A method and system for diagnosing battery micro-short circuit faults

Abstract

The present invention belongs to the field of battery management technology, and specifically discloses a battery micro-short circuit fault diagnosis method and system, the method comprising: obtaining real-time voltage data of each battery in a battery module, denoising to obtain a voltage matrix; calculating a dynamic reference voltage sequence of each battery in the voltage matrix within a sliding window length, extracting the characteristic value of each battery at different times; calculating the correlation coefficient between the voltage of each battery at different times in the sliding window and the reference voltage; forming a characteristic point matrix of the battery module based on the characteristic value and the correlation coefficient; calculating a dynamic reference characteristic point sequence based on the characteristic point matrix, calculating an improved Fréchet distance between the characteristic point sequence of each battery and the dynamic reference characteristic point sequence, and using the distance value as an abnormal score of each battery; and then judging whether each battery has a short circuit fault. The present invention can effectively amplify fault features through feature extraction, has high sensitivity to micro short circuit faults, and can quickly detect and locate the fault location.

Description

Technical Field

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

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

[0003] In the practical applications of electric vehicles and energy storage systems, various internal and external faults may occur during the operation of the battery, 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 the short circuit may cause thermal runaway and result in serious safety accidents. Therefore, it is crucial to detect tiny short circuit faults in the initial stage of the short circuit fault in a timely manner to prevent its further development and ensure the safe and reliable operation of the battery.

[0004] Currently, the methods for diagnosing battery micro-short circuit faults mainly include the following categories:

[0005] (1) Model-based method: First, a non-linear model of the battery is constructed, and then the faults are diagnosed by using the anomalies of the model parameters or the estimated states. However, this method has problems such as high complexity, large computational amount and susceptibility to external interference.

[0006] (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 the non-linear fitting degree of a single signal processing method for battery data is relatively low, resulting in insufficient fault diagnosis accuracy.

[0007] (3) Machine learning-based method: Machine learning technologies, such as the powerful non-linear fitting ability of neural networks, are used to learn the battery fault patterns for fault diagnosis, with relatively high accuracy. However, this method highly depends on high-quality fault data, has poor effects when the data is lacking, and has high requirements for computing power and is difficult to be applied on embedded devices. Summary of the Invention

[0008] In order to solve the above problems, the present invention proposes a method and system for diagnosing battery micro-short circuit faults, which can quickly and accurately detect micro-short circuit faults in the initial stage of the short circuit fault while having low computational complexity, and ensure the safe operation of the battery system.

[0009] In some embodiments, the following technical solutions are adopted:

[0010] A method for diagnosing battery micro-short circuit faults includes:

[0011] Obtain the real-time voltage data of each battery in the electric vehicle battery module, and use variational mode decomposition for denoising to obtain a voltage matrix;

[0012] Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k. Based on the dynamic reference voltage sequence, extract the eigenvalue of each battery at different times; at the same time, calculate the correlation coefficient between the voltage of each battery at different times in the sliding window and the reference voltage; form a feature point matrix of the battery module based on the eigenvalue and the correlation coefficient;

[0013] Calculate the dynamic reference feature point sequence based on the feature point matrix, and calculate the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence. The distance value is used as the anomaly score of each battery;

[0014] Compare the anomaly score with a set threshold. If the anomaly score of a certain battery is greater than the set threshold, it is determined that the battery has a fault, and at the same time, send the fault judgment result to the electric vehicle battery management system BMS;

[0015] The electric vehicle battery management system BMS issues a fault alarm. At the same time, according to the received fault judgment result, block the battery that is about to have a fault or remove the faulty battery from the battery module.

[0016] Optionally, using variational mode decomposition for denoising to obtain a voltage matrix, specifically:

[0017] Use the variational mode decomposition method to process the voltage data of each battery in the battery module. The voltage sequence of each battery will obtain corresponding m modal components: { };

[0018] Remove the highest frequency band component among them Then perform an inverse transform on the remaining components to obtain the voltage sequence after denoising for each battery, and jointly form a voltage matrix:

[0019] ;

[0020] Among them, is the standard voltage matrix of the battery module voltage sequence, represents the voltage sequence of the i th battery, represents the voltage of the i th point in the sequence collected by the j th battery; i = 1, 2, …, n ; j = 1, 2, …, k ; n is the number of batteries,k is the sequence length extracted by the sliding window.

