A Battery Fault Detection Method Based on Quantile and Exponential Weighting and Adaptive Normalization

Abstract

This invention relates to a battery fault detection method based on quantile and exponential weighting adaptive normalization, belonging to the field of battery fault detection technology. The method includes the following steps: S1: Collecting electric vehicle operating data, including voltage data and time parameters of each cell in the battery pack, and preprocessing it; S2: Performing Kalman filtering on the cell voltage data; S3: Normalizing the filtered cell voltage data, including using quantiles to determine the normal value range, applying an exponential decay function to normal values, and applying an exponential growth function to outliers; S4: Calculating the DTW distance between each pair of normalized cell voltage time series and calculating the standard score of the distance; S5: Setting a fixed threshold and determining whether the standard score of the cell voltage exceeds the set threshold, thereby determining whether each cell battery has failed.

Description

Technical Field

[0001] This invention belongs to the field of battery fault detection technology, and relates to a battery fault detection method based on quantile and exponential weighting adaptive normalization. Background Technology

[0002] With the widespread application of electric vehicles, the performance and safety of power batteries, as their core components, have attracted much attention. Accurate and timely detection of battery faults is crucial for ensuring the reliable operation and lifespan of electric vehicles. Existing battery fault detection methods mainly rely on threshold judgments based on direct measurements such as voltage, current, and temperature, as well as some traditional data processing techniques, such as Z-score normalization and Min-Max normalization. These methods can identify abnormal battery states to some extent, but they have certain limitations. For example, Z-score normalization is sensitive to outliers, which may distort the normalized data and affect the accuracy of fault detection. Methods such as Min-Max normalization may not effectively amplify the differences between outliers and normal values, thus masking potential fault characteristics. Furthermore, existing normalization methods often use a uniform scale, failing to distinguish between normal values ​​and potential outliers, lacking differentiated modeling for the heterogeneity of both, exhibiting poor adaptability to differences between different battery cells, and failing to adequately handle the characteristic differences of individual cells within a battery pack. Moreover, they lack robustness in the face of complex operating conditions and noise interference, easily leading to false alarms or missed alarms.

[0003] To address the aforementioned issues, there is an urgent need for a more effective normalized real-time power battery fault detection method for electric vehicles that combines robustness with anomaly sensitivity. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a battery fault detection method based on quantile and exponential weighting with adaptive normalization, which can effectively suppress the interference of outliers on the overall distribution and enhance the expression of potential weak anomaly information, thereby improving the performance and stability of data-driven methods in battery fault detection, especially achieving online lightweight deployment without the need for a large amount of historical data and complex models.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A battery fault detection method based on quantile and exponential weighting with adaptive normalization includes the following steps:

[0007] S1: Collect electric vehicle operating data, including voltage data and time parameters of each cell in the battery pack, and perform preprocessing;

[0008] S2: Perform Kalman filtering on the individual battery cell voltage data;

[0009] S3: Normalize the filtered individual voltage data, including using quantiles to determine the normal range, applying an exponential decay function to the normal values, and applying an exponential growth function to the outliers.

[0010] S4: Calculate the DTW distance between each pair of normalized individual voltage time series and calculate the standard score of the distance;

[0011] S5: Set a fixed threshold and determine whether the standard score of the individual cell voltage exceeds the set threshold, thereby determining whether each individual cell has failed.

[0012] Furthermore, step S1 specifically includes the following steps:

[0013] S11: Collects normal operating data of electric vehicles, including total voltage, individual cell voltage, time parameters, and BMS alarm signals;

[0014] S12: Perform data processing and cleaning on the extracted unit voltages, and fill in missing values ​​in the data;

[0015] S13: Establish a power battery operation database based on the collected operating data of the power battery pack and individual battery cells.

[0016] Furthermore, step S2 specifically includes the following steps:

[0017] S21: Based on the processed individual unit voltage data, determine the initial state estimate and the estimation error covariance;

[0018] S22: Use the mean and variance of the first few data points as the initial state estimate, use the state estimate of the previous time step to predict the state at the current time step, and update the estimation error covariance;

[0019] S23: Calculate the Kalman gain based on the current measurement, correct the prediction state to obtain a more accurate estimate, and update the estimation error covariance;

[0020] S24: Filter the time series of each individual voltage, and output a smooth and stable data sequence at each time point during the filtering process.

