Fault diagnosis method of power battery pack and related equipment

Abstract

The invention relates to a fault diagnosis method for a power battery pack and related equipment, and the method comprises the following steps: carrying out the electrochemical impedance spectroscopy analysis of the power battery pack based on a phase characteristic spectrum, and obtaining an impedance characteristic spectrum of the power battery pack; whether the power battery pack has a fault is diagnosed based on the impedance characteristic spectrum of the power battery pack; if the impedance characteristic spectrum exists, performing fault characteristic mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault characteristic vector of the power battery pack; and performing fault positioning analysis on the power battery pack based on the fault feature vector to obtain a fault positioning result of the power battery pack, thereby solving the problems that the traditional method mainly depends on simple voltage and current monitoring and periodic maintenance inspection, the efficiency is low, and the reliability is high. And the technical problem that irreversible damage is caused when a fault occurs due to the fact that internal potential problems often cannot be found and processed in time is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power battery packs, and in particular to a fault diagnosis method for a power battery pack and related equipment. Background Art

[0002] In the context of the rapid development of electric vehicles and renewable energy systems, the performance and reliability of power battery packs, as one of the core components of these technologies, are directly related to the stability and safety of the entire system. However, with the continuous advancement of battery technology, traditional fault diagnosis methods have been unable to meet the needs of modern high-performance battery packs. Traditional methods mainly rely on simple voltage and current monitoring and periodic maintenance inspections, which are not only inefficient, but also often fail to detect and handle potential internal problems in a timely manner, resulting in irreversible damage when a fault occurs.

[0003] In addition, the complex structure of power battery packs, which contain many single battery cells and complex thermal management systems, makes accurate status assessment and fault diagnosis extremely challenging. Especially in the face of factors such as battery aging, temperature changes, and differences in usage environments, how to accurately identify various failure modes that may occur in battery packs and effectively monitor their health status has become an important topic in current research. Existing diagnostic methods usually lack an in-depth understanding of the dynamic processes inside the battery, making it difficult to provide comprehensive and accurate fault information, which limits the optimization and development of battery management systems.

[0004] In response to the above problems, researchers have begun to explore more advanced fault diagnosis technologies, aiming to improve the fault detection accuracy and response speed of power battery packs. By adopting advanced analysis methods such as phase spectrum analysis and electrochemical impedance spectroscopy analysis, we can have a deeper understanding of the working state of the battery pack and the physical and chemical changes occurring inside it, thereby achieving early warning and precise positioning of faults. Such methods can overcome the limitations of traditional technologies and provide new solutions for improving the safety of battery packs and extending their service life. However, to successfully apply these cutting-edge technologies to actual battery management systems, a series of technical challenges such as high data processing complexity and strict algorithm real-time requirements still need to be addressed. Summary of the invention

[0005] The main purpose of the present invention is to provide a fault diagnosis method and related equipment for a power battery pack, which solves the technical problem that the traditional method mainly relies on simple voltage and current monitoring and periodic maintenance inspections. This method is not only inefficient, but also often cannot detect and deal with potential internal problems in a timely manner, resulting in irreversible damage when a fault occurs.

[0006] To achieve the above object, the present invention provides a method for diagnosing a fault of a power battery pack, comprising the following steps: The electrical parameters of the power battery pack are collected by a preset data collection device to obtain the electrical collection parameters of the power battery pack; Performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack; Performing electrochemical impedance spectroscopy analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnosing whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack; If so, performing fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack; A fault location analysis is performed on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack.

[0007] Furthermore, the electrical parameters of the power battery pack are collected by a preset data collection device to obtain the electrical collection parameters of the power battery pack, including: Collecting electrical parameters of the power battery pack through a preset data acquisition device to obtain preliminary collected electrical parameters of the power battery pack; Performing discrete wavelet transform on the preliminary collected electrical parameters to obtain multi-resolution electrical parameter data of the power battery pack; wherein the preliminary collected electrical parameters include voltage parameters and current collection parameters; Performing singular value decomposition on the power battery pack based on the multi-resolution electrical parameter data to obtain electrical singular value data; The electrical singular value data are denoised and subjected to principal component analysis to obtain electrical acquisition parameters of the power battery pack.

[0008] Furthermore, the performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack includes: Based on the electrical acquisition parameters, the power battery pack is converted into a time-frequency domain to obtain time-frequency characteristic data of the power battery pack, and the time-frequency characteristic data is mapped into a complex domain to obtain a complex domain characteristic sequence of the power battery pack; Performing phase angle extraction on the complex domain characteristic sequence to obtain an instantaneous phase sequence of the power battery pack; and performing phase unwrapping processing on the instantaneous phase sequence to obtain a continuous phase sequence; Performing phase difference calculation on the continuous phase sequence to obtain a phase difference feature of the power battery pack, and performing multi-scale decomposition based on the phase difference feature to obtain a multi-scale phase component; Performing energy distribution calculation on the multi-scale phase component to obtain a phase energy spectrum of the power battery pack, and performing nonlinear feature extraction on the phase energy spectrum to obtain a nonlinear phase feature; Phase reconstruction is performed based on the nonlinear phase characteristics to obtain a phase characteristic spectrum of the power battery pack; wherein the phase characteristic spectrum is used to reflect the spectrum data of the overall phase change characteristics in the power battery pack.

