A battery pack fault diagnosis method and system based on multi-domain features

Abstract

The present invention provides a battery pack fault diagnosis method and system based on multi-domain features, which relates to the field of battery management technology. The method includes: obtaining the median voltage curve of the battery pack and the single cell voltage curve of the single cell in the battery pack; decomposing the single cell voltage curve with the median voltage curve as the decomposition reference; optimizing the key parameters of the variational mode decomposition algorithm by the gray wolf optimization algorithm with the goal of minimizing the sum of the envelope entropies of the decomposed components of the decomposed single cell voltage curve; decomposing each single cell voltage curve according to the optimized variational mode decomposition algorithm to obtain a high-frequency fault information component; extracting the characteristics of the high-frequency fault information component; performing dimensionality reduction fusion on the time domain characteristics and the frequency domain characteristics through the principal component analysis algorithm to generate a comprehensive fault coefficient; identifying the faulty battery single cell in combination with the local gravity algorithm according to the time domain characteristics, the frequency domain characteristics and the comprehensive fault coefficient. The present invention can improve the fault diagnosis capability and help to give early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to a battery pack fault diagnosis method and system based on multi-domain features. Background Art

[0002] Multi-domain features refer to features in multiple different fields (such as time domain features, frequency domain features, etc.), and the battery pack fault diagnosis method uses multiple features (time domain features, frequency domain features, etc.) to comprehensively analyze the operating status of the battery pack, help identify and diagnose faults in the battery pack, and can capture the health status of the battery pack more comprehensively and accurately.

[0003] With the expansion of the new energy vehicle market and the increase in the number of electric vehicles, power ion battery safety accidents occur frequently, seriously threatening people’s lives and property safety. Therefore, fault diagnosis of battery packs can effectively avoid battery safety accidents, which is of great significance for saving lives, preventing property losses, reducing maintenance costs, extending the service life of battery packs, improving system reliability and stability, and promoting the advancement of battery technology.

[0004] However, due to the nonlinear and time-varying characteristics of battery voltage, the analysis of its abnormal changes faces great challenges. The existing battery pack fault diagnosis algorithms are relatively simple and usually rely on a certain feature (such as voltage, current, temperature, etc.). Especially when the battery pack state is complex and the environmental interference is strong, the accuracy and robustness are insufficient, thereby ignoring multi-dimensional information, resulting in reduced accuracy and efficiency of fault diagnosis, difficulty in handling complex nonlinear fault modes, increased labor costs, increased probability of misjudgment or missed judgment, and failure to detect faulty batteries in a timely manner. Summary of the invention

[0005] In order to solve the technical problems that the existing battery pack fault diagnosis algorithms in the prior art are relatively simple and usually rely on a certain feature (such as voltage, current, temperature, etc.), especially when the battery pack state is complex and the environmental interference is strong, the accuracy and robustness are insufficient, thereby ignoring multi-dimensional information, resulting in reduced accuracy and efficiency of fault diagnosis, difficulty in handling complex nonlinear fault modes, increased labor costs, increased probability of misjudgment or missed judgment, and failure to detect faulty batteries in time, the present invention provides a battery pack fault diagnosis method and system based on multi-domain features.

[0006] The technical solution provided by the embodiment of the present invention is as follows:

[0007] First aspect

[0008] An embodiment of the present invention provides a battery pack fault diagnosis method based on multi-domain features, including:

[0009] S1: Obtaining a median voltage curve of a battery pack and a single cell voltage curve of each single cell in the battery pack;

[0010] S2: Decomposing the single voltage curve using the median voltage curve as the decomposition benchmark;

[0011] S3: Taking the sum of the envelope entropies of the decomposed components of the single-cell voltage curve obtained by decomposition as the minimum, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm;

[0012] S4: Decompose each monomer voltage curve according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components;

[0013] S5: extracting time domain features and frequency domain features of high-frequency fault information components, wherein the time domain features include root mean square coefficients and pulse coefficients, and the frequency domain features include centroid frequency coefficients;

[0014] S6: Through the principal component analysis algorithm, the time domain features and frequency domain features are dimensionality reduced and integrated to generate a comprehensive fault coefficient;

[0015] S7: According to the time domain characteristics, frequency domain characteristics and comprehensive fault coefficient, the faulty battery cells are identified in combination with the local gravity algorithm.

[0016] Second aspect

[0017] An embodiment of the present invention provides a battery pack fault diagnosis system based on multi-domain features, including:

[0018] processor;

[0019] A memory having computer-readable instructions stored therein, which, when executed by a processor, implements the battery pack fault diagnosis method based on multi-domain features as in the first aspect.