[0021] Furthermore, step S3 specifically includes the following steps:

[0022] S31: Use quartiles to determine the normal range. The expression for calculating the lower limit of the normal range is as follows:

[0023] Q1-1.5·IQR

[0024] The expression for calculating the upper limit of the normal range interval is as follows:

[0025] Q3+1.5·IQR

[0026] Where Q1 represents the first quartile, which is the 25th percentile of the individual voltage data; Q3 represents the third quartile, which is the 75th percentile of the individual voltage data; IQR is the difference between the third quartile Q3 and the first quartile Q1 in the individual voltage data set, expressed as follows:

[0027] IQR = Q3 - Q1

[0028] S32: Based on the determined normal value range, determine the normal and abnormal values ​​of the individual unit voltage. Individual unit voltage data within the upper and lower limits of the normal value range are considered normal values, and individual unit voltage data outside the upper and lower limits of the normal value range are considered abnormal values.

[0029] S33: Apply the exponential decay function to the normal value, and the calculation expression is as follows:

[0030]

[0031] Among them, X i Z represents the voltage value of each individual cell i; median represents the median voltage of the cell, which is the cell voltage in the middle position after all cell voltage data are sorted from smallest to largest; α represents the normal value linear adjustment parameter, α>1; α X represents the linear scaling factor for normal values; normal_norm This represents the value of a single-unit voltage after exponential decay from its normal value.

[0032] S34: Apply the exponential growth function to the outliers, and the calculation expression is as follows:

[0033]

[0034] Where β represents the outlier linear adjustment parameter, 0 < β < 1; Z β X represents the linear scaling factor for outliers; outlier_norm This represents the value after the abnormal voltage of a single unit has increased exponentially.

[0035] S35: Combine the processed normal and outlier values ​​into a new normalized dataset.

[0036] Furthermore, the calculations of normal and abnormal values ​​are both based on the current point in time.

[0037] Furthermore, step S4 specifically includes the following steps:

[0038] S41: Apply the dynamic time warping algorithm to calculate the DTW distance between the normalized individual cell voltage time series at each time point; select the normalized median voltage value and compare it with the normalized individual cell voltage values ​​to realize battery fault detection;

[0039] S42: Calculate the standard score Z-score of the distance between each individual voltage curve and the median voltage curve, as shown in the following expression:

[0040]

[0041] Among them, X i_dtw This represents the distance between the voltage curve of individual cell i and the median voltage curve. σ represents the arithmetic mean of the distances between all individual voltage curves and the median voltage curve, σ represents the standard deviation of the distances between all individual voltage curves and the median voltage curve, and Z is the standard fraction of the distance between each individual voltage curve and the median voltage curve.

[0042] S43: Based on the standard score of the distance between each individual cell voltage curve and the median voltage curve, establish a real-time online fault detection model for power batteries based on voltage curve similarity.

[0043] Furthermore, the dynamic time warping algorithm described in step S41 is as follows:

[0044] Dynamic time warping uses dynamic programming to find the shortest path between two sequences. The calculation formula is shown below:

[0045] S(i,j)=d(x i ,y i )+min(S(i-1,j),S(i,j-1),S(i-1,j-1))

[0046]

[0047] Where S(i,j) represents the Euclidean distance between the i-th element of sequence X and the j-th element of sequence Y, and d(x i ,y i ) represents an element x in sequence X. i and the element y of sequence Y i The distance between them;

[0048] Each individual voltage sequence is selected as sequence X, and the median voltage sequence is selected as sequence Y;

[0049] Dynamic time warping backtracks from S(i,j) to S(1,1) to find the shortest path between two curves, representing the best match between the two curves on the time axis, thereby capturing the similarity between the curves.

[0050] Furthermore, in step S42, the arithmetic mean of the distances between all individual voltage curves and the median voltage curve is calculated. The expression obtained through the dynamic time warping algorithm is as follows:

[0051]

[0052] Where N is the total number of individuals;

[0053] The standard deviation σ of the distance between all individual voltage curves and the median voltage curve is obtained by a dynamic time warping algorithm, and the expression is as follows:

[0054]

[0055] Furthermore, step S5 specifically includes the following steps:

[0056] S51: Set an appropriate model threshold based on the standard fraction of the distance between each individual voltage curve and the median voltage curve;

[0057] S52: By judging whether the standard score of the distance between the real-time individual cell voltage curve and the median voltage curve exceeds the threshold set by the model, it is determined whether each individual cell has failed. If it exceeds the threshold, it indicates that a failure has occurred; otherwise, it indicates that no failure has occurred.

