Battery micro short circuit fault diagnosis method and system

Abstract

The invention belongs to the technical field of battery management, and particularly discloses a battery micro-short-circuit fault diagnosis method and system, and the method comprises the steps: obtaining the real-time voltage data of each battery in a battery module, and carrying out the denoising of the real-time voltage data, and obtaining a voltage matrix; calculating a dynamic reference voltage sequence of each battery in the voltage matrix within the length of the sliding window, and extracting characteristic values of each battery at different moments; calculating a correlation coefficient between the voltage of each battery in the sliding window at different moments and the reference voltage; forming a feature point matrix of the battery module based on the feature values and the correlation coefficients; calculating a dynamic reference feature point sequence based on the feature point matrix, calculating an improved Frechet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, and taking the distance value as an abnormal score of each battery; and whether each battery has a short-circuit fault or not is judged. Fault features can be effectively amplified through feature extraction, sensitivity to tiny short circuit faults is high, and fault occurrence positions can be rapidly detected and positioned.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to a battery micro-short circuit fault diagnosis method and system. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In the actual application of electric vehicles and energy storage systems, various internal and external faults may occur during battery operation, leading to performance problems and even serious consequences such as thermal runaway, fire or explosion. Among them, short circuit is the most common and representative safety failure mode, and one of the most important factors causing battery failure and safety problems. The continuous development and intensification of short circuit may cause thermal runaway and cause serious safety accidents. Therefore, it is crucial to detect small short circuit faults in time in the early stage of short circuit faults and prevent them from further development to ensure the safe and reliable operation of batteries.

[0004] At present, there are mainly the following methods for diagnosing micro-short circuit faults of batteries: (1) Model-based method: First, a nonlinear model of the battery is constructed, and then the model parameters or estimated state anomalies are used for fault diagnosis. However, this method has problems such as high complexity, large amount of calculation, and susceptibility to external interference.

[0005] (2) Signal processing-based method: Signal processing methods, such as correlation coefficient and sample entropy, are used to process and analyze battery data such as voltage and current for fault diagnosis. However, this method has the problem that a single signal processing method has a low nonlinear fit for battery data, resulting in insufficient fault diagnosis accuracy.

[0006] (3) Machine learning-based methods: This method uses machine learning techniques such as neural networks to learn battery failure modes and perform fault diagnosis, which is highly accurate. However, this method is highly dependent on high-quality failure data, and has poor results when data is lacking. It also has high computing power requirements and is difficult to apply on embedded devices. Summary of the invention

[0007] In order to solve the above problems, the present invention proposes a battery micro-short circuit fault diagnosis method and system, which has low computational complexity and can quickly and accurately detect micro-short circuit faults in the early stage of short circuit faults to ensure the safe operation of the battery system.

[0008] In some embodiments, the following technical solutions are adopted: A battery micro-short circuit fault diagnosis method, comprising: Obtain the real-time voltage data of each battery in the electric vehicle battery module, and use variational mode decomposition to denoise and obtain the voltage matrix; Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the characteristic value of each battery at different times based on the dynamic reference voltage sequence; at the same time, calculate the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage; form a characteristic point matrix of the battery module based on the characteristic value and the correlation coefficient; Calculating a dynamic reference feature point sequence based on the feature point matrix, calculating a modified Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, and using the distance value as an abnormality score of each battery; Compare the abnormality score with a set threshold value. If the abnormality score of a battery is greater than the set threshold value, the battery is determined to be faulty, and the fault judgment result is sent to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS issues a fault alarm and, based on the received fault judgment results, blocks the battery that is about to fail or removes the failed battery from the battery module.