[0009] Further, the performing electrochemical impedance spectrum analysis on the power battery pack based on the phase characteristic spectrum to obtain the impedance characteristic spectrum of the power battery pack includes: Performing a complex domain transformation on the power battery pack based on the phase characteristic spectrum to obtain a complex impedance sequence of the power battery pack, and performing a multi-scale discrete Fourier analysis on the complex impedance sequence to obtain a frequency domain impedance component; Performing a time-frequency joint decomposition on the frequency domain impedance component by a preset adaptive kernel function to obtain a time-frequency impedance matrix, and performing a bispectral analysis on the time-frequency impedance matrix to obtain a high-order impedance feature; Performing variational mode decomposition on the power battery pack based on the high-order impedance characteristics to obtain an intrinsic impedance function of the power battery pack, and performing Volterra series expansion on the intrinsic impedance function to obtain a nonlinear impedance kernel function of the power battery pack; The nonlinear impedance kernel function is eigen-decomposed by a preset covariance matrix to obtain an impedance eigenvector, and a spectrum is reconstructed based on the impedance eigenvector to obtain an impedance characteristic spectrum of the power battery pack.

[0010] Furthermore, the performing fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack includes: Based on the impedance characteristic spectrum, a spectral domain hierarchical analysis is performed on the power battery pack to obtain a multi-layer impedance characteristic set of the power battery pack, and based on the multi-layer impedance characteristic set, a high-dimensional feature transformation is performed on the power battery pack to obtain a transformation characteristic space of the power battery pack, and a fault correlation analysis is performed on the transformation characteristic space to obtain a fault correlation spectrum of the power battery pack; Decoupling the fault mode of the power battery pack based on the fault correlation spectrum to obtain a decoupling feature matrix, and performing fuzzy clustering analysis on the decoupling feature matrix to obtain a fault clustering spectrum of the power battery pack; The fault cluster spectrum is subjected to feature fusion through adaptive fusion to obtain a fused feature vector, and the fused feature vector is used as a fault feature vector of the power battery pack.

[0011] Further, the performing fault location analysis on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack includes: Performing a fault sensitivity analysis on the power battery pack based on the fault feature vector to obtain a fault sensitivity matrix of the power battery pack, and performing a singular value decomposition on the fault sensitivity matrix to obtain a fault feature space of the power battery pack; Performing envelope extraction on the fault feature space by Hilbert transform to obtain a fault envelope spectrum of the power battery pack, and performing bidirectional recursive analysis on the fault envelope spectrum to obtain a fault propagation link of the power battery pack; Performing fault delimitation mapping on the power battery pack based on the fault propagation link to obtain a fault boundary sequence, and topologically reconstructing the fault boundary sequence to obtain a fault topology network; Calculating the fault probability of the fault topology network by kernel density estimation to obtain a fault probability distribution, and performing maximum likelihood positioning on the fault probability distribution to obtain a set of fault source points of the power battery pack; Performing a fault influence domain analysis on the power battery pack based on the fault source point set to obtain a fault diffusion matrix of the power battery pack, and performing a spectral clustering analysis on the fault diffusion matrix to obtain a fault area spectrum of the power battery pack; A fault location analysis is performed on the power battery pack using the fault area spectrum to obtain a fault location result of the power battery pack.

[0012] Furthermore, the performing of a fault influence domain analysis on the power battery pack based on the fault source point set to obtain a fault diffusion matrix of the power battery pack includes: Based on the fault source point set, the power battery pack is grid-segmented to obtain a multi-layer fault grid map, and a topological connectivity analysis is performed on the multi-layer fault grid map to obtain a fault conduction path set; Performing fault diffusion dynamics modeling based on the fault conduction path set to obtain a fault diffusion dynamics equation, and performing fault evolution analysis on the fault diffusion dynamics equation to obtain a fault evolution characteristic diagram; Calculating the fault influence intensity of the power battery pack based on the fault evolution characteristic diagram to obtain a fault weight vector, and performing nonlinear coupling analysis on the fault weight vector to obtain a fault coupling network; A causal association analysis is performed on the fault coupling network to obtain a fault transmission chain, and the fault impact of the power battery pack is quantified based on the fault transmission chain to obtain a fault impact matrix, and a spectral clustering analysis is performed on the fault impact matrix to obtain a fault impact area set of the power battery pack, and a fault diffusion matrix of the power battery pack is obtained based on the fault impact area set.

[0013] The present invention also provides a fault diagnosis device for a power battery pack, comprising: The acquisition module is used to acquire electrical parameters of the power battery pack through a preset data acquisition device to obtain electrical acquisition parameters of the power battery pack; A first analysis module, configured to perform phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack; A second analysis module is used to perform electrochemical impedance spectroscopy analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnose whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack; A mapping module, configured to perform fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack, if any; The positioning module is used to perform fault positioning analysis on the power battery pack based on the fault feature vector to obtain a fault positioning result of the power battery pack.