[0058] S53: If the battery malfunctions, locate the faulty cell based on the voltage of the faulty cell.

[0059] The beneficial effects of this invention are as follows:

[0060] 1) The proposed quantile normalization method standardizes battery measurement data based on the median and interquartile range, which can significantly improve the robustness of features to outliers, avoid the interference of extreme data on fault identification results, and improve the stability and accuracy of the data preprocessing stage.

[0061] 2) Introducing an exponentially enhanced normalization strategy, by applying higher normalization weights to suspected outliers outside the statistical boundary, the potential early minor faults are effectively amplified in the standardized feature space, thereby improving the identifiability of weak anomalies and facilitating fine-grained fault detection.

[0062] 3) The method proposed in this invention does not rely on a large number of historical samples or complex models. It is suitable for online fault detection scenarios during the operation of electric vehicles, has good real-time performance and engineering feasibility, and can accurately locate early anomalies at the cell level, thereby enhancing the overall safety assurance capability of the system.

[0063] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0064] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0065] Figure 1 This is a flowchart illustrating the overall method of the present invention;

[0066] Figure 2 Here is a diagram of the algorithm framework for an example;

[0067] Figure 3 The figures show the cell voltage curves of normal and faulty cells in the embodiments.

[0068] Figure 4 The Kalman filter curves of the voltage of the normal and faulty cells in the embodiment are shown.

[0069] Figure 5 The normalized voltage curves of the normal and faulty cells in the embodiments are shown.

[0070] Figure 6 The results are DTW fault detection results for normal and faulty cells in the embodiments. Detailed Implementation

[0071] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0072] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0073] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0074] Example 1:

[0075] like Figure 1-2 As shown, the present invention provides a battery fault detection method based on quantile and exponential weighting with adaptive normalization, which specifically includes the following steps:

[0076] Step S1: Collect electric vehicle operating data, including the voltage data of each cell in the battery pack and other characteristics, and perform data filtering and cleaning, specifically including the following steps:

[0077] Step S11: Collect normal operation data of a certain electric vehicle, including parameters such as total voltage, voltage of each individual cell, time, and BMS alarm signals;

[0078] Step S12: Perform data processing and cleaning on the extracted individual unit voltages, and fill in the missing values ​​in the data;

[0079] Data cleaning involves removing rows with a total voltage of 0, and missing data values ​​can be filled using the average value interpolation method between adjacent cells.

[0080] Step S13: Establish a power battery operation database based on the collected operating data of the power battery pack and individual battery cells.

[0081] Step S2: Perform Kalman filtering on the battery cell voltage data, specifically including the following steps:

[0082] Step S21: Based on the processed individual unit voltage data, determine the initial state estimate and the estimation error covariance;

[0083] Step S22: Use the mean and variance of the previous few data points as the initial state estimate, use the state estimate from the previous time step to predict the state at the current time step, and update the estimation error covariance.

[0084] Step S23: Calculate the Kalman gain based on the current measurement value, correct the prediction state to obtain a more accurate estimate, and update the estimation error covariance;

[0085] Step S24: Filter the voltage time series of each individual unit, and output a smooth and stable data sequence at each time point during the filtering process.

[0086] Step S3: Based on the filtered individual voltage data, perform normalization processing, including using quantiles to determine the normal value range, applying an exponential decay function to the normal values, and applying an exponential growth function to the outliers. Specifically, this includes the following steps:

[0087] Step S31: Use quartiles to determine the normal range. The expression for calculating the lower limit of the normal range is as follows:

[0088] Q1-1.5·IQR

[0089] The expression for calculating the upper limit of the normal range interval is as follows:

[0090] Q3+1.5·IQR

[0091] Where Q1 represents the first quartile, which is the 25th percentile of the individual voltage data; Q3 represents the third quartile, which is the 75th percentile of the individual voltage data; IQR is the difference between the third quartile (Q3) and the first quartile (Q1) in the individual voltage data set, expressed as follows:

[0092] IQR = Q3 - Q1

[0093] Step S32: Based on the determined normal value range, determine the normal and abnormal values ​​of the individual unit voltage. Individual unit voltage data within the upper and lower limits of the normal value range are considered normal values, and individual unit voltage data outside the upper and lower limits of the normal value range are considered abnormal values.