[0020] The third aspect

[0021] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the battery pack fault diagnosis method based on multi-domain features as in the first aspect is implemented.

[0022] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0023] In the present invention, based on the median voltage curve of the battery pack and the single cell voltage curve of each single cell, the single cell voltage curve is decomposed, and the sum of the envelope entropies of the decomposed components of the single cell voltage curve obtained by decomposition is minimized as the goal, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm, and each single cell voltage curve is decomposed according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components, and time domain features and frequency domain features are further extracted. The time domain features and frequency domain features are reduced in dimension and fused by the principal component analysis algorithm to generate a comprehensive fault coefficient, and finally the extracted two-domain three-dimensional features and the comprehensive fault coefficient fused by the principal component analysis algorithm are input into the local gravity algorithm for fault detection, and the faulty battery cell is output. The present invention proposes a fault diagnosis method based on signal decomposition and multi-domain feature fusion. By fusing the three-dimensional features of time domain and frequency domain extracted from the battery voltage signal, the non-stationarity and nonlinear behavior of the battery can be accurately identified, the accuracy and reliability of fault diagnosis can be improved, the expression ability of fault characteristics can be enhanced, the influence of noise can be effectively suppressed, the sensitivity of the system to early faults such as thermal runaway can be enhanced, the fault diagnosis capability can be significantly improved, the labor cost can be reduced, the potential problems of the battery can be discovered in time, and early warning can be provided to ensure the safety of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 A schematic diagram of a flow chart of a battery pack fault diagnosis method based on multi-domain features provided by an embodiment of the present invention;

[0026] Figure 2 A schematic diagram of the structure of a battery pack fault diagnosis system based on multi-domain features provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0028] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0029] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0030] Reference Manual Attached Figure 1 , shows a flow chart of a battery pack fault diagnosis method based on multi-domain features provided in an embodiment of the present invention.

[0031] The embodiment of the present invention provides a battery pack fault diagnosis method based on multi-domain features, which can be implemented by a battery pack fault diagnosis device based on multi-domain features, and the battery pack fault diagnosis device based on multi-domain features can be a terminal or a server. The processing flow of the battery pack fault diagnosis method based on multi-domain features may include the following steps:

[0032] S1: Obtain a median voltage curve of a battery pack and a single cell voltage curve of each single cell in the battery pack.

[0033] Among them, the median voltage curve is a curve composed of the median values ​​at each moment based on the original voltage signals of multiple batteries. It represents a reference line of "normal state" and avoids the influence of noise or abnormal values ​​in certain battery signals. The single cell voltage curve records the voltage signal of a single battery cell changing over time, reflecting the real-time state of the single cell battery.

[0034] Optionally, the battery pack is a lithium-ion battery pack.

[0035] It should be noted that by obtaining the median voltage curve and single cell voltage curve of the battery pack, the interference of noise and abnormal values ​​in individual battery signals can be eliminated. The median voltage curve can be used as a stable "normal state" reference to effectively improve the signal's noise resistance and the reliability of analysis. In addition, this method can fully capture the state of single cells, and combined with overall trend analysis, it lays a solid foundation for subsequent fault diagnosis, especially in complex multi-battery systems with strong applicability and accuracy.

[0036] S2: Decompose the monomer voltage curve using the median voltage curve as the decomposition basis.

[0037] Specifically, the voltage signal of the battery is often affected by external interference (such as sudden fluctuations or noise). Directly processing the original signal may cause distortion of the result. Taking the median voltage as the benchmark is equivalent to using it to filter out these interferences and focus on deviations from the median (such as fault signals). Variational mode decomposition requires a stable starting point for decomposition, and the median voltage curve is used as a benchmark to help the decomposed signal more accurately reflect the battery status. All signal features (such as high-frequency parts) will be extracted relative to this median curve to ensure that the results are more representative. "Taking the median voltage as the benchmark" is to ensure the stability and reliability of signal decomposition. The median voltage curve provides a unified reference point, allowing variational mode decomposition to focus on extracting the truly important features of the signal without being disturbed by noise or outliers.

[0038] It should be noted that decomposing the cell voltage curve based on the median voltage curve provides an effective way to eliminate noise and interference in the signal while capturing key deviation features. Using the median voltage as a reference can stabilize the decomposition results and ensure that the decomposed signal accurately reflects the changes in the battery state, especially the extraction of abnormal signals (such as fault signals) is more accurate.