[0021] Optionally, calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, specifically:

[0022] Obtain the sum of voltages of each battery in the voltage matrix within the sliding window length k: ;

[0023] For the sum of voltages of each battery obtained make a comparison to find the batteries B max and B min corresponding to the maximum and minimum values, and remove the voltage sequences of these batteries from the voltage matrix and ;

[0024] Calculate the average value of each element in the voltage sequences of the remaining batteries at the same time:

[0025] ;

[0026] Thus, obtain the dynamic reference voltage sequence , indicating the dynamic reference voltage at the j th moment within the sliding window, indicating the voltage at the i th point in the sequence collected by the j th battery; i = 1, 2, …, n ; j = 1, 2, …, k ; n is the number of batteries, k is the sequence length extracted by the sliding window.

[0027] Optionally, extract the characteristic values of each battery at different moments, specifically:

[0028] ;

[0029] where is the characteristic value of the i th battery at the j th moment.

[0030] Optionally, calculate the correlation coefficient between the voltage and the reference voltage of each battery at different moments within the sliding window, specifically:

[0031] ;

[0032] where , are respectively thei Voltage sequence of a battery and the average value of the dynamic reference voltage sequence , is the i th battery at the j th moment of the correlation coefficient value, and are respectively the voltage sequence and the dynamic reference voltage sequence in the pth value, p = j - k + 1, j - k + 2, …, j.

[0033] Optionally, calculate the improved Fréchet distance between the eigenvalue sequence of each battery and the dynamic reference eigenvalue, specifically:

[0034] ;

[0035] wherein, represents the Fréchet distance between two trajectories, represents each battery feature point sequence within the sliding window , represents the dynamic reference feature point sequence obtained based on each battery feature point ; and =( ) respectively represent the ath point in the trajectory and the bth point in the trajectory ; a = 1, 2, …, g; b = 1, 2, …, h; g and h respectively represent the lengths of the battery feature point sequence and the dynamic reference feature point sequence; and are respectively the eigenvalue and the correlation coefficient value corresponding to the ath point in the trajectory , and are respectively the eigenvalue and the correlation coefficient value corresponding to the bth point in the trajectory ; respectively represent the trajectories composed of the remaining points after removing the gth point and the hth point in the trajectory and the trajectory ;

[0036] wherein, , , is the component and standard deviation, is the component and standard deviation.

[0037] Optionally, the process of determining the set threshold is specifically as follows:

[0038] Using the most recent complete battery cycle, calculate the anomaly scores within each sliding window to obtain all the anomaly scores for this battery cycle;

[0039] Based on the absolute median difference of the anomaly scores of each battery under a certain sliding window e j and the median M j , calculate the reference threshold for this sliding window T j as:

[0040] ;

[0041] Calculate the reference thresholds for each sliding window in the most recent complete battery cycle T j , and select the maximum value among them as the set threshold T;

[0042] If no battery failure is detected in the next complete battery cycle, update the set threshold T using the data of the next complete battery cycle.

[0043] In some other embodiments, the following technical solution is adopted:

[0044] A battery micro-short circuit fault diagnosis system, comprising:

[0045] A data acquisition module, configured to acquire the real-time voltage data of each battery in the electric vehicle battery module, and perform denoising using variational mode decomposition to obtain a voltage matrix;

[0046] A feature calculation module, configured to calculate the dynamic reference voltage sequence of each battery in the voltage matrix within a sliding window length k, and based on the dynamic reference voltage sequence, extract the eigenvalue of each battery at different moments; at the same time, calculate the correlation coefficient between the voltage of each battery at different moments within the sliding window and the reference voltage; form a feature point matrix of the battery module based on the eigenvalue and the correlation coefficient;

[0047] An anomaly score calculation module, configured to calculate a dynamic reference feature point sequence based on the feature point matrix, and calculate the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, and use the distance value as the anomaly score of each battery;

[0048] A fault judgment module, configured to compare the anomaly score with a set threshold, and if the anomaly score of a certain battery is greater than the set threshold, determine that the battery has a fault, and at the same time send the fault judgment result to the electric vehicle battery management system BMS;

[0049] The battery management system BMS of the electric vehicle performs fault alarm, and at the same time, according to the received fault judgment result, blocks the battery that will have a fault or removes the faulty battery from the battery module.