[0009] Optionally, variational mode decomposition is used to perform denoising to obtain a voltage matrix, specifically: The voltage data of each battery in the battery module is processed using the variational mode decomposition method, and the voltage sequence of each battery will obtain corresponding m modal components: }; Remove the highest frequency component Then, the remaining components are inversely transformed to obtain the denoised voltage sequence of each battery, which together form the voltage matrix: ; in, It is the standard voltage matrix of the battery module voltage sequence. Indicates i The voltage sequence of each battery, Indicates i The first battery in the sequence of collection j The voltage at each point; i =1,2,…, n ; j =1,2,…, k ; n is the number of batteries, k is the length of the sequence extracted by the sliding window.

[0010] Optionally, a dynamic reference voltage sequence of each battery in the voltage matrix within a sliding window length k is calculated, specifically: Calculate the voltage sum of each battery in the voltage matrix within the sliding window length k: ; For each battery voltage and Compare and find the battery B corresponding to the maximum and minimum values max and B min , and remove the voltage sequence of the battery in the voltage matrix and ; Find the average value of each element in the voltage series of the remaining batteries at the same time: ; Thus, the dynamic reference voltage sequence is obtained , Indicates the first j The dynamic reference voltage at each moment, Indicates i The first battery in the sequence of collection j The voltage at each point; i =1,2,…, n ; j =1,2,…, k ; n is the number of batteries, k is the length of the sequence extracted by the sliding window.

[0011] Optionally, extract the characteristic value of each battery at different times, specifically: ; in, It is i The battery in j The characteristic value at a moment.

[0012] Optionally, the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage is calculated, specifically: ; in, , Respectively i Voltage sequence of the battery The average value and dynamic reference voltage sequence The average value of It is i The battery in j The correlation coefficient value at each moment is and The voltage series and dynamic reference voltage sequence The p-th value in , p=j-k+1,j-k+2,…,j.

[0013] Optionally, the modified Fréchet distance between the eigenvalue sequence of each battery and the dynamic reference eigenvalue is calculated, specifically: ; in, express The Fréchet distance between two trajectories, Represents each battery feature point sequence in the sliding window , Represents a dynamic reference feature point sequence obtained based on each battery feature point ; and =( ) represent the trajectory The ath point and trajectory in The b-th point in ; a=1,2,…,g; b=1,2,…,h; g and h represent the length of the battery feature point sequence and the length of the dynamic reference feature point sequence respectively; and The trajectory The eigenvalue and correlation coefficient corresponding to the a-th point in , and The trajectory The eigenvalue and correlation coefficient corresponding to the b-th point in , Respectively represent the trajectory and trajectory The trajectory of the remaining points after removing the g-th point and the h-th point; in, , , For the weight and The standard deviation of For the weight and The standard deviation of .

[0014] Optionally, the process of determining the set threshold is specifically as follows: Using the most recent complete battery cycle, calculate the anomaly score in each sliding window to obtain all anomaly scores under the battery cycle; Based on the absolute median difference of the abnormality score of each battery in a sliding window e j and the median M j , calculate the reference threshold under the sliding window T j for: ; Calculate the reference threshold for each sliding window in the most recent complete battery cycle T j , select the maximum value as the threshold value T; If no battery failure is detected in the next complete battery cycle, the set threshold T is updated using the next complete battery cycle data.

[0015] In other embodiments, the following technical solutions are adopted: A battery micro-short circuit fault diagnosis system, comprising: The data acquisition module is used to obtain the real-time voltage data of each battery in the battery module of the electric vehicle, and use variational mode decomposition to perform denoising to obtain the voltage matrix; A feature calculation module is used to calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the characteristic value of each battery at different times based on the dynamic reference voltage sequence; at the same time, calculate the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage; and form a characteristic point matrix of the battery module based on the characteristic value and the correlation coefficient; an abnormality score calculation module, used to calculate a dynamic reference feature point sequence based on the feature point matrix, and calculate a modified Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, wherein the distance value is used as an abnormality score of each battery; A fault judgment module is used to compare the abnormality score with a set threshold value. If the abnormality score of a battery is greater than the set threshold value, the battery is judged to be faulty and the fault judgment result is sent to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS issues a fault alarm and, based on the received fault judgment results, blocks the battery that is about to fail or removes the failed battery from the battery module.