[0094] Step S33: Apply the exponential decay function to the normal value, and calculate the expression as follows:

[0095]

[0096] Among them, X i Z represents the voltage value of each individual cell i; median represents the median voltage of the cell, which is the cell voltage in the middle position after all cell voltage data are sorted from smallest to largest; α represents the normal value linear adjustment parameter, α>1; α X represents the linear scaling factor for normal values; normal_norm This represents the value of a single-unit voltage after exponential decay from its normal value.

[0097] Step S34: Apply the exponential growth function to the outlier, and calculate the expression as follows:

[0098]

[0099] Where β represents the outlier linear adjustment parameter, 0 < β < 1; Z β X represents the linear scaling factor for outliers; outlier_norm This represents the value after the abnormal voltage of a single unit has increased exponentially.

[0100] Step S35: Combine the processed normal values ​​and outliers into a new normalized dataset.

[0101] The calculations for normal and outlier values ​​are based on the current time point, including determining the range of normal values, applying an exponential decay function to the normal values, and applying an exponential growth function to the outlier values.

[0102] Step S4: Calculate the DTW distance between each pair of normalized individual voltage time series and calculate the standard score of the distance, specifically including the following steps:

[0103] Step S41: Apply the dynamic time warping algorithm to calculate the DTW distance between the normalized unit voltage time series at each time point;

[0104] To reduce the impact of abnormal individual battery cells on the average voltage, the normalized median voltage value is compared with the normalized voltage values ​​of each individual cell to achieve battery fault detection. The dynamic time warping algorithm flow is as follows:

[0105] Dynamic time warping uses dynamic programming to find the shortest path between two sequences. The calculation formula is shown below:

[0106] S(i,j)=d(x i ,y i )+min(S(i-1,j),S(i,j-1),S(i-1,j-1))

[0107]

[0108] Where S(i,j) represents the distance between the i-th element of sequence X and the j-th element of sequence Y, usually using Euclidean distance, d(x i ,y i ) represents an element x in sequence X. i and the element y of sequence Y i The distance between them.

[0109] Typically, the individual voltage sequences are selected as sequence X, and the median voltage sequence is selected as sequence Y.

[0110] Dynamic time warping backtracks from S(i,j) to S(1,1) to find the shortest path between two curves, representing the best match between the two curves on the time axis, thereby capturing the similarity between the curves.

[0111] Step S42: Calculate the standard score (Z-score) of the distance between each individual cell voltage curve and the median voltage curve, as shown in the following expression:

[0112]

[0113] Among them, X i_dtw This represents the distance between the voltage curve of individual cell i and the median voltage curve. σ represents the arithmetic mean of the distances between all individual voltage curves and the median voltage curve, σ represents the standard deviation of the distances between all individual voltage curves and the median voltage curve, and Z represents the standard fraction of the distance between each individual voltage curve and the median voltage curve.

[0114] Among them, the arithmetic mean of the distances between all individual voltage curves and the median voltage curve. It is obtained through the dynamic time warping algorithm, and the expression is as follows:

[0115]

[0116] Where N is the total number of individuals.

[0117] The standard deviation σ of the distance between all individual voltage curves and the median voltage curve is obtained through a dynamic time warping algorithm, and the expression is as follows:

[0118]

[0119] Step S43: Based on the standard score of the distance between each individual cell voltage curve and the median voltage curve, establish a real-time online fault detection model for power batteries based on voltage curve similarity.

[0120] Step S5: Set a fixed threshold. Based on whether the standard score of the individual cell voltage exceeds the set threshold, determine whether each individual cell has failed. This specifically includes the following steps:

[0121] Step S51: Set an appropriate model threshold based on the standard fraction of the distance between each individual voltage curve and the median voltage curve;

[0122] Step S52: By judging whether the standard score of the distance between the real-time individual cell voltage curve and the median voltage curve exceeds the threshold set by the model, it is determined whether each individual cell has failed. If it exceeds the threshold, it indicates that a failure has occurred; otherwise, it indicates that no failure has occurred.