[0039] S3: With the goal of minimizing the sum of the envelope entropies of the decomposed components of the single-cell voltage curve, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm.

[0040] Among them, the decomposed components are signal components of different frequency ranges extracted from the original battery signal by signal decomposition algorithms (such as variational mode decomposition, VMD). Envelope entropy is an indicator to measure the complexity and randomness of the signal. The smaller the envelope entropy, the stronger the regularity of the signal, and the more it can reflect the key fault information. The gray wolf optimization algorithm is a bionic optimization algorithm that simulates the hunting behavior of gray wolf groups and finds the optimal solution to the problem through group collaboration and iterative updates. The variational mode decomposition algorithm is an adaptive signal decomposition algorithm used to extract different frequency components in the signal. Key parameters refer to important parameters in the variational mode decomposition algorithm, such as the number of decompositions (number of modes) and the penalty factor, which directly affect the quality of the decomposition results.

[0041] It should be noted that by taking the minimum envelope entropy as the goal and combining the Gray Wolf Optimization algorithm to optimize the key parameters of variational mode decomposition, adaptive adjustment of the decomposition parameters is achieved, which not only avoids the complexity of manual parameter adjustment, but also ensures that the decomposed signal is more regular and representative. The Gray Wolf Optimization algorithm is efficient and robust, and can quickly find the global optimal solution in a complex parameter space, thereby improving the accuracy and stability of signal decomposition.

[0042] In a possible implementation, key parameters include: decomposition quantity and penalty factor.

[0043] The number of decompositions refers to how many sub-modes or intrinsic mode functions the original signal is decomposed into. These sub-modes represent different frequency components in the signal. Each mode should capture part of the basic characteristics of the signal. The penalty factor is used to adjust the weight of the smoothing term in the algorithm. It is a regularization parameter used to control the bandwidth allocation in the modal decomposition process to ensure that the bandwidth of each mode remains within a reasonable range.

[0044] In a possible implementation, S3 includes:

[0045] S301: Optimizing key parameters using the envelope entropy of the decomposed components of each monomer voltage curve as the fitness function of the Grey Wolf optimization algorithm, wherein the sum of each decomposed component is equal to the monomer voltage signal corresponding to the monomer voltage curve:

[0046] ;

[0047] in, k represents the current number of decompositions of the variational mode decomposition algorithm, represents the current penalty factor of the variational mode decomposition algorithm, express k The optimal number of decompositions obtained after optimization, express The best penalty factor obtained after optimization, E p represents envelope entropy, and min represents the minimum value.

[0048] S302: Initialize key parameters of the variational mode decomposition algorithm through the Grey Wolf Optimization Algorithm, and set the optimization range of the key parameters.

[0049] S303: Calculate the fitness of each group of parameters within the optimization range.

[0050] S304: updating key parameters according to the fitness, and calculating the fitness of the updated parameters, wherein the updated parameters include the best parameters and the worst parameters.

[0051] S305: Determine whether the updated parameters have reached the maximum number of iterations. If so, output the optimal parameters. Otherwise, return to S303.

[0052] It should be noted that the Gray Wolf Optimization Algorithm has become an effective tool for solving complex optimization problems with its simple structure, powerful global search capability, strong adaptability and efficient optimization capabilities. It is particularly suitable for multi-dimensional, high-noise and nonlinear optimization scenarios, such as optimizing the parameters of the variational mode decomposition algorithm in battery pack fault diagnosis, which can significantly improve the diagnosis accuracy and efficiency.

[0053] S4: Decompose each monomer voltage curve according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components.

[0054] Among them, the high-frequency fault information component is a high-frequency signal component extracted by the variational mode decomposition algorithm, which contains possible fault characteristics and is the key data for battery fault diagnosis. By combining the optimized decomposition number and penalty factor, the VMD algorithm is used to decompose the single cell voltage curve to accurately extract high-frequency information containing fault characteristics, which can effectively filter out low-frequency noise and other irrelevant information, and highlight the signal characteristics related to the fault.

[0055] In a possible implementation, S4 specifically includes:

[0056] S401: Decomposing each monomer voltage curve by using a variational mode decomposition algorithm with an optimal decomposition quantity and an optimal penalty factor to extract a high-frequency component of each monomer voltage curve.

[0057] S402: Superimposing high-frequency components belonging to the same battery cell voltage curve to obtain high-frequency fault information components of each battery cell.