[0050] In some other embodiments, the following technical solutions are adopted:

[0051] A terminal device includes a processor and a memory. The processor is used to implement instructions; the memory is used to store multiple instructions, and the instructions are suitable for being loaded and executed by the processor to perform the above-mentioned battery micro-short circuit fault diagnosis method.

[0052] In some other embodiments, the following technical solutions are adopted:

[0053] A computer-readable storage medium stores multiple instructions, and the instructions are suitable for being loaded and executed by the processor of the terminal device to perform the above-mentioned battery micro-short circuit fault diagnosis method.

[0054] Compared with the prior art, the beneficial effects of the present invention are:

[0055] (1) Through feature extraction, the present invention can effectively amplify the fault features, has high sensitivity to micro-short circuit faults, can quickly detect the fault and locate the fault position after the fault occurs, that is, determine which specific battery of the electric vehicle has a fault.

[0056] (2) The present invention improves the Fréchet distance algorithm, modifies the kernel function therein, can amplify the fault features while reducing the inconsistent interference, and effectively improves the fault detection speed.

[0057] (3) The present invention proposes a calculation method for the dynamic reference voltage, and based on the dynamic reference voltage, eigenvalue extraction and correlation coefficient calculation are performed. The change law between the voltage sequence and the dynamic reference voltage sequence is characterized by the correlation coefficient, and the eigenvalue and the correlation coefficient are combined to form a feature point, which can more comprehensively reflect the battery abnormality to obtain a more reliable fault diagnosis result.

[0058] Based on the dynamic reference voltage, the present invention performs feature extraction and improved Fréchet distance calculation, which can effectively reduce the interference of the abnormal score calculation result by battery inconsistency or abnormal individual batteries, and improve the sensitivity to abnormal batteries; solves the problem that the calculation result is inaccurate due to the influence of inconsistent batteries or abnormal batteries when the prior art performs fault judgment calculation.

[0059] (4) The method developed by the present invention has the characteristics of data-driven, does not involve complex battery electrochemistry mechanisms, and can be conveniently used for the fault diagnosis of various different types of batteries without the need to construct different battery models in different situations like the model-based method.

[0060] Other features and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of this aspect. Brief Description of the Drawings

[0061] Figure 1 It is a flowchart of a battery micro-short circuit fault diagnosis method in an embodiment of the present invention. Detailed Embodiments

[0062] It should be noted that the following detailed description is illustrative and is 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 meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs.

[0063] 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 "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0064] Embodiment 1

[0065] In one or more embodiments, a battery micro-short circuit fault diagnosis method is disclosed, in combination with Figure 1 , which specifically includes the following processes:

[0066] S101: Obtain the real-time voltage data of each battery in the electric vehicle battery module, and use variational mode decomposition for denoising to obtain a voltage matrix.

[0067] In this embodiment, the number of batteries in the battery module n and the cycle condition are selected according to actual applications. The cycle conditions of the condition include the current rate of battery charging and discharging, discharge depth, temperature, 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 acquisition time.

[0068] The real-time voltage data refers to the latest voltage data generated by the operating battery, which can be collected through a voltage sensor.

[0069] Use the variational mode decomposition algorithm to process the voltage data of each battery in the battery module to obtain m (adjusted according to actual conditions) modal components: { };

[0070] Remove the highest frequency band component among them After that, an inverse transform is performed on the remaining components, and the denoising work can be completed, and finally the voltage matrix is obtained:

[0071] ;

[0072] Among them, is the standard voltage matrix of the battery module voltage sequence, represents the voltage sequence of the i th battery, represents the voltage of the i th point in the sequence collected by the j th battery; 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.

[0073] S102: Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k, and extract the eigenvalue of each battery at different times .

[0074] Among them, the dynamic reference voltage sequence will be updated according to the change of the real-time collected data.