[0014] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0015] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0016] A fault diagnosis method for a power battery pack provided by the present invention comprises the following steps: collecting electrical parameters of the power battery pack through a preset data acquisition device to obtain electrical acquisition parameters of the power battery pack; performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack; performing electrochemical impedance spectrum analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnosing whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack; if so, performing fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack; performing fault location analysis on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack, thereby solving the technical problem that the traditional method mainly relies on simple voltage and current monitoring and periodic maintenance inspections, which is not only inefficient, but also often fails to timely discover and handle potential internal problems, resulting in irreversible damage when the fault occurs, and realizing that once a fault is detected, fault feature mapping is performed based on the impedance characteristic spectrum, and further fault location analysis is performed using the fault feature vector, so that the fault position can be accurately locked. This is crucial for developing targeted maintenance plans, helping to reduce unnecessary disassembly and inspection work, saving time and costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a schematic diagram of the steps of a method for diagnosing a fault of a power battery pack in one embodiment of the present invention; Figure 2 is a structural block diagram of a fault diagnosis device for a power battery pack in one embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0018] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0020] like Figure 1 As shown, Figure 1 This is a schematic diagram of the steps of a method for diagnosing a fault of a power battery pack in one embodiment of the present invention; In one embodiment of the present invention, a method for diagnosing a fault of a power battery pack is provided, comprising the following steps: Step S1, collecting electrical parameters of the power battery pack through a preset data collection device to obtain electrical collection parameters of the power battery pack.

[0021] Specifically, the process of collecting electrical parameters of the power battery pack through a preset data acquisition device to obtain the electrical acquisition parameters of the power battery pack is the basic step of the entire fault diagnosis method, and its purpose is to obtain accurate and comprehensive power battery pack operating status information. First, the data acquisition device must be configured according to the preset standards and requirements, which includes selecting the appropriate sensor type, determining the sampling frequency, and setting the data transmission protocol. For example, in the application scenario of electric vehicles, in order to monitor the working condition of the power battery pack, a combination of various types of sensors such as voltage sensors, current sensors, and temperature sensors are usually used to monitor the key indicators of the battery pack in real time. In the specific implementation process, these sensors are installed at key positions of the power battery pack, such as the input and output ports of each battery module, to ensure that the most realistic electrical parameter changes can be captured. Once the vehicle is started or charging begins, the data acquisition device will automatically collect parameters such as voltage, current, and temperature according to the preset sampling frequency, and transmit these raw data to the central processing unit by wire or wireless means. It should be noted here that the sampling frequency needs to be adjusted according to the needs of the actual application. For high-precision requirements, a higher sampling frequency may be required to capture rapidly changing dynamic processes. The collected data will then be stored and preliminarily processed for use in subsequent analysis steps. For example, by comparing and analyzing the voltage and current data collected continuously over a period of time, it can be found whether there are abnormal fluctuations, thereby providing the necessary basic data for phase spectrum analysis. Suppose in a specific application example, the power battery pack of an electric bus has an unexplained problem of excessive power consumption during a certain driving process. Technicians can trace back the electrical acquisition parameters during this period of time and combine other diagnostic methods to gradually find out whether the root cause of the problem lies in the aging of certain cells inside the battery or poor contact of the connecting lines. It can be seen that accurate and comprehensive electrical parameter acquisition not only provides reliable data support for subsequent fault diagnosis, but also is an important guarantee for ensuring the safe and stable operation of the entire power battery system. In the whole process, the data acquisition device plays a vital role. It is not only a bridge connecting the physical world and the digital world, but also the cornerstone of realizing the intelligent battery management system. By continuously optimizing the data acquisition solution, the accuracy and efficiency of fault diagnosis can be further improved, contributing to the development of new energy vehicles.

[0022] Step S2: performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack.

[0023] Specifically, the phase spectrum analysis of the power battery pack is performed based on the electrical acquisition parameters to obtain the phase characteristic spectrum of the power battery pack. This process is achieved through complex mathematical and signal processing techniques, aiming to extract key features that can reflect the changes in the internal state of the battery from the acquired electrical parameters. First, the acquired electrical acquisition parameters, including voltage, current and other data, will be input into a specially designed algorithm for processing. These algorithms are usually based on fast Fourier transform (FFT) or other similar frequency domain analysis methods to convert time domain signals into frequency domain representations, thereby revealing periodic and non-periodic components hidden in the original data. Specifically, in the application scenario of the power battery pack of an electric vehicle, it is assumed that we have obtained the voltage and current data of the battery pack for a certain period of time through a preset data acquisition device. Next, in order to perform phase spectrum analysis, we will feed these time series data into the phase analysis algorithm. The algorithm first preprocesses the data, such as filtering and smoothing, to remove noise and unnecessary interference. Then, using techniques such as FFT or wavelet transform, these preprocessed time series data are converted into frequency domain representations. In this process, not only the frequency distribution can be obtained, but also the phase information corresponding to each frequency component, the so-called phase spectrum, can be obtained. Phase spectrum is a tool to describe the response characteristics of a system at different frequencies. It can reflect the changes in the internal structure of the battery pack and potential problems. For example, in a specific case, if the power battery pack of an electric bus exhibits abnormal charging and discharging behavior, technicians can identify whether there is an abnormal phase shift at a specific frequency by analyzing the phase spectrum of the battery pack during this period. This shift may be caused by the aging of certain cells inside the battery or poor connection. Through a detailed analysis of the phase spectrum, these problems can be discovered and provide a basis for further diagnosis. In addition, phase spectrum analysis can also help identify changes in the characteristics of the battery pack under different working conditions. For example, the phase spectrum under high load conditions and low load conditions may show different patterns, which helps to understand the performance of the battery pack under various operating conditions. In short, by performing phase spectrum analysis based on electrical acquisition parameters, not only can the working status information of the battery pack be deeply mined, but also the foundation can be laid for subsequent electrochemical impedance spectroscopy analysis, and finally a comprehensive assessment of the health status of the power battery pack and fault warning can be achieved. This method greatly improves the accuracy and reliability of fault detection, which is of great significance for ensuring the safe operation of electric vehicles. Throughout the entire process, accurate electrical acquisition parameters and effective phase spectrum analysis technology work together to enable even tiny changes in battery status to be captured and analyzed in a timely manner, providing maintenance personnel with a powerful tool.