[0058] It should be noted that extracting the high-frequency fault information of the single battery after superimposing the high-frequency components can fully reflect the health status of each single battery and provide a more reliable basis for subsequent feature analysis. Compared with directly processing the original signal, it improves the accuracy and stability of data processing and significantly improves the sensitivity and robustness of battery fault diagnosis.

[0059] S5: extracting time domain features and frequency domain features of the high-frequency fault information component, wherein the time domain features include a root mean square coefficient and a pulse coefficient, and the frequency domain features include a centroid frequency coefficient.

[0060] Among them, the time domain feature is the characteristic of the signal in the time domain, which is used to describe the amplitude and fluctuation of the signal. The frequency domain feature is the characteristic of the signal in the frequency domain, which is used to describe the frequency distribution of the signal. The root mean square coefficient is used to measure the overall energy intensity of the signal. The pulse coefficient is used to evaluate abnormal peaks or fluctuations in the signal. The center of gravity frequency coefficient is the weighted average of the signal frequency, reflecting the center of gravity position of the spectrum.

[0061] It should be noted that by extracting time domain and frequency domain features from high-frequency fault information components, the abnormal characteristics of the signal can be characterized in multiple dimensions, thereby accurately locating the source of the fault. Time domain features (such as the root mean square coefficient and pulse coefficient) reveal the strength and fluctuation of the signal, while frequency domain features (such as the center of gravity frequency coefficient) provide global information on the frequency distribution. Combining time domain and frequency domain features can more comprehensively capture the characteristic signals of battery faults.

[0062] In a possible implementation, S5 specifically includes:

[0063] S501: Extract the root mean square coefficient and pulse coefficient of the time domain feature from the high-frequency fault information component:

[0064] ;

[0065] in, x rms represents the root mean square coefficient, n Indicates the size of the window, specifically 30. x i Indicates the window i The reconstructed signal value, I represents the pulse coefficient, Indicates the maximum value of the data in the window. Indicates the minimum value of the data in the window. represents the average value of the data in the window. Indicates the absolute value sign.

[0066] S502: Convert the time domain signal into a frequency domain signal through Fourier transform, and extract the centroid frequency coefficient of the frequency domain signal:

[0067] ;

[0068] in, represents the Fourier transform, x ( t )express t The reconstruction signal of the moment, represents a complex signal, e represents the base of natural logarithms, j represents the imaginary unit, represents the angular frequency, dt represents the differential element, FC represents the center of gravity frequency coefficient, represents the power spectrum of the signal, f Indicates frequency, represents the integral of the power spectrum weighted by all frequencies, represents the unweighted power spectrum integral, df Represents a differential element.

[0069] It should be noted that Fourier transform is a mathematical tool that converts time domain signals into frequency domain signals and is used to extract the frequency components of signals. This multi-domain feature extraction method enhances the reliability and sensitivity of diagnosis, especially for complex and subtle fault signals. It has a strong recognition ability, laying a solid foundation for subsequent fault classification and location.

[0070] S6: Through the principal component analysis algorithm, the time domain features and frequency domain features are reduced in dimension and fused to generate a comprehensive fault coefficient.

[0071] Among them, the principal component analysis algorithm is a statistical method used to reduce the dimension of the data while retaining the variability in the original data as much as possible. By extracting the main components in the data, the inherent structure of the data is discovered. The comprehensive failure coefficient refers to one or several quantitative indicators obtained after dimensionality reduction and data fusion processing. These indicators comprehensively reflect the failure status of the battery pack.

[0072] It should be noted that the main advantage of this method is that it can effectively simplify the data structure, reduce the complexity of data processing, and retain the important information in the original data to the maximum extent. Through dimensionality reduction, PCA helps identify which features are the most critical factors affecting battery performance, thereby generating a comprehensive fault coefficient. This not only improves the efficiency of data processing, but also enhances the sensitivity and accuracy of the model for fault detection.

[0073] In a possible implementation, S6 specifically includes:

[0074] S601: performing normalization processing on the root mean square coefficient, the pulse coefficient and the center of gravity frequency coefficient.