[0075] In this embodiment, the process of calculating the dynamic reference voltage sequence is specifically as follows:

[0076] Obtain the sum of the voltages of each battery in the voltage matrix within the sliding window length k: ;

[0077] Compare the obtained sum of the voltages of each battery to find the batteries B max and B min corresponding to the maximum value and the minimum value, and remove the voltage sequences of the batteries in the voltage matrix and ;

[0078] Calculate the average value of each element in the voltage sequences of the remaining batteries at the same time:

[0079] ;

[0080] Thus, the dynamic reference voltage sequence is obtained; among them, represents the dynamic reference voltage at the j th moment within the sliding window.

[0081] Since inconsistent or abnormal batteries usually exhibit maximum or minimum values in terms of voltage, the dynamic reference voltage obtained by removing the maximum and minimum values and then calculating the average can better represent the state of normal battery cells in the battery module. Furthermore, when calculating the characteristic values of each battery, the impact of inconsistent or abnormal individual batteries on the calculation results can be reduced. Using the dynamic reference eigenvalue sequence for feature extraction and abnormal score calculation can effectively reduce the interference of the abnormal score by battery inconsistency or abnormal individual batteries and improve the sensitivity to abnormal batteries.

[0082] In this embodiment, the characteristic values of each battery at different times are extracted using the feature extraction formula; meanwhile, the correlation coefficient between the voltage of each battery at different times within the sliding window and the reference voltage is calculated; a feature point matrix of the battery module is formed based on the characteristic values and the correlation coefficient. F ;

[0083] Specifically, the feature extraction formula is specifically as follows:

[0084] ;

[0085] In the formula, is the characteristic value of the i th battery at the j th time, is the dynamic reference voltage at the j th time within the sliding window, is the i th battery, and j is the voltage value of the

[0086] In this embodiment, the feature extraction formula adopts an exponential form, and the exponential part is always positive. As the difference between the dynamic reference voltage and the battery voltage increases, the extracted characteristic value also increases significantly, enabling the characteristic value to more clearly represent the fault characteristics and facilitating the accurate detection of micro-short circuit faults.

[0087] Furthermore, the correlation coefficient between the voltage of each battery at different times within the sliding window and the reference voltage is calculated , specifically as follows:

[0088] ;

[0089] Among them, , are the averages of the voltage sequence i of the th battery and the dynamic reference voltage sequence respectively, is the correlation coefficient value of the i th battery at the j th time, and are the p-th values of the voltage sequence and the dynamic reference voltage sequence respectively, where p = j - k + 1, j - k + 2, …, j.

[0090] The change law between the -th battery voltage sequence i and the dynamic reference voltage sequence can be characterized by the correlation coefficient. When the battery is normal, the correlation coefficient is close to 1. After a fault occurs, the correlation between the voltage sequences decreases, resulting in a smaller correlation coefficient.

[0091] Combining ( , ) forms a characteristic point of the battery, and further forms the characteristic point matrix F of the battery module; this can more comprehensively reflect the battery abnormality to obtain a more reliable fault diagnosis result.

[0092] Therefore, the characteristic point matrix F is:

[0093] ;

[0094] where , representing the characteristic point sequence of the i -th battery from time 1 to time k, i = 1, 2, …, n .

[0095] S103: Calculate the dynamic reference characteristic points based on the characteristic point matrix, and calculate the improved Fréchet distance between the characteristic point sequences of each battery and the dynamic reference characteristic point sequence. The distance value is used as the abnormality score of each battery.

[0096] In this embodiment, for the obtained characteristic point matrix , the dynamic reference characteristic point is obtained; the specific calculation process is similar to the process of calculating the dynamic reference voltage, and the specific process is as follows:

[0097] Find the sum of the characteristics of each battery in the sliding window length k in the characteristic point matrix: ; 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.

[0098] Find the maximum and minimum values after the eigenvalues are added, remove the eigenvalue sequences corresponding to the maximum and minimum values, and average each element in the remaining eigenvalue sequences at the same time: , so as to obtain the eigenvalue at each moment;

[0099] ​​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 sequences at the same time:

[0100] , and obtain the correlation coefficient at each moment.

[0101] 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: .

[0102] Then, calculate the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence, specifically:

[0103] ;

[0104] where represents the Fréchet distance between two trajectories, represents each battery feature point sequence within the sliding window , represents the dynamic reference feature point sequence obtained based on the battery feature points ; thus, by calculating the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence can be obtained.