[0024] Step S3, performing electrochemical impedance spectroscopy analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnosing whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack.

[0025] Specifically, the power battery pack is subjected to electrochemical impedance spectroscopy analysis based on the phase characteristic spectrum to obtain the impedance characteristic spectrum of the power battery pack, and the power battery pack is diagnosed to be faulty based on the impedance characteristic spectrum of the power battery pack. This process is achieved through a series of complex signal processing and data analysis techniques, aiming to further analyze the internal state changes of the battery. First, after obtaining the phase characteristic spectrum, these data will be used to construct an electrochemical impedance spectroscopy (EIS) model. EIS is a powerful tool that evaluates the electrochemical properties of a battery by applying a small amplitude sinusoidal voltage or current disturbance and measuring the corresponding response. In this process, the phase characteristic spectrum provides important frequency domain information, which helps to more accurately explain the behavior of the battery at different frequencies. Specifically, in the application scenario of the power battery pack of an electric vehicle, it is assumed that we have obtained the phase characteristic spectrum of the battery pack from the previous step. Next, using these phase characteristic spectra as input, we can apply the electrochemical impedance spectroscopy analysis method to further analyze the internal state of the battery pack. First, by performing frequency domain analysis on the phase characteristic spectrum, the main frequency components and their corresponding phase relationships are identified. Then, based on these frequency components, the impedance response of the battery pack at different frequencies is simulated and calculated. This usually involves complex mathematical modeling and simulation processes, in which equivalent circuit models (such as Randles circuit models) are used to describe the electrochemical behavior of various components inside the battery (such as electrolytes, electrode materials, etc.). For example, in a specific case, if the power battery pack of an electric bus shows an abnormal decrease in charge and discharge efficiency, technicians can determine the problem by analyzing the phase characteristic spectrum and impedance characteristic spectrum of the battery pack during this period. By comparing the impedance characteristic spectrum in the normal state with the impedance characteristic spectrum in the current state, it can be found that the impedance value in certain specific frequency ranges has increased or decreased significantly, which may be caused by changes in the internal structure of the battery (such as electrolyte decomposition, active material shedding, etc.). This change causes the battery to exhibit different impedance characteristics in the high-frequency and low-frequency regions, providing important clues for fault diagnosis. In addition, when fault diagnosis is performed based on the impedance characteristic spectrum, historical data and other sensor information can also be combined for comprehensive judgment. For example, if the temperature of the battery pack increases within a certain period of time and the impedance characteristic spectrum shows abnormal fluctuations, it may indicate that the battery is at risk of overheating or other potential faults. Through in-depth analysis of these data, not only can the specific fault type be identified, but also the specific location of the fault can be located, providing guidance for subsequent maintenance and repair work. In summary, by performing electrochemical impedance spectroscopy analysis based on the phase characteristic spectrum and diagnosing whether the power battery pack has a fault based on the impedance characteristic spectrum, this method not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles.During the entire process, the phase characteristic spectrum and the impedance characteristic spectrum work together to enable even tiny changes in the battery status to be captured and analyzed in a timely manner, providing a powerful tool for maintenance personnel.

[0026] Step S4: If yes, perform fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack.