[0075] S602: According to the standardized coefficients, a covariance matrix is ​​constructed to analyze the correlation between the coefficients:

[0076] ;

[0077] in, X represents the original feature matrix consisting of the RMS coefficient, pulse coefficient, and centroid frequency coefficient, x mw Represents the first m Line w The elements of the column, x 1 represents the RMS coefficient, x 2 represents the pulse coefficient, x 3 represents the center of gravity frequency coefficient, O The number of samples representing the feature, , Indicates the total length of the segment sampling points, Indicates the window size, X stand represents the standardized feature matrix, x st1 represents the normalized root mean square coefficient, x st2 represents the normalized pulse coefficient, xst3 represents the normalized centroid frequency coefficient, C represents the covariance matrix, T represents transpose, represents the eigenvector of the covariance matrix, represents the eigenvalues ​​of the covariance matrix, The vector obtained by normalizing the eigenvector corresponding to the maximum eigenvalue. Represents the feature after dimensionality reduction, that is, the comprehensive failure coefficient.

[0078] It should be noted that the comprehensive fault coefficient, as a single metric, simplifies the evaluation and comparison of fault states, making the fault diagnosis process more intuitive and easy to understand. This method is particularly suitable for processing high-dimensional data, can significantly improve the speed and quality of diagnostic decisions, and provides an efficient fault prediction tool for battery management systems.

[0079] S7: According to the time domain characteristics, frequency domain characteristics and comprehensive fault coefficient, the faulty battery cells are identified in combination with the local gravity algorithm.

[0080] Among them, the local gravity algorithm is a heuristic optimization algorithm that evaluates and compares the relative differences between different objects (battery cells in this article) by simulating the gravitational force between objects.

[0081] Specifically, common battery failures include overvoltage, external short circuit, internal short circuit, sensor failure, etc.

[0082] It should be noted that there are many advantages to using the local gravity algorithm to identify faulty battery cells. First, this method uses time domain, frequency domain characteristics and comprehensive fault coefficients to evaluate the relative difference between each battery cell and other cells in the group, providing a dynamic and relative evaluation method. The introduction of the local gravity algorithm allows the health status of each cell in the battery pack to be compared with each other, and more intuitively and accurately identifies the unique faulty cells.

[0083] In actual applications, battery failures are usually caused by mechanical, electrical or thermal abuse during vehicle operation, or by the gradual deterioration of small defects introduced during the manufacturing process. The failure of a single cell not only leads to a degradation of the battery pack performance, but may also cause thermal runaway, leading to serious accidents such as fire and explosion.

[0084] In a possible implementation, S7 specifically includes:

[0085] S701: Combined with the local gravity algorithm, according to the time domain characteristics, frequency domain characteristics and comprehensive fault coefficient, calculate the gravitational force change rate of each single cell in the battery pack relative to other single cells in the group:

[0086] ;

[0087] in, Represents a monomer l At the same time relative to the total number of single batteries M The rate of change of the resultant gravitational force, M Indicates the total number of single cells in the battery pack. z j and z l Respectively represent single battery j and single battery l Fault characteristics at discrete moments, where the fault characteristics are time domain characteristics, frequency domain characteristics, or comprehensive fault coefficients.

[0088] S702: When the gravitational force change rate is greater than a preset gravitational force change rate, the corresponding single battery is output as a faulty battery single battery.

[0089] Among them, those skilled in the art can set the magnitude of the preset rate of change of the resultant gravitational force according to actual conditions, and the present invention does not limit this.

[0090] It should be noted that the concept of gravitational force provides a quantitative tool. By setting the threshold of the gravitational force change rate, the sensitivity of diagnosis can be flexibly adjusted to adapt to different application scenarios and requirements. Overall, it enhances the targeting and real-time nature of battery fault diagnosis and provides an effective fault monitoring and early warning mechanism for the battery management system.

[0091] In a possible implementation manner, after S7, the method further includes:

[0092] S8: Output corresponding warning according to the faulty battery cell.

[0093] Among them, a faulty battery cell refers to a single battery in the battery pack that is identified as having potential faults or has already failed after the fault diagnosis process. The early warning is a warning signal or notification issued in advance by the system for the detected faulty battery cell, which aims to notify maintenance personnel or system operators to take necessary maintenance measures or preventive measures to avoid the fault from worsening or affecting the safety and performance of the entire battery pack.

[0094] It should be noted that by giving a warning after identifying a faulty battery cell, not only the initiative and efficiency of battery pack fault handling are enhanced, but also solid support is provided for the efficient operation and safety of the battery management system.

[0095] Specifically, the present invention is further explained by taking the voltage data of a vehicle that suffered a thermal runaway failure and caught fire as an example.