[0123] Step S53: If the battery malfunctions, locate the faulty cell based on the voltage of the faulty cell.

[0124] The faulty unit is located by its unit number and the time when the unit failed. The time difference between the fault detection time and the warning time is calculated and displayed.

[0125] To illustrate the effectiveness of the present invention, the present invention provides operating data of a faulty vehicle, in which one battery cell experienced capacity decay and another battery cell malfunctioned and was alarmed by the BMS. Figure 3 This section presents the individual cell voltage curves for normal and faulty cells in the battery of a faulty vehicle. Kalman filtering and normalization were then applied to the data. Figure 4 The Kalman filter curves show the voltage of normal and faulty cells in the battery of the faulty vehicle. Figure 5 This section presents the normalized voltage curves of normal and faulty cells in the battery of the faulty vehicle. The normalized data is processed using the DTW algorithm, with a standard score detection threshold set to 19. Figure 6 The present invention provides DTW (Deep Time Warp) fault detection results for both normal and faulty cells in a faulty vehicle battery. It detects no false alarms for normal cells and identifies faults earlier than the BMS (Battery Management System) alarm time. Therefore, this invention effectively suppresses the interference of outliers on the overall distribution and enhances the representation of potential weak anomalies, enabling better real-time fault detection of electric vehicle power batteries and accurate location of faulty cells.

[0126] Example 2:

[0127] An electronic device, comprising a memory and a processor;

[0128] The memory is used to store computer programs;

[0129] The processor is configured to implement the method described in Embodiment 1 when executing the computer program.

[0130] Example 3:

[0131] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in Embodiment 1.

[0132] Example 4:

[0133] A computer program product includes a computer program that, when executed by a processor, implements the method described in Example 1.

[0134] In the above embodiments, the reference to "this embodiment" in the specification indicates that a specific feature, structure, or characteristic described in connection with the embodiment is included in at least some embodiments, but not necessarily all embodiments. Multiple appearances of "this embodiment" do not necessarily refer to the same embodiment.

[0135] In the above embodiments, although the invention has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory structures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed. The embodiments of the invention are intended to cover all such substitutions, modifications, and variations falling within the broad scope of the appended claims.

[0136] As will be understood by those skilled in the art, the computer-readable storage medium described in this embodiment allows for the implementation of all or part of the steps in the above method embodiments by computer program-related hardware. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0137] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication between them. The memory is used to store computer programs, the communication interface is used to perform communication, and the processor and the transceiver are used to run the computer programs, so that the electronic terminal performs the steps of the above method.

[0138] In this embodiment, the memory may include random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device.

[0139] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0140] This invention can be used in a wide range of general-purpose or special-purpose computing system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0141] This invention can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A battery fault detection method based on quantile and exponentially weighted adaptive normalization, characterized by: Includes the following steps: S1: Collect electric vehicle operating data, including voltage data and time parameters of each cell in the battery pack, and perform preprocessing; S2: Perform Kalman filtering on the individual battery cell voltage data; S3: Normalize the filtered individual voltage data, including using quantiles to determine the normal range, applying an exponential decay function to the normal values, and applying an exponential growth function to the outliers. S4: Calculate the DTW distance between each pair of normalized individual voltage time series and calculate the standard score of the distance; S5: Set a fixed threshold and determine whether the standard score of the individual cell voltage exceeds the set threshold, thereby determining whether each individual cell has failed. Step S3 specifically includes the following steps: S31: Use quartiles to determine the normal range. The expression for calculating the lower limit of the normal range is as follows: The expression for calculating the upper limit of the normal range interval is as follows: wherein, Q1 represents the first quartile, the 25th percentile of the monobloc voltage data; Q3 represents the third quartile, the 75th percentile of the monobloc voltage data; Q3 is the third quartile of the monobloc voltage data set the difference between the first quartile and the third quartile, expressed as follows: S32: Based on the determined normal value range, determine the normal and abnormal values ​​of the individual unit voltage. Individual unit voltage data within the upper and lower limits of the normal value range are considered normal values, and individual unit voltage data outside the upper and lower limits of the normal value range are considered abnormal values. S33: Apply the exponential decay function to the normal value, and the calculation expression is as follows: wherein, represents the voltage value of each monomer ; represents the median of the cell voltage, i.e. the cell voltage value located in the middle position after sorting all cell voltage data from small to large; represents the normal value linear adjustment parameter, ; represents the normal value linear scaling coefficient; represents the value of the normal value of the cell voltage after exponential decay. S34: Apply the exponential growth function to the outliers, and the calculation expression is as follows: in, This represents the linear adjustment parameter for outliers. ; This represents the linear scaling factor for outliers; This represents the value after the abnormal voltage of a single unit has increased exponentially. S35: Combine the processed normal and outlier values ​​into a new normalized dataset; Step S4 specifically includes the following steps: S41: Apply the dynamic time warping algorithm to calculate the DTW distance between the normalized individual cell voltage time series at each time point; select the normalized median voltage value and compare it with the normalized individual cell voltage values ​​to realize battery fault detection; S42: Calculate the standard score Z-score of the distance between each individual voltage curve and the median voltage curve, as shown in the following expression: in, Indicates monomer i The distance between the voltage curve and the median voltage curve. This represents the arithmetic mean of the distances between all individual voltage curves and the median voltage curve. This represents the standard deviation of the distance between all individual voltage curves and the median voltage curve. The standard fraction representing the distance between each individual voltage curve and the median voltage curve; S43: Based on the standard score of the distance between each individual cell voltage curve and the median voltage curve, establish a real-time online fault detection model for power batteries based on voltage curve similarity.