[0016] In other embodiments, the following technical solutions are adopted: A terminal device includes a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store multiple instructions, and the instructions are suitable for being loaded by the processor and executing the above-mentioned battery micro-short circuit fault diagnosis method.

[0017] In other embodiments, the following technical solutions are adopted: A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded and executed by a processor of a terminal device for the above-mentioned battery micro-short circuit fault diagnosis method.

[0018] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention can effectively amplify fault features through feature extraction and has high sensitivity to small short-circuit faults. It can quickly detect the fault and locate the fault location after the fault occurs, that is, determine which specific battery of the electric vehicle has a fault.

[0019] (2) The present invention improves the Fréchet distance algorithm by modifying the kernel function therein, which can reduce inconsistent interference while amplifying fault characteristics, thereby effectively improving the fault detection speed.

[0020] (3) The present invention proposes a method for calculating a dynamic reference voltage, and performs feature value extraction and correlation coefficient calculation based on the dynamic reference voltage. The correlation coefficient is used to characterize the variation law between the voltage sequence and the dynamic reference voltage sequence. The feature value and the correlation coefficient are combined to form a feature point, which can more comprehensively reflect the battery abnormality and obtain a more reliable fault diagnosis result.

[0021] The present invention performs feature extraction and improved Fréchet distance calculation on the basis of a dynamic reference voltage, which can effectively reduce the interference of battery inconsistency or abnormal individual batteries on the abnormal score calculation results, and improve the sensitivity to abnormal batteries; it solves the problem that the prior art is easily affected by inconsistent batteries or abnormal batteries when performing fault judgment calculations, resulting in inaccurate calculation results.

[0022] (4) The method developed in the present invention has data-driven characteristics and does not involve complex battery electrochemical mechanisms. It can be conveniently used for fault diagnosis of various types of batteries without the need to construct different battery models in different situations like model-based methods.

[0023] Other features and advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart of a battery micro-short circuit fault diagnosis method in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0026] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0027] Embodiment 1 In one or more embodiments, a battery micro short circuit fault diagnosis method is disclosed, combining Figure 1 , specifically including the following process: S101: Acquire real-time voltage data of each battery in the battery module of the electric vehicle, and use variational mode decomposition to perform denoising to obtain a voltage matrix.

[0028] In this embodiment, the number of batteries in the battery module n The cycle conditions are selected according to the actual application. The cycle conditions of the working conditions include the current rate of battery charging and discharging, discharge depth, temperature and cut-off voltage, etc. The collected battery test data includes but is not limited to the voltage data of each battery in the module and the corresponding data collection time.

[0029] Real-time voltage data refers to the latest voltage data generated by the running battery, which can be collected and obtained through a voltage sensor.

[0030] The voltage data of each battery in the battery module is processed using the variational mode decomposition algorithm to obtain m (adjusted according to actual conditions) modal components: }; Remove the highest frequency component After that, the remaining components are inversely transformed to complete the denoising work and finally obtain the voltage matrix: ; in, It is the standard voltage matrix of the battery module voltage sequence. Indicates i The voltage sequence of each battery, Indicates i The first battery in the sequence of collection j The voltage at each point; i =1,2,…, n ; j =1,2,…, k ; n is the number of batteries, k is the length of the sequence extracted by the sliding window.

[0031] S102: Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the characteristic value of each battery at different times based on the dynamic reference voltage sequence. .

[0032] Among them, the dynamic reference voltage sequence will be updated according to the changes of real-time collected data.

[0033] In this embodiment, the process of calculating the dynamic reference voltage sequence is specifically as follows: Calculate the voltage sum of each battery in the voltage matrix within the sliding window length k: ; For each battery voltage and Compare and find the battery B corresponding to the maximum and minimum values max and B min , and remove the voltage sequence of the battery in the voltage matrix and ; Find the average value of each element in the voltage series of the remaining batteries at the same time: ; Thus, the dynamic reference voltage sequence is obtained ;in, Indicates the first j The dynamic reference voltage at a given moment.