[0096] First, the Gray Wolf Optimization Algorithm is used to optimize the key parameters of the variational mode decomposition (VMD) algorithm, taking the median voltage as the benchmark. The VMD algorithm decomposes the signal into k modal components with a central frequency, with the goal of minimizing the sum of their bandwidths and making the sum of all components equal to the original signal. The Gray Wolf Optimization Algorithm optimizes the parameters of the VMD algorithm by minimizing the envelope entropy of the decomposed components, simulating the hunting behavior of the gray wolf, and has a strong global search capability, which can effectively avoid local optimal solutions and finally obtain the optimized parameter combination. The optimized parameters are applied to the voltage curves of different monomers, and the high-frequency components are extracted and retained. The key parts are then superimposed to finally construct a high-frequency signal containing key fault information.

[0097] Next, the characteristic curve extracted from the high-frequency component curve is shown, with the time window size set to 30, including: RMS coefficient, pulse coefficient, center of gravity frequency coefficient, and comprehensive fault coefficient obtained after dimensionality reduction by principal component analysis algorithm. From the extracted characteristic curve, the fluctuation trend of the faulty monomer is obviously different from that of other monomers, indicating the existence of a fault.

[0098] Finally, based on the results of abnormal diagnosis of various features by the local gravity algorithm, since the amplitude of the abnormal diagnosis coefficient obtained is inconsistent, the time point that deviates from the overall data curve is used as the starting mark of fault detection. The VMD+root mean square coefficient algorithm can detect the fault of cell 76 at the 196th sampling point; the VMD+pulse coefficient algorithm can detect the fault of cell 76 at the 179th sampling point; the VMD+center of gravity frequency coefficient algorithm can detect the fault of cell 76 at the 181st sampling point; the VMD+comprehensive fault coefficient algorithm can detect the fault of cell 76 at the 170th sampling point. In the battery management system (BMS), the time point when the fault is discovered is the 293rd sampling point. In contrast, the algorithm combining signal decomposition with multi-domain feature fusion can detect the fault about 120 sampling points (about 20 minutes) in advance, thereby effectively improving safety and providing timely safety warnings. By comparison, it can be seen that the comprehensive fault coefficient algorithm can detect faults earlier than the single feature algorithm, indicating that the algorithm has obvious advantages in improving fault detection sensitivity and early warning capabilities.

[0099] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0100] In the present invention, based on the median voltage curve of the battery pack and the single cell voltage curve of each single cell, the single cell voltage curve is decomposed, and the sum of the envelope entropies of the decomposed components of the single cell voltage curve obtained by decomposition is minimized as the goal, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm, and each single cell voltage curve is decomposed according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components, and time domain features and frequency domain features are further extracted. The time domain features and frequency domain features are reduced in dimension and fused by the principal component analysis algorithm to generate a comprehensive fault coefficient, and finally the extracted two-domain three-dimensional features and the comprehensive fault coefficient fused by the principal component analysis algorithm are input into the local gravity algorithm for fault detection, and the faulty battery cell is output. The present invention proposes a fault diagnosis method based on signal decomposition and multi-domain feature fusion. By fusing the three-dimensional features of time domain and frequency domain extracted from the battery voltage signal, the non-stationarity and nonlinear behavior of the battery can be accurately identified, the accuracy and reliability of fault diagnosis can be improved, the expression ability of fault characteristics can be enhanced, the influence of noise can be effectively suppressed, the sensitivity of the system to early faults such as thermal runaway can be enhanced, the fault diagnosis capability can be significantly improved, the labor cost can be reduced, the potential problems of the battery can be discovered in time, and early warning can be provided to ensure the safety of the system.

[0101] Reference Manual Attached Figure 2 , showing a structural schematic diagram of a battery pack fault diagnosis system based on multi-domain features provided by the present invention.

[0102] The present invention further provides a battery pack fault diagnosis system 20 based on multi-domain features, which is applied to the above-mentioned battery pack fault diagnosis method based on multi-domain features, and comprises:

[0103] Processor 201.

[0104] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201 , a battery pack fault diagnosis method based on multi-domain features as in the method embodiment is implemented.