2. The battery fault detection method based on quantile and exponential weighting with adaptive normalization according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11: Collects normal operating data of electric vehicles, including total voltage, individual cell voltage, time parameters, and BMS alarm signals; S12: Perform data processing and cleaning on the extracted unit voltages, and fill in missing values ​​in the data; S13: Establish a power battery operation database based on the collected operating data of the power battery pack and individual battery cells.

3. The battery fault detection method based on quantile and exponential weighting with adaptive normalization according to claim 1, characterized in that: Step S2 specifically includes the following steps: S21: Based on the processed individual unit voltage data, determine the initial state estimate and the estimation error covariance; S22: Use the mean and variance of the first few data points as the initial state estimate, use the state estimate of the previous time step to predict the state at the current time step, and update the estimation error covariance; S23: Calculate the Kalman gain based on the current measurement, correct the prediction state to obtain a more accurate estimate, and update the estimation error covariance; S24: Filter the time series of each individual voltage, and output a smooth and stable data sequence at each time point during the filtering process.

4. The battery fault detection method based on quantile and exponential weighting with adaptive normalization according to claim 1, characterized in that: The calculations of normal and abnormal values ​​are based on the current point in time.

5. The battery fault detection method based on quantile and exponential weighting with adaptive normalization according to claim 1, characterized in that: The dynamic time warping algorithm described in step S41 is as follows: Dynamic time warping uses dynamic programming to find the shortest path between two sequences. The calculation formula is shown below: in, Represents a sequence The Elements and sequences The Euclidean distance between elements Represents a sequence elements and sequence elements The distance between them; Select the voltage sequence of each individual cell as the sequence The median voltage sequence is a sequence ; Dynamic time warping from Back to The goal is to find the shortest path between two curves, representing the best match between the two curves on the time axis, thereby capturing the similarity between the curves.

6. The battery fault detection method based on quantile and exponential weighting with adaptive normalization according to claim 1, characterized in that: In step S42, the arithmetic mean of the distances between all individual voltage curves and the median voltage curve is calculated. The expression obtained through the dynamic time warping algorithm is as follows: in, The total number of individual units; Standard deviation of the distance between all individual voltage curves and the median voltage curve The expression obtained through the dynamic time warping algorithm is as follows:

7. The battery fault detection method based on quantile and exponential weighting with adaptive normalization according to claim 1, characterized in that: Step S5 specifically includes the following steps: S51: Set an appropriate model threshold based on the standard fraction of the distance between each individual voltage curve and the median voltage curve; S52: By judging whether the standard score of the distance between the real-time individual cell voltage curve and the median voltage curve exceeds the threshold set by the model, it is determined whether each individual cell has failed. If it exceeds the threshold, it indicates that a failure has occurred; otherwise, it indicates that no failure has occurred. S53: If the battery malfunctions, locate the faulty cell based on the voltage of the faulty cell.

Images