[0105] In the above formula, and =( ) respectively represent the a-th point in trajectory (i.e., the battery feature point sequence ) and the b-th point in trajectory (i.e., the dynamic reference feature point sequence ); a = 1, 2, …, g; b = 1, 2, …, h; g and h respectively represent the length of the battery feature point sequence and the length of the dynamic reference feature point sequence.

[0106] and are respectively the eigenvalue and the value of the correlation coefficient corresponding to the a-th point in trajectory , and are respectively the eigenvalue and the value of the correlation coefficient corresponding to the b-th point in trajectory , respectively represent the trajectories composed of the remaining points after removing the g-th point and the h-th point from trajectory and trajectory .

[0107] In this embodiment, the Fréchet distance has the same basic calculation form as that of the prior art. The main improvement lies in the calculation method of .

[0108] denotes and ; in the conventional Fréchet distance calculation, denotes calculating the Euclidean distance between the two:

[0109] ;

[0110] However, since the values in the feature points are calculated by different methods and the scales of the component values are different, it may lead to a certain error in the conventional Euclidean distance used, restricting the calculation difference of the Fréchet distance between the faulty battery and the normal battery, and being unfavorable for the identification of micro-short circuit faults.

[0111] Therefore, this embodiment improves the distance calculation method and uses the following distance formula:

[0112] ;

[0113] In the formula, is the standard deviation of the components and , is the standard deviation of the components and . This can normalize the scales of the two values in the feature points, ensure that the scales of the two are consistent, and ensure the calculation accuracy.

[0114] As a further solution, to increase the difference between the anomaly scores of the normal battery and the abnormal battery, is further improved by making , where is related to the lengths g and h of the eigenvalue sequences, specifically . In this embodiment, g = h = the sliding window length k.

[0115] The improved Fréchet distance algorithm of this embodiment can effectively amplify anomalies and improve the sensitivity of the anomaly score to faulty batteries. The value of q is related to the lengths g and h of the eigenvalue sequences, and can effectively curb the attenuation of the algorithm's sensitivity to anomalies when the sequence length increases.

[0116] This embodiment calculates the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence , and takes this distance as the anomaly score value of each battery:

[0117] ;

[0118] In the formula, is the anomaly score of the nth battery at time j, and its value is the improved Fréchet distance value calculated for the nth battery at time j.

[0119] S104: Compare the anomaly score with a set threshold to determine whether each battery has a short - circuit fault.

[0120] In this embodiment, the set threshold is calculated based on the absolute median difference and the median of the anomaly scores. The specific process is as follows:

[0121] S1041: Use the most recent complete battery cycle to calculate the anomaly score within each sliding window, obtaining all the anomaly scores for this battery cycle;

[0122] S1042: Based on the absolute median difference e j and the median M j of the anomaly scores of each battery under a certain sliding window, calculate the reference threshold T j for this sliding window as:

[0123] ;

[0124] S1043: Calculate the reference threshold T j for each sliding window in the most recent complete battery cycle, and select the maximum value as the set threshold T;

[0125] S1044: If no battery fault is detected in the next complete battery cycle, update the set threshold T using the data of the next complete battery cycle.

[0126] In this embodiment, the method of using the absolute median difference and the median to calculate the threshold ensures that a small number of outliers do not affect the final threshold setting, guarantees the accuracy of the threshold setting, is more sensitive to outliers, and improves the accuracy of detecting micro - short - circuit faults.

[0127] Compare the anomaly score obtained by processing the real - time voltage data with the threshold to complete the fault diagnosis. Among them, comparing the anomaly score obtained by processing the real - time voltage data with the threshold to complete the fault diagnosis is specifically: process the real - time collected data to obtain the anomaly score, and compare it with the set threshold; if the anomaly score of a certain battery is detected to be greater than the set threshold, it is determined that the battery has a fault.

[0128] Finally, it is possible to obtain the judgment result of whether there is a short - circuit fault point in each battery. This fault judgment result will be transmitted to the electric - vehicle battery management system BMS. The electric - vehicle battery management system BMS processes according to the fault judgment results of each battery. For example, it can give a fault alarm, or block the faulty battery or remove the faulty battery from the battery module, so as to detect short - circuit faults in time, handle them in time, avoid affecting other batteries, and ensure the normal operation of the battery module.