[0105] The battery pack fault diagnosis system 20 based on multi-domain features provided by the present invention can execute the above-mentioned battery pack fault diagnosis method based on multi-domain features and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0106] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0107] In the present invention, based on the median voltage curve of the battery pack and the single cell voltage curve of each single cell, the single cell voltage curve is decomposed, and the sum of the envelope entropies of the decomposed components of the single cell voltage curve obtained by decomposition is minimized as the goal, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm, and each single cell voltage curve is decomposed according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components, and time domain features and frequency domain features are further extracted. The time domain features and frequency domain features are reduced in dimension and fused by the principal component analysis algorithm to generate a comprehensive fault coefficient, and finally the extracted two-domain three-dimensional features and the comprehensive fault coefficient fused by the principal component analysis algorithm are input into the local gravity algorithm for fault detection, and the faulty battery cell is output. The present invention proposes a fault diagnosis method based on signal decomposition and multi-domain feature fusion. By fusing the three-dimensional features of time domain and frequency domain extracted from the battery voltage signal, the non-stationarity and nonlinear behavior of the battery can be accurately identified, the accuracy and reliability of fault diagnosis can be improved, the expression ability of fault characteristics can be enhanced, the influence of noise can be effectively suppressed, the sensitivity of the system to early faults such as thermal runaway can be enhanced, the fault diagnosis capability can be significantly improved, the labor cost can be reduced, the potential problems of the battery can be discovered in time, and early warning can be provided to ensure the safety of the system.

[0108] It should be understood that the processor in the embodiment of the present invention 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), field 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.

[0109] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0110] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0111] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0112] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0113] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0114] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0115] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0116] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0117] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0118] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0119] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0120] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, a battery pack fault diagnosis method based on multi-domain features as in the method embodiment is implemented.

[0121] A computer-readable storage medium provided by the present invention can implement the steps and effects of the battery pack fault diagnosis method based on multi-domain features of the above method embodiment. To avoid repetition, the present invention will not go into details.

[0122] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0123] In the present invention, based on the median voltage curve of the battery pack and the single cell voltage curve of each single cell, the single cell voltage curve is decomposed, and the sum of the envelope entropies of the decomposed components of the single cell voltage curve obtained by decomposition is minimized as the goal, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm, and each single cell voltage curve is decomposed according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components, and time domain features and frequency domain features are further extracted. The time domain features and frequency domain features are reduced in dimension and fused by the principal component analysis algorithm to generate a comprehensive fault coefficient, and finally the extracted two-domain three-dimensional features and the comprehensive fault coefficient fused by the principal component analysis algorithm are input into the local gravity algorithm for fault detection, and the faulty battery cell is output. The present invention proposes a fault diagnosis method based on signal decomposition and multi-domain feature fusion. By fusing the three-dimensional features of time domain and frequency domain extracted from the battery voltage signal, the non-stationarity and nonlinear behavior of the battery can be accurately identified, the accuracy and reliability of fault diagnosis can be improved, the expression ability of fault characteristics can be enhanced, the influence of noise can be effectively suppressed, the sensitivity of the system to early faults such as thermal runaway can be enhanced, the fault diagnosis capability can be significantly improved, the labor cost can be reduced, the potential problems of the battery can be discovered in time, and early warning can be provided to ensure the safety of the system.

[0124] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0125] There are a few points to note:

[0126] (1) The drawings of the embodiments of the present invention only involve structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0127] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or there may be intermediate elements.

[0128] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.

[0129] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A battery pack fault diagnosis method based on multi-domain features, characterized in that: include: S1: Acquire a median voltage curve of the battery pack and a single cell voltage curve of each single cell in the battery pack; S2: Decomposing the monomer voltage curve using the median voltage curve as a decomposition reference; S3: Taking the sum of the envelope entropies of the decomposed components of the single-cell voltage curve obtained by decomposition as the minimum, the key parameters of the variational mode decomposition algorithm are optimized by the Grey Wolf Optimization Algorithm; S4: Decompose each monomer voltage curve according to the optimized variational mode decomposition algorithm to obtain high-frequency fault information components; S5: extracting the time domain features and frequency domain features of the high-frequency fault information component, wherein the time domain features include a root mean square coefficient and a pulse coefficient, and the frequency domain features include a centroid frequency coefficient; S6: Performing dimensionality reduction fusion on the time domain features and the frequency domain features through a principal component analysis algorithm to generate a comprehensive fault coefficient; S7: Identify the faulty battery cell according to the time domain characteristics, the frequency domain characteristics and the comprehensive fault coefficient in combination with a local gravity algorithm; Wherein, the S3 includes: S301: Optimizing the key parameters by using the envelope entropy of the decomposed components of each cell voltage curve as the fitness function of the grey wolf optimization algorithm, wherein the sum of the decomposed components is equal to the cell voltage signal corresponding to the cell voltage curve: in, represents the current number of decompositions of the variational mode decomposition algorithm, represents the current penalty factor of the variational mode decomposition algorithm, express The optimal number of decompositions obtained after optimization, express After optimization The best penalty factor is obtained, represents the envelope entropy, Indicates taking the minimum value; S302: Initializing key parameters of the variational mode decomposition algorithm through the Grey Wolf Optimization Algorithm, and setting the optimization range of the key parameters; S303: Calculating the fitness of each group of parameters within the optimization range; S304: updating the key parameters according to the fitness, and calculating the fitness of the updated parameters, wherein the updated parameters include the best parameters and the worst parameters; S305: Determine whether the updated parameters have reached the maximum number of iterations. If so, output the optimal parameters. Otherwise, return to S303. Wherein, the S7 specifically includes: S701: In combination with the local gravitational algorithm, according to the time domain characteristics, the frequency domain characteristics and the comprehensive fault coefficient, the gravitational force change rate of each single cell in the battery pack relative to other single cells in the pack is calculated: in, Represents a monomer At the same time relative to the total number of single batteries The rate of change of the resultant gravitational force, Indicates the total number of single cells in the battery pack. and Represents single battery j and single battery l Fault characteristics at discrete moments, wherein the fault characteristics are time domain characteristics, frequency domain characteristics or comprehensive fault coefficients; S702: When the gravitational force change rate is greater than a preset gravitational force change rate, the corresponding single battery is output as a faulty battery single battery.