[0027] Specifically, if there is a fault, the power battery pack is subjected to fault feature mapping based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack. This process aims to convert complex electrochemical impedance data into a feature representation that is easy to understand and process. After obtaining the impedance characteristic spectrum of the battery pack, the next key step is to identify and extract specific patterns or features that can characterize the fault type through fault feature mapping technology. Specifically, it is necessary to first establish a fault feature library, which contains various known fault types and their corresponding impedance characteristic spectrum templates. These templates are usually constructed based on a large amount of experimental data and historical fault records, and are verified to ensure their accuracy and reliability. In the application scenario of the power battery pack of electric vehicles, it is assumed that we have determined that the battery pack has some abnormality through electrochemical impedance spectroscopy analysis. In order to further clarify the specific type and location of the fault, the technician will compare the impedance characteristic spectrum of the current battery pack with the template in the fault feature library. For example, when the power battery pack of an electric bus shows an abnormal decrease in charging and discharging efficiency, the technician can input the impedance characteristic spectrum of the battery pack during this period into the fault feature mapping algorithm. This algorithm calculates the similarity or distance measure between the current impedance characteristic spectrum and each fault template in the library, such as Euclidean distance, cosine similarity, etc. Through this comparison and calculation, the algorithm can identify the fault template closest to the current impedance characteristic spectrum and extract the corresponding fault feature vector from it. The fault feature vector is a multidimensional array, in which each dimension represents a specific fault feature parameter, such as a change in resistance value, a change in capacitance value, a phase shift, etc. These feature parameters together constitute a complete fault descriptor that can fully reflect the specific problems inside the battery pack. For example, in the above case, if it is found that the impedance of the high-frequency area of ​​the battery pack increases significantly, while the impedance of the low-frequency area is relatively stable, it may indicate that there is a problem of electrolyte decomposition inside the battery pack. At this time, the fault feature vector will contain multiple feature parameters related to electrolyte decomposition, such as high-frequency impedance value, temperature change rate, etc. In addition, in order to improve the accuracy and robustness of fault feature mapping, machine learning and artificial intelligence technologies can also be introduced. For example, models such as support vector machines (SVM) and neural networks are used to classify and regress fault features. These models can be trained based on a large amount of historical data, so that they can provide more accurate diagnostic results when faced with new fault conditions. In this way, not only can the specific fault type be identified, but also the development trend of the fault can be predicted, providing a scientific basis for subsequent maintenance and repair work. In short, the process of fault feature mapping based on the impedance characteristic spectrum and obtaining the fault feature vector not only improves the accuracy and reliability of fault detection, but also provides strong support for optimizing the battery management system.Throughout the process, the impedance characteristic spectrum and fault characteristic vector work together to capture and analyze even tiny changes in the battery state in a timely manner, providing maintenance personnel with a powerful tool. This step is crucial to ensuring the safety and efficient operation of electric vehicles, especially when dealing with complex and changing actual application environments.

[0028] Step S5, performing fault location analysis on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack.

[0029] Specifically, the power battery pack is subjected to fault location analysis based on the fault feature vector to obtain the fault location result of the power battery pack. This process is achieved by comprehensively applying a variety of data analysis and pattern recognition technologies, aiming to accurately determine the specific fault location inside the battery pack. After obtaining the fault feature vector, the next key task is to map these feature parameters to specific physical locations or components, so as to provide clear guidance for subsequent repairs and maintenance. First, the fault feature vector contains feature parameters of multiple dimensions, each of which represents a specific electrochemical or physical property change. For example, in the application scenario of the power battery pack of an electric bus, it is assumed that we have obtained a fault feature vector through the previous steps, which contains multiple feature parameters such as high-frequency impedance value, temperature change rate, and phase shift. In order to realize fault location analysis, it is usually necessary to construct a detailed battery pack structure model and combine the relationship between the various components in the model (such as single battery cells, connectors, thermal management systems, etc.) and the fault feature vector. Specifically, technicians use computer-aided engineering (CAE) tools or specially developed fault diagnosis software to simulate the working state of the battery pack and adjust the model parameters according to the parameters in the fault feature vector. For example, if the fault feature vector shows a significant increase in impedance in the high-frequency region, while the low-frequency region is relatively stable, it may indicate that the problem occurs in certain specific cells or connection points of the battery pack. At this point, technicians can simulate these changes by adjusting the corresponding parameters in the model and observe their impact on the overall system. In practical applications, such as the abnormal decrease in charging and discharging efficiency of the power battery pack of an electric bus, technicians can use the fault feature vector for detailed analysis. First, they will enter the fault feature vector in the fault diagnosis software and run the simulation program. During the simulation process, the software will automatically adjust the various parameters in the battery pack model according to the data in the fault feature vector and monitor the output results in real time. In this way, technicians can intuitively see which parts have abnormal changes in impedance or temperature, thereby gradually narrowing the scope of the fault. For example, if the simulation results show that the impedance of a specific battery module is significantly higher than that of other modules, it can be preliminarily determined that the module has a potential fault. In addition, in order to further improve the accuracy of fault location, other sensor data and historical records can also be combined for comprehensive analysis. For example, by comparing the historical operating data of the battery pack, technicians can find that the performance of a module has shown a significant downward trend in a certain period of time, which provides important clues for further confirming the fault location. At the same time, using infrared imagers or other detection equipment to conduct on-site inspections of suspected fault areas can more accurately verify the fault diagnosis results. Ultimately, after a series of data analysis and on-site verification, technicians can obtain an accurate fault location result, pointing out the specific faulty component or location.For example, in the above case, if it is determined that the electrolyte decomposition of a battery module causes the increase in high-frequency impedance, the module can be replaced in a targeted manner to restore the overall performance of the battery pack. In short, the process of fault location analysis based on fault feature vectors not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Throughout the process, the fault feature vector and the detailed battery pack structure model work together to enable even minor changes in battery status to be captured and analyzed in a timely manner, providing a powerful tool for maintenance personnel.