[0034] Because inconsistent or abnormal batteries usually show maximum or minimum values ​​in voltage, the dynamic reference voltage obtained by removing the maximum and minimum values ​​and taking the average value can better represent the state of normal battery cells in the battery module, thereby reducing the impact of inconsistent batteries or abnormal individual batteries on the calculation results when calculating the characteristic values ​​of each battery. Using the dynamic reference characteristic value sequence for feature extraction and abnormal score calculation can effectively reduce the interference of battery inconsistency or abnormal individual batteries on the abnormal score and improve the sensitivity to abnormal batteries.

[0035] In this embodiment, the feature extraction formula is used to extract the eigenvalue of each battery at different times; at the same time, the correlation coefficient between the voltage of each battery at different times and the reference voltage in the sliding window is calculated; based on the eigenvalue and the correlation coefficient, the feature point matrix of the battery module is formed. F ; Specifically, the feature extraction formula is: ; In the formula, It is i The battery in j The characteristic value at the moment, is the first jThe dynamic reference voltage at each moment, It is i The battery in j The voltage value at a moment.

[0036] The feature extraction formula of this embodiment adopts an exponential form, and the exponential part is always a positive value. As the difference between the dynamic reference voltage and the battery voltage increases, the extracted eigenvalue also increases significantly, so that the eigenvalue can more clearly characterize the fault characteristics, which is conducive to the accurate detection of micro-short circuit faults.

[0037] Furthermore, the correlation coefficient between the voltage of each battery at different times in the sliding window and the reference voltage is calculated. , specifically: ; in, , Respectively i The voltage sequence of the battery The average value and dynamic reference voltage sequence The average value of It is i Batteries in j The correlation coefficient value at time, and The voltage series and dynamic reference voltage sequence The p-th value in , p=j-k+1,j-k+2,…,j.

[0038] By correlation coefficient Can characterize the i Battery voltage sequence and dynamic reference voltage sequence When the battery is normal, the correlation coefficient will be close to 1. After a fault occurs, the correlation between the voltage sequences will decrease, resulting in a smaller correlation coefficient.

[0039] Will( , ) together constitute a characteristic point of the battery, and then form the characteristic point matrix F of the battery module; this can more comprehensively reflect the battery abnormality so as to obtain more reliable fault diagnosis results.

[0040] Therefore, the feature point matrix F is: ; in, , indicating the i The characteristic point sequence of a battery from time 1 to time k, i =1, 2, …, n .

[0041] S103: Calculate dynamic reference feature points based on the feature point matrix, calculate the modified Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, and use the distance value as the abnormality score of each battery.

[0042] In this embodiment, for the feature point matrix obtained , find the dynamic reference feature points ; The specific calculation process is similar to the process of calculating the dynamic reference voltage, and the specific process is as follows: Calculate the sum of the features of each battery in the feature point matrix within the sliding window length k: ; When summing, each eigenvalue component is added, and each correlation coefficient component is added, that is, the eigenvalues ​​of battery i at each moment are added, and the correlation coefficients of battery i at each moment are added.

[0043] Find the maximum and minimum values ​​after adding the eigenvalues, remove the eigenvalue sequences corresponding to the maximum and minimum values, and average the elements in the remaining eigenvalue sequences at the same time: , thus obtaining the eigenvalue at each moment; Similarly, find the maximum and minimum values ​​after adding the correlation coefficients, remove the correlation coefficient sequences corresponding to the maximum and minimum values, and calculate the average value of each element in the remaining correlation coefficient sequence at the same time: , and get the correlation coefficient at each moment.

[0044] The eigenvalue and correlation coefficient at each moment together constitute the dynamic reference feature point at that moment, thus forming a sequence of dynamic reference feature points at each moment: .

