.

[0077] In summary, the present invention provides a vehicle power battery fault diagnosis method based on hierarchical identification. This method can rapidly identify faults in vehicles with only a few months of data. By analyzing actual operating data, it screens out abnormal segments and uses these to determine whether the vehicle's power battery is faulty. Furthermore, it can further identify suspected faulty cells within the power battery, facilitating subsequent analysis of the battery fault type and cause.

[0078] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, any equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. A vehicle power battery fault diagnosis method based on hierarchical identification, characterized in that: include: Obtain the actual operation data of the vehicle; Dividing the actual operation data into a plurality of data segments, and extracting one-dimensional or multi-dimensional feature quantities from the data segments; Screening out abnormal data segments from all data segments according to the feature quantities of the data segments, and the steps include: For any two data segments, the Euclidean distance between the different data segments with respect to the same feature quantity is calculated, and when the Euclidean distance is abnormal, the abnormality count of the two data segments is incremented by one; the abnormal data segment is identified based on the cumulative abnormality count of each data segment, and all abnormal data segments are integrated to diagnose whether the power battery is faulty; For abnormal data segments, the vertical abnormal average value of the power battery cell voltage is calculated to identify the battery cell suspected of failure in the power battery, and the fault type of the power battery is analyzed based on the performance of the battery cell suspected of failure; Among them, whether the Euclidean distance is abnormal is determined as follows; For any type of feature quantity, calculate the Euclidean distance between any two different data segments with respect to the feature quantity based on all data segments, until all Euclidean distances with respect to the feature quantity are obtained; Calculating corresponding quartiles, quartiles, and standard deviations based on all Euclidean distances of the feature quantities, and calculating first and second thresholds based on the quartiles, quartiles, and standard deviations as reference standards for determining whether the Euclidean distances of the feature quantities are abnormal; When the Euclidean distance is lower than the corresponding first threshold or higher than the corresponding second threshold, the Euclidean distance is considered abnormal.

2. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 1 is characterized in that: After obtaining the actual operating data of the vehicle, it also includes: The actual operation data is cleaned to remove failure values ​​in the original data.

3. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 1 is characterized in that: The steps of dividing the actual operation data into a plurality of data segments and extracting one-dimensional or multi-dimensional feature quantities from the data segments include: For any data segment, extract the corresponding multi-dimensional feature quantity based on the sampling parameters of the power battery; The sampling parameters of the power battery include: the total voltage of the power battery, the total current of the power battery, the state of charge of the power battery, the temperature extremes of the power battery, the insulation resistance of the power battery, and the single cell voltage of the power battery.

4. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 3 is characterized in that: For any data segment, the steps of extracting corresponding multi-dimensional feature quantities based on the sampling parameters of the power battery therein include: For the total voltage, total current, state of charge, or insulation resistance of the power battery, obtain the sampled values ​​at all times and calculate the mean, standard deviation, quartile, quartile, minimum, and maximum values ​​of the sampled values ​​as one of the one-dimensional feature quantities; For the temperature extremes of the power battery, obtain the sampled values ​​at all times and calculate the minimum, maximum, and average values ​​of the lowest sampled temperature, as well as the minimum, maximum, and average values ​​of the highest sampled temperature, as one of the one-dimensional feature quantities; For the single cell voltage of the power battery, the sampling values ​​at all times are obtained, and the maximum value and standard deviation of the single cell highest voltage, the average value and standard deviation of the single cell voltage range, and the average value and maximum value of the voltage entropy are calculated as one of the one-dimensional feature quantities.

5. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 1 is characterized in that: The calculation formulas of the first threshold and the second threshold are as follows: x1=Q1-1.5*Std, x2=Q2+1.5*Std, Wherein, x1 represents the first threshold, x2 represents the second threshold, Q1 represents the quartile, Q2 represents the quartile, and Std represents the standard deviation.

6. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 1 is characterized in that: For abnormal data segments, the steps of calculating the longitudinal abnormal average value of the power battery cell voltage to identify the battery cell suspected of failure in the power battery, and analyzing the power battery failure type based on the performance of the battery cell suspected of failure include: For each abnormal data segment, it is divided into several charging segments and discharging segments: Calculate the longitudinal abnormality average value of each cell of the power battery during each of the charging / discharging segments, and when the longitudinal abnormality average value is abnormal, increase the abnormality count of the corresponding cell by one; Identify suspected faulty battery cells based on the cumulative number of abnormalities of each power battery cell.

7. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 6 is characterized in that: The calculation formula for the longitudinal anomaly mean is as follows: Among them, LOA j Indicates the average longitudinal abnormality of the jth cell of the power battery, V ij Represents the cell voltage of the jth cell at the i-th moment, V imid It represents the median of all cell voltages at the i-th moment, and n represents the number of cells in the power battery.

8. The vehicle power battery fault diagnosis method based on hierarchical identification according to claim 7 is characterized in that: The steps for determining whether the longitudinal anomaly average is abnormal include: For each charging / discharging segment, a corresponding standard interval is determined using the three sigma principle; The longitudinal anomaly average value calculated using the charging segment / discharging segment is compared with the standard interval: If the longitudinal abnormal average value exceeds the standard range, the longitudinal abnormal average value is considered abnormal.

Images