[0030] In a specific embodiment, the electrical parameters of the power battery pack are collected by a preset data collection device to obtain the electrical collection parameters of the power battery pack, including: Collecting electrical parameters of the power battery pack through a preset data acquisition device to obtain preliminary collected electrical parameters of the power battery pack; Performing discrete wavelet transform on the preliminary collected electrical parameters to obtain multi-resolution electrical parameter data of the power battery pack; wherein the preliminary collected electrical parameters include voltage parameters and current collection parameters; Performing singular value decomposition on the power battery pack based on the multi-resolution electrical parameter data to obtain electrical singular value data; The electrical singular value data are denoised and subjected to principal component analysis to obtain electrical acquisition parameters of the power battery pack.

[0031] Specifically, the process of collecting electrical parameters of the power battery pack through a preset data acquisition device to obtain the electrical acquisition parameters of the power battery pack is the basic step of the entire fault diagnosis method, and its purpose is to obtain accurate and comprehensive power battery pack operation status information. First, the data acquisition device is configured according to the preset standards and requirements, which includes selecting the appropriate sensor type, determining the sampling frequency, and setting the data transmission protocol. For example, in the application scenario of electric vehicles, in order to monitor the working condition of the power battery pack, a combination of various types of sensors such as voltage sensors, current sensors, and temperature sensors are usually used to monitor the key indicators of the battery pack in real time. In the specific implementation process, these sensors are installed at key positions of the power battery pack, such as the input and output ports of each battery module, to ensure that the most realistic electrical parameter changes can be captured. Once the vehicle is started or charging begins, the data acquisition device will automatically collect parameters such as voltage, current, and temperature according to the preset sampling frequency, and transmit these raw data to the central processing unit by wire or wireless means. It should be noted here that the sampling frequency needs to be adjusted according to the needs of the actual application. For high-precision requirements, a higher sampling frequency may be required to capture rapidly changing dynamic processes. The electrical parameters collected initially are called preliminary collected electrical parameters. These data contain rich information, but are also accompanied by noise and unnecessary interference. In order to improve the quality of the data, these preliminary collected electrical parameters will be subjected to discrete wavelet transform (DWT) to obtain multi-resolution electrical parameter data. Discrete wavelet transform is an effective signal processing technology that can analyze signals at different scales, retain important feature information, and remove noise at the same time. In this process, the voltage parameters and current acquisition parameters are decomposed into different frequency components to form a multi-resolution representation. This multi-resolution representation not only helps to identify the detailed features in the signal, but also effectively separates the noise components, providing a clearer data basis for subsequent analysis. Based on the multi-resolution electrical parameter data, singular value decomposition (SVD) is further performed on it. Singular value decomposition is a matrix decomposition method that can decompose complex multidimensional data into several main components that represent the main features of the data. Through singular value decomposition, the most important feature information, namely electrical singular value data, can be extracted from the multi-resolution electrical parameter data. This process is similar to extracting the most representative features from an image, so that subsequent analysis focuses more on the key parts and ignores those unimportant details. However, even after singular value decomposition, there may still be some noise and redundant information in the data. Therefore, the electrical singular value data needs to be denoised and principal component analysis (PCA) performed next. Denoising is intended to further reduce noise interference in the data and ensure the accuracy of subsequent analysis results.The principal component analysis is used for dimensionality reduction and feature extraction. By calculating the covariance matrix and finding its eigenvector, high-dimensional data can be projected into a low-dimensional space to retain the most important feature information. Finally, through denoising and principal component analysis, the final electrical acquisition parameters of the power battery pack are obtained. These parameters have a high signal-to-noise ratio and representativeness, and can truly reflect the actual working state of the battery pack. For example, in a specific case, it is assumed that the power battery pack of an electric bus has an unexplained problem of excessive power consumption during a certain driving process. The technicians can conduct a detailed analysis through the above steps: first, the voltage and current parameters of the battery pack are collected using a preset data acquisition device; then, these preliminary collected electrical parameters are subjected to discrete wavelet transform to obtain multi-resolution electrical parameter data; then, singular value decomposition is used to extract electrical singular value data; then, the electrical singular value data is denoised and principal component analysis is performed on the electrical singular value data to obtain the final electrical acquisition parameters. Through in-depth analysis of these electrical acquisition parameters, technicians can find out whether there are abnormal fluctuations or phase shifts at specific frequencies, thereby providing reliable data support for subsequent phase spectrum analysis and electrochemical impedance spectrum analysis. In addition, this method can be applied not only to fault diagnosis, but also to battery performance evaluation and optimization management. For example, by regularly collecting the electrical parameters of the battery pack and performing the above-mentioned processing, a long-term monitoring system for the battery health status can be established to detect potential problems in a timely manner and take preventive measures. In short, the electrical parameters of the power battery pack are collected by a preset data acquisition device, and after a series of complex signal processing and data analysis steps, high-quality electrical acquisition parameters are finally obtained. This process not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Throughout the process, accurate electrical acquisition parameters and effective data processing technology work together, so that even small changes in battery status can be captured and analyzed in a timely manner, providing a powerful tool for maintenance personnel. This method greatly enhances the adaptability and response speed of the system, and contributes an important force to the development of new energy vehicles.