[0045] Then, the improved Fréchet distance between the characteristic point sequence of each battery and the dynamic reference characteristic point sequence is calculated, specifically: ; in, express The Fréchet distance between two trajectories, Represents each battery feature point sequence in the sliding window , Represents a dynamic reference feature point sequence obtained based on each battery feature point ; thus by calculating Then the improved Fréchet distance between the characteristic point sequence of each battery and the dynamic reference characteristic point sequence can be obtained. In the above formula, and =( ) represent the trajectory (i.e. battery feature point sequence ) in the ath point and trajectory (i.e., dynamic reference feature point sequence ), a=1,2,…,g, b=1,2,…,h, g and h represent the length of battery feature point sequence and the length of dynamic reference feature point sequence, respectively.

[0046] and The trajectory The eigenvalue and correlation coefficient corresponding to the a-th point in , and The trajectory The eigenvalue and correlation coefficient corresponding to the b-th point in , Respectively represent the trajectory and trajectory The trajectory of the remaining points after removing the g-th point and the h-th point.

[0047] In this embodiment, the Fréchet distance The basic calculation form is the same as the prior art, and the main improvement is The calculation method of .

[0048] express and The distance between; in the conventional Fréchet distance calculation, Indicates calculating the Euclidean distance between the two: ; However, since the values ​​in the feature points are calculated by different methods and the scales of the component values ​​are different, the conventional Euclidean distance used may have certain errors, which limits the calculated difference in the Fréchet distance between faulty batteries and normal batteries, and is not conducive to the identification of micro-short circuit faults.

[0049] Therefore, this embodiment improves the distance calculation method and uses the following distance formula: ; In the formula, For the weight and The standard deviation of For the weight and This can normalize the scales of the two values ​​in the feature point to ensure that the scales of the two are consistent and the accuracy of the calculation is guaranteed.

[0050] As a further solution, in order to increase the difference between the abnormal scores of normal batteries and abnormal batteries, To make further improvements, ,in, It is related to the length g and h of the eigenvalue sequence, specifically In this embodiment, g=h=sliding window length k.

[0051] The improved Fréchet distance algorithm of this embodiment can effectively amplify anomalies and improve the sensitivity of anomaly scores to faulty batteries. The value of q is related to the lengths g and h of the eigenvalue sequence, which can effectively curb the attenuation of the algorithm's sensitivity to anomalies when the sequence length increases.

[0052] In this embodiment, the characteristic point sequence of each battery is calculated. With dynamic reference feature point sequence The improved Fréchet distance between them is used as the abnormality score of each battery: ; In the formula, is the abnormal score of the n-th battery at time j, and its value is the improved Fréchet distance value calculated for the n-th battery at time j.

[0053] S104: Compare the abnormality score with a set threshold value to determine whether a short circuit fault occurs in each battery.

[0054] In this embodiment, the threshold is set based on the absolute median difference and median of the abnormal score. The specific process is: S1041: Using the most recent complete battery cycle, calculate the anomaly score in each sliding window to obtain all anomaly scores in the battery cycle; S1042: Absolute median difference of abnormal scores of each battery in a sliding window e j and the median M j , calculate the reference threshold under the sliding window T j for: ; S1043: Calculate the reference threshold value of each sliding window in the most recent complete battery cycle T j , select the maximum value as the threshold value T; S1044: If no battery failure is detected in the next complete battery cycle, the set threshold T is updated using the next complete battery cycle data.

[0055] In this embodiment, the method of using the absolute median difference and the median to calculate the threshold value does not affect the final threshold setting due to a small number of abnormal values, thereby ensuring the accuracy of the threshold setting, being more sensitive to abnormal values, and improving the accuracy of micro-short circuit fault detection.

[0056] The abnormality score obtained after processing the real-time voltage data is compared with the threshold to complete the fault diagnosis. Specifically, the abnormality score obtained after processing the real-time voltage data is compared with the threshold to complete the fault diagnosis: the real-time collected data is processed to obtain the abnormality score, and the score is compared with the set threshold; if the abnormality score of a battery is detected to be greater than the set threshold, the battery is determined to be faulty.