2. The battery pack fault diagnosis method based on multi-domain features according to claim 1 is characterized in that: The key parameters include: decomposition quantity and penalty factor.

3. The battery pack fault diagnosis method based on multi-domain features according to claim 1, characterized in that: The S4 specifically includes: S401: decomposing each monomer voltage curve by using a variational mode decomposition algorithm with the optimal decomposition number and the optimal penalty factor to extract a high-frequency component of each monomer voltage curve; S402: Superimposing high-frequency components belonging to the same battery cell voltage curve to obtain high-frequency fault information components of each battery cell.

4. The battery pack fault diagnosis method based on multi-domain features according to claim 1, characterized in that: The S5 specifically includes: S501: Extracting the root mean square coefficient and the pulse coefficient of the time domain feature from the high-frequency fault information component: in, represents the root mean square coefficient, Indicates the size of the window, specifically 30. Indicates the window i The reconstructed signal value, represents the pulse coefficient, Indicates the maximum value of the data in the window. Indicates the minimum value of the data in the window. represents the average value of the data in the window. Indicates the absolute value sign; S502: Convert the time domain signal into a frequency domain signal by Fourier transform, and extract the centroid frequency coefficient of the frequency domain signal: in, represents the Fourier transform, express The reconstruction signal of the moment, represents a complex signal, represents the base of natural logarithms, represents the imaginary unit, represents the angular frequency, represents the differential element, represents the center of gravity frequency coefficient, represents the power spectrum of the signal, Indicates frequency, represents the integral of the power spectrum weighted by all frequencies, represents the unweighted power spectrum integral, Represents a differential element.

5. The battery pack fault diagnosis method based on multi-domain features according to claim 1, characterized in that: The S6 specifically includes: S601: performing standardization processing on the root mean square coefficient, the pulse coefficient and the center of gravity frequency coefficient; S602: According to the standardized coefficients, a covariance matrix is ​​constructed to analyze the correlation between the coefficients: in, represents the original feature matrix consisting of the RMS coefficient, pulse coefficient, and centroid frequency coefficient, Represents the first m Line w The elements of the column, represents the root mean square coefficient, represents the pulse coefficient, represents the center of gravity frequency coefficient, The number of samples representing the feature, Indicates the total length of the segment sampling points, Indicates the window size, represents the standardized feature matrix, represents the normalized root mean square coefficient, represents the normalized pulse coefficient, represents the normalized centroid frequency coefficient, represents the covariance matrix, represents transpose, represents the eigenvector of the covariance matrix, represents the eigenvalues ​​of the covariance matrix, The vector obtained by normalizing the eigenvector corresponding to the maximum eigenvalue. Represents the feature after dimensionality reduction, that is, the comprehensive failure coefficient.

6. The battery pack fault diagnosis method based on multi-domain features according to claim 1 is characterized in that: The S7 also includes: S8: Outputting a corresponding warning according to the faulty battery cell.

7. A battery pack fault diagnosis system based on multi-domain features, characterized in that: include: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the battery pack fault diagnosis method based on multi-domain features as described in any one of claims 1 to 6 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a battery pack fault diagnosis method based on multi-domain features as described in any one of claims 1 to 6 is implemented.

Images