[0032] In a specific embodiment, performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack includes: Based on the electrical acquisition parameters, the power battery pack is converted into a time-frequency domain to obtain time-frequency characteristic data of the power battery pack, and the time-frequency characteristic data is mapped into a complex domain to obtain a complex domain characteristic sequence of the power battery pack; Performing phase angle extraction on the complex domain characteristic sequence to obtain an instantaneous phase sequence of the power battery pack; and performing phase unwrapping processing on the instantaneous phase sequence to obtain a continuous phase sequence; Performing phase difference calculation on the continuous phase sequence to obtain a phase difference feature of the power battery pack, and performing multi-scale decomposition based on the phase difference feature to obtain a multi-scale phase component; Performing energy distribution calculation on the multi-scale phase component to obtain a phase energy spectrum of the power battery pack, and performing nonlinear feature extraction on the phase energy spectrum to obtain a nonlinear phase feature; Phase reconstruction is performed based on the nonlinear phase characteristics to obtain a phase characteristic spectrum of the power battery pack; wherein the phase characteristic spectrum is used to reflect the spectrum data of the overall phase change characteristics in the power battery pack.

[0033] Specifically, the phase spectrum analysis of the power battery pack is performed based on the electrical acquisition parameters to obtain the phase characteristic spectrum of the power battery pack. This process is achieved through a series of complex signal processing and data analysis technologies, aiming to extract key features that can reflect the changes in the internal state of the battery from the acquired electrical parameters. First, the acquired electrical acquisition parameters, including voltage, current and other data, will be input into a specially designed algorithm for processing. These algorithms are usually based on time-frequency domain conversion methods to convert time series signals into time-frequency representations that are easier to analyze. Specifically, in the application scenario of the power battery pack of an electric vehicle, it is assumed that we have obtained the voltage and current data of the battery pack for a certain period of time through a preset data acquisition device. In order to perform phase spectrum analysis, it is first necessary to perform time-frequency domain conversion on these electrical acquisition parameters to obtain the time-frequency characteristic data of the power battery pack. For example, by using methods such as short-time Fourier transform (STFT) or wavelet transform, the original time series data can be decomposed into different frequency components and their changes can be displayed on the time axis. The purpose of this is to reveal the frequency characteristics of the battery pack at different time points and their changing trends over time. In this process, suppose we select a 5-minute time window, collect data once a second, and collect a total of 300 data points. Through time-frequency domain conversion, we can generate a two-dimensional matrix containing 300 time points and multiple frequency components, each element represents the energy value at a certain frequency at a certain moment. Next, the obtained time-frequency feature data is mapped to the complex domain, that is, the real and imaginary parts are separated and mapped to the complex plane to form a complex domain feature sequence. This step helps to retain the phase information in the original signal, making the subsequent phase analysis more accurate. For example, suppose we obtain a 128×300 time-frequency feature matrix in the above 5-minute time window. By performing complex domain mapping on each frequency component, we can obtain a 128×300 complex matrix. Each complex number represents the instantaneous phase information at a certain frequency at a certain moment. Then, the complex domain feature sequence is subjected to phase angle extraction to obtain an instantaneous phase sequence. Phase angle extraction is the process of separating phase information from the complex domain feature sequence, which can reflect the phase changes of the battery pack at different frequencies. In order to ensure the continuity of phase information, it is also necessary to perform phase unwrapping on the instantaneous phase sequence to obtain a continuous phase sequence. For example, in the above case, suppose we obtain a 128×300 instantaneous phase sequence, in which some phase values ​​may jump due to calculation errors. By unwrapping these phase values, we can eliminate such jumps and ensure the continuity and accuracy of phase changes. Furthermore, the continuous phase sequence is subjected to phase differential calculation to obtain phase differential features. The phase differential feature reflects the difference in phase changes between adjacent time points and can capture subtle changes in the internal state of the battery pack.For example, in the above case, suppose we perform phase difference calculation on the 128×300 continuous phase sequence and obtain a new 128×299 phase difference feature matrix. Each element represents the phase change rate between two adjacent time points at a certain frequency. Subsequently, the phase difference feature is multi-scale decomposed to obtain multi-scale phase components. Multi-scale decomposition is a method of decomposing a signal into components of different scales, which can reveal the behavioral characteristics of the battery pack at different time scales. For example, suppose we perform multi-scale decomposition on the above 128×299 phase difference feature matrix and obtain three-scale phase components: high-frequency component (corresponding to fast changes), medium-frequency component (corresponding to medium-speed changes) and low-frequency component (corresponding to slow changes). Each scale component reflects the dynamic characteristics of the battery pack at different time scales. Then, the energy distribution of the multi-scale phase components is calculated to obtain the phase energy spectrum. The energy distribution calculation is to quantify the energy contribution of each frequency component, so as to better understand the overall behavior of the battery pack. For example, in the above case, suppose we perform energy distribution calculation on the three-scale phase components respectively and obtain three energy spectra. Each energy spectrum shows the energy distribution at the corresponding scale, which can help identify the frequency ranges in which energy changes are most significant. Finally, the phase energy spectrum is subjected to nonlinear feature extraction to obtain nonlinear phase features, and phase reconstruction is performed based on these features to obtain the final phase characteristic spectrum of the power battery pack. Nonlinear feature extraction is to capture the complex and nonlinear dynamic change patterns within the battery pack, thereby providing a more comprehensive state description. For example, in the above case, suppose we extracted several nonlinear features, such as kurtosis, skewness, etc., from the three energy spectra, and performed phase reconstruction based on these features. The reconstructed phase characteristic spectrum not only contains the overall phase change characteristics of the battery pack, but also reflects the interaction relationship between the various frequency components. In a specific example, suppose that the power battery pack of an electric bus has an abnormal decrease in charging and discharging efficiency during a certain driving process. Technicians can conduct detailed analysis through the above steps: first, use the preset data acquisition device to collect the voltage and current parameters of the battery pack; then, perform time-frequency domain conversion on these preliminary collected electrical parameters to obtain time-frequency feature data; then, obtain a continuous phase sequence through complex domain mapping, phase angle extraction and phase unwrapping processing; then, perform phase difference calculation on the continuous phase sequence to obtain phase difference features; then perform multi-scale decomposition on the phase difference features to obtain multi-scale phase components; then, perform energy distribution calculation on the multi-scale phase components to obtain a phase energy spectrum; finally, perform nonlinear feature extraction and phase reconstruction on the phase energy spectrum to obtain the final phase feature spectrum. Through in-depth analysis of these phase feature spectra, technicians can find out whether there are abnormal phase fluctuations at specific frequencies, thereby providing reliable data support for subsequent electrochemical impedance spectroscopy analysis.For example, in the above case, suppose we find that the phase value in a certain frequency range increases significantly, while other frequency ranges are relatively stable, which may indicate problems such as electrolyte decomposition or electrode material aging inside the battery pack. This change will cause the battery to exhibit different phase characteristics in the high-frequency and low-frequency regions, providing important clues for fault diagnosis. In addition, this method can be applied not only to fault diagnosis, but also to battery performance evaluation and optimization management. For example, by regularly collecting the electrical parameters of the battery pack and performing the above processing, a long-term monitoring system for the battery health status can be established to detect potential problems in time and take preventive measures. In summary, the process of phase spectrum analysis based on electrical acquisition parameters not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Throughout the process, the precise electrical acquisition parameters and effective phase spectrum analysis technology work together to capture and analyze even minor changes in battery status in a timely manner, providing a powerful tool for maintenance personnel. For example, suppose we monitor the power battery pack of an electric bus for a month, collecting data once a day. After processing through the above steps, we find that the phase characteristic spectrum on the 15th day shows that the phase values ​​in certain frequency ranges are significantly higher than those in other time periods. Further analysis shows that this may be caused by the aging of certain cells inside the battery pack. By replacing these aging cells in time, the overall performance of the battery pack can be effectively restored, its service life can be extended, and the overall performance and safety of the vehicle can be improved. This method not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Specifically, in the above case, assume that we collected voltage and current data within 5 minutes, once per second, and a total of 300 data points were collected. Through time-frequency domain conversion, we generated a 128×300 time-frequency feature matrix. Then, by mapping the matrix to the complex domain, a 128×300 complex matrix was obtained. Then, the complex matrix was phase-extracted to obtain a 128×300 instantaneous phase sequence. By performing phase unwrapping on the instantaneous phase sequence, a 128×300 continuous phase sequence was obtained. Next, the continuous phase sequence was phase-differentially calculated to obtain a 128×299 phase-differential feature matrix. Then, the matrix was decomposed at multiple scales to obtain phase components at three scales: high frequency (1kHz-10kHz), medium frequency (100Hz-1kHz) and low frequency (10Hz-100Hz). Each scale component reflects the dynamic characteristics of the battery pack at different time scales. Then, the energy distribution of the phase components at these three scales was calculated to obtain three energy spectra. Each energy spectrum shows the energy distribution at the corresponding scale, helping to identify which frequency ranges have the most significant energy changes.Finally, the nonlinear feature extraction of the energy spectrum was performed to obtain several nonlinear phase features, and phase reconstruction was performed based on these features to obtain the final phase characteristic spectrum. This phase characteristic spectrum not only contains the overall phase change characteristics of the battery pack, but also reflects the interaction relationship between each frequency component, providing a reliable basis for subsequent fault diagnosis. In this way, even the slightest battery state change can be captured and analyzed in time, providing a powerful tool for maintenance personnel. This method greatly enhances the adaptability and response speed of the system, and contributes significantly to the development of new energy vehicles.

[0034] In a specific embodiment, the electrochemical impedance spectrum analysis of the power battery pack is performed based on the phase characteristic spectrum to obtain the impedance characteristic spectrum of the power battery pack, including: Performing a complex domain transformation on the power battery pack based on the phase