[0057] Eventually, the judgment result of whether a short circuit fault occurs in each battery can be obtained, and the fault judgment result will be transmitted to the electric vehicle battery management system BMS. The electric vehicle battery management system BMS processes the fault judgment result of each battery. For example, a fault alarm can be issued, or the faulty battery can be blocked or removed from the battery module, so as to timely discover the short circuit fault and handle it in time to avoid affecting other batteries and ensure the normal operation of the battery module.

[0058] Of course, the battery micro-short circuit fault diagnosis method of this embodiment is not only applicable to the fault diagnosis of electric vehicle battery modules, but also applicable to the fault diagnosis of other energy storage system battery modules.

[0059] Embodiment 2 In one or more embodiments, a battery micro-short circuit fault diagnosis system is disclosed, comprising: The data acquisition module is used to obtain the real-time voltage data of each battery in the battery module of the electric vehicle, and use variational mode decomposition to perform denoising to obtain the voltage matrix; A feature calculation module is used to calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the characteristic value of each battery at different times based on the dynamic reference voltage sequence; at the same time, calculate the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage; and form a characteristic point matrix of the battery module based on the characteristic value and the correlation coefficient; an abnormality score calculation module, used to calculate a dynamic reference feature point sequence based on the feature point matrix, and calculate a modified Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, wherein the distance value is used as an abnormality score of each battery; A fault judgment module is used to compare the abnormality score with a set threshold value. If the abnormality score of a battery is greater than the set threshold value, the battery is judged to be faulty and the fault judgment result is sent to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS issues a fault alarm and, based on the received fault judgment results, blocks the battery that is about to fail or removes the failed battery from the battery module.

[0060] The specific implementation of the above modules is the same as that in Example 1 and will not be described in detail.

[0061] Embodiment 3 In one or more embodiments, a terminal device is disclosed, which includes a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store multiple instructions, and the instructions are suitable for being loaded by the processor and executing the battery micro-short circuit fault diagnosis method described in Example 1.

[0062] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0063] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0064] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0065] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A battery micro-short circuit fault diagnosis method, characterized in that: include: Obtain the real-time voltage data of each battery in the electric vehicle battery module, and use variational mode decomposition to denoise and obtain the voltage matrix; Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the characteristic value of each battery at different times based on the dynamic reference voltage sequence; at the same time, calculate the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage; Forming a characteristic point matrix of the battery module based on the characteristic values ​​and the correlation coefficients; Calculating a dynamic reference feature point sequence based on the feature point matrix, calculating a modified Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, and using the distance value as an abnormality score of each battery; Compare the abnormality score with a set threshold value. If the abnormality score of a battery is greater than the set threshold value, the battery is determined to be faulty, and the fault judgment result is sent to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS issues a fault alarm and, based on the received fault judgment results, blocks the battery that is about to fail or removes the failed battery from the battery module.

2. A battery micro-short circuit fault diagnosis method as claimed in claim 1, characterized in that: The voltage matrix is ​​obtained by denoising using variational mode decomposition, which is: The voltage data of each battery in the battery module is processed using the variational mode decomposition method, and the voltage sequence of each battery will obtain corresponding m modal components: }; Remove the highest frequency component Then, the remaining components are inversely transformed to obtain the denoised voltage sequence of each battery, which together form the voltage matrix: ； in, It is the standard voltage matrix of the battery module voltage sequence. Indicates i The voltage sequence of each battery, Indicates i The first battery in the sequence of collection j The voltage at each point; i =1,2,…, n ; j =1,2,…, k ; n is the number of batteries, k is the length of the sequence extracted by the sliding window.