[0129] Certainly, the battery micro - short - circuit fault diagnosis method in this embodiment is not only applicable to the fault diagnosis of electric - vehicle battery modules, but also applicable to the fault diagnosis of battery modules in other energy - storage systems.

[0130] Embodiment 2

[0131] In one or more embodiments, a battery micro - short - circuit fault diagnosis system is disclosed, including:

[0132] A data acquisition module, which is used to acquire the real - time voltage data of each battery in the electric - vehicle battery module, and perform denoising using variational mode decomposition to obtain a voltage matrix;

[0133] A feature calculation module, which 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 eigenvalue 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 at different times and the reference voltage within the sliding window; form a feature - point matrix of the battery module based on the eigenvalue and the correlation coefficient;

[0134] An abnormal - score calculation module, which is used to calculate a dynamic reference feature - point sequence based on the feature - point matrix, calculate the improved Fréchet distance between the feature - point sequence of each battery and the dynamic reference feature - point sequence, and use the distance value as the abnormal score of each battery;

[0135] A fault judgment module, which is used to compare the abnormal score with a set threshold. If the abnormal score of a certain battery is greater than the set threshold, it is determined that the battery has a fault, and at the same time, send the fault judgment result to the electric - vehicle battery management system BMS;

[0136] The electric - vehicle battery management system BMS gives a fault alarm, and at the same time, according to the received fault judgment result, blocks the battery about to have a fault or removes the faulty battery from the battery module.

[0137] The specific implementation manners of the above - mentioned modules are the same as those in Embodiment 1 and will not be elaborated here.

[0138] Embodiment 3

[0139] In one or more embodiments, a terminal device is disclosed, which includes a processor and a memory. The processor is configured to implement instructions, and the memory is configured to store multiple instructions, and the instructions are adapted to be loaded and executed by the processor to perform the battery micro-short circuit fault diagnosis method described in the first embodiment.

[0140] 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 (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), 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.

[0141] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0142] In the implementation process, each step of the above method may be completed by the integrated logic circuit in the hardware of the processor or the instructions in the form of software.

[0143] Although the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. Those skilled in the art should understand that based on the technical solutions of the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts are still within the protection scope of the present invention.

Claims

1. A method for diagnosing battery micro-short circuit faults, characterized in that, Including: Obtain the real-time voltage data of each battery in the electric vehicle battery module, and use variational mode decomposition for denoising to obtain a voltage matrix; Calculate the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k. Based on the dynamic reference voltage sequence, extract the eigenvalue of each battery at different times; meanwhile, calculate the correlation coefficient between the voltage of each battery at different times and the reference voltage within the sliding window; form a feature point matrix of the battery module based on the eigenvalue and the correlation coefficient; The feature point matrix is: ； Among them, , represents the characteristic point sequence of the i -th battery from time 1 to time k, i = 1, 2, …, n ; Calculate the dynamic reference feature point sequence based on the feature point matrix, and calculate the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence. The distance value is used as the anomaly score of each battery; Compare the anomaly score with a set threshold. If the anomaly score of a certain battery is greater than the set threshold, it is determined that the battery has a fault, and at the same time, send the fault judgment result to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS issues a fault alarm. At the same time, according to the received fault judgment result, block the battery that is about to have a fault or remove the faulty battery from the battery module.

2. The battery micro-short circuit fault diagnosis method according to claim 1, wherein Use variational mode decomposition for denoising to obtain a voltage matrix, specifically: The variational mode decomposition method is used to process the voltage data of each battery in the battery module, and the voltage sequence of each battery will obtain the corresponding m modal components: { }; Remove the highest frequency band component among them After performing inverse transformation on the remaining components, the voltage sequences after denoising for each battery are obtained, jointly constituting a voltage matrix: ； Among them, is the standard voltage matrix of the battery module voltage sequence, represents the voltage sequence of the i th battery, represents the voltage of the i th point in the sequence collected by the j th battery; 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. The battery micro-short circuit fault diagnosis method according to 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 sum of the voltages of each battery in the voltage matrix within the sliding window length k: ; For each obtained battery voltage sum Compare them to find the batteries B corresponding to the maximum and minimum values max and B min , and remove the voltage sequences of the said batteries from the voltage matrix and ; Take the average value of each element in the voltage sequence of the remaining batteries at the same time: ； Thus, a dynamic reference voltage sequence is obtained , represents the dynamic reference voltage at the j -th moment within the sliding window, represents the voltage at the i -th point in the sequence collected from the j -th battery; 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. The method for diagnosing the battery micro-short circuit fault according to claim 3, wherein, Extract the eigenvalue of each battery at different times, specifically: ； Among them, is the i th eigenvalue of the j th battery at the 5. The battery micro-short circuit fault diagnosis method according to claim 1, characterized in that, Calculate the correlation coefficient between the voltage of each battery at different times and the reference voltage within the sliding window, specifically: ； Among them, , are respectively the average value of the voltage sequence of the i -th battery and the average value of the dynamic reference voltage sequence . , is the correlation coefficient value of the i -th battery at the j -th moment. and are respectively the p-th values in the voltage sequence and the dynamic reference voltage sequence , where p = j - k + 1, j - k + 2, …, j.