3. A battery micro-short circuit fault diagnosis method as claimed in claim 1, characterized in that: Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, specifically: Calculate the voltage sum of each battery in the voltage matrix within the sliding window length k: ; For each battery voltage and Compare and find the battery B corresponding to the maximum and minimum values max and B min , and remove the voltage sequence of the battery in the voltage matrix and ; Find the average value of each element in the voltage series of the remaining batteries at the same time: ； Thus, the dynamic reference voltage sequence is obtained , Indicates the first j The dynamic reference voltage at each moment, Indicates i The first battery in the sequence of collection j The voltage at each point; i =1,2,…, n ; j =1,2,…, k ; n is the number of batteries, k is the length of the sequence extracted by the sliding window.

4. A battery micro-short circuit fault diagnosis method as claimed in claim 3, characterized in that: Extract the characteristic values ​​of each battery at different times, specifically: ； in, It is i The battery in j The characteristic value at a moment.

5. A battery micro-short circuit fault diagnosis method as claimed in claim 1, characterized in that: Calculate the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage, specifically: ； in, , Respectively i The voltage sequence of the battery The average value and dynamic reference voltage sequence The average value of It is i The battery in j The correlation coefficient value at each moment is and The voltage series and dynamic reference voltage sequence The p-th value in , p=j-k+1,j-k+2,…,j.

6. A battery micro-short circuit fault diagnosis method as claimed in claim 1, characterized in that: The improved Fréchet distance between the eigenvalue sequence of each battery and the dynamic reference eigenvalue is calculated as follows: ； in, express The Fréchet distance between two trajectories, Represents each battery feature point sequence in the sliding window , Represents a dynamic reference feature point sequence obtained based on each battery feature point ; and =( ) represent the trajectory The ath point and trajectory in The b-th point in ; a=1,2,…,g; b=1,2,…,h; g and h represent the length of the battery feature point sequence and the length of the dynamic reference feature point sequence respectively; and The trajectory The eigenvalue and correlation coefficient corresponding to the a-th point in , and The trajectory The eigenvalue and correlation coefficient corresponding to the b-th point in , Respectively represent the trajectory and trajectory The trajectory of the remaining points after removing the g-th point and the h-th point; in, , , For the weight and The standard deviation of For the weight and The standard deviation of .

7. A battery micro-short circuit fault diagnosis method as claimed in claim 1, characterized in that: The process of determining the threshold is specifically as follows: Using the most recent complete battery cycle, calculate the anomaly score in each sliding window to obtain all anomaly scores under the battery cycle; Based on the absolute median difference of the abnormality score of each battery in a sliding window e j and the median M j , calculate the reference threshold under the sliding window T j for: ； Calculate the reference threshold for each sliding window in the most recent complete battery cycle T j , select the maximum value as the threshold value T; If no battery failure is detected in the next complete battery cycle, the set threshold T is updated using the next complete battery cycle data.

8. A battery micro-short circuit fault diagnosis system, characterized in that: include: The data acquisition module is used to obtain the real-time voltage data of each battery in the battery module of the electric vehicle, and use variational mode decomposition to perform denoising to obtain the voltage matrix; A feature calculation module is used to calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the feature value of each battery at different times based on the dynamic reference voltage sequence; at the same time, calculate the correlation coefficient between the voltage of each battery in the sliding window at different times and the reference voltage; Forming a characteristic point matrix of the battery module based on the characteristic values ​​and the correlation coefficients; an abnormality score calculation module, used to calculate a dynamic reference feature point sequence based on the feature point matrix, and calculate a modified Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, wherein the distance value is used as an abnormality score of each battery; A fault judgment module is used to compare the abnormality score with a set threshold value. If the abnormality score of a battery is greater than the set threshold value, the battery is judged to be faulty and the fault judgment result is sent to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS issues a fault alarm and, based on the received fault judgment results, blocks the battery that is about to fail or removes the failed battery from the battery module.

9. A terminal device, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the battery micro-short circuit fault diagnosis method described in any one of claims 1-7.

10. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instruction is suitable for being loaded by a processor of a terminal device and executing the battery micro-short circuit fault diagnosis method described in any one of claims 1-7.

Images