6. The battery micro-short circuit fault diagnosis method according to claim 1, wherein, Calculate the improved Fréchet distance between the eigenvalue sequence of each battery and the dynamic reference eigenvalue, specifically: ； Among them, represents the Fréchet distance between two trajectories, represents each sequence of battery feature points within the sliding window , represents the sequence of dynamic reference feature points obtained based on each battery feature point ; and =( ) represent the a-th point in trajectory and the b-th point in trajectory respectively; a = 1, 2, …, g; b = 1, 2, …, h; g and h represent the lengths of the battery feature point sequence and the dynamic reference feature point sequence respectively; and are the eigenvalue and the value of the correlation coefficient corresponding to the a-th point in trajectory respectively, and are the eigenvalue and the value of the correlation coefficient corresponding to the b-th point in trajectory respectively, represent the trajectories composed of the remaining points after removing the g-th point and the h-th point from trajectory and trajectory respectively; Among them, , , is the standard deviation of components and . is the standard deviation of components and .

7. The method for diagnosing the micro-short circuit fault of a battery according to claim 1, wherein, The determination process of the set threshold is specifically: Use the most recent complete battery cycle to calculate the anomaly score within each sliding window, and obtain all the anomaly scores in this battery cycle; Absolute median difference of the anomaly scores of each battery under a certain sliding window e j and the median M j , the reference threshold under the sliding window is calculated T j as follows: ； Calculate the reference threshold under each sliding window in the most recent complete battery cycle T j , and select the maximum value among them as the set threshold T; If no battery fault is detected in the next complete battery cycle, update the set threshold T using the data of the next complete battery cycle.

8. A battery micro-short circuit fault diagnosis system, characterized in that, Including: A data acquisition module for obtaining the real-time voltage data of each battery in the electric vehicle battery module, and using variational mode decomposition for denoising to obtain a voltage matrix; A feature calculation module for calculating the dynamic reference voltage sequence of each battery in the voltage matrix within the sliding window length k. Based on the dynamic reference voltage sequence, extract the eigenvalue of each battery at different times; meanwhile, calculate the correlation coefficient between the voltage of each battery at different times and the reference voltage within the sliding window; form a feature point matrix of the battery module based on the eigenvalue and the correlation coefficient; The feature point matrix is: ； Among them, , representing the sequence of characteristic points of the i -th battery from time 1 to time k, i = 1, 2, …, n ; An anomaly score calculation module for calculating the dynamic reference feature point sequence based on the feature point matrix, and calculating the improved Fréchet distance between the feature point sequence of each battery and the dynamic reference feature point sequence. The distance value is used as the anomaly score of each battery; A fault judgment module, which is used to compare the abnormal score with a set threshold. If the abnormal score of a certain battery is greater than the set threshold, it is determined that the battery has a fault, and at the same time, the fault judgment result is sent to the electric vehicle battery management system BMS; The electric vehicle battery management system BMS performs a fault alarm. At the same time, according to the received fault judgment result, the battery that is about to have a fault is blocked or the faulty battery is removed from the battery module.

9. A terminal device, comprising a processor and a memory, the processor being configured to implement instructions; the memory being configured to store a plurality of instructions, characterized in that, The instruction is suitable for being loaded and executed by a processor to perform the battery micro-short circuit fault diagnosis method according to any one of claims 1-7.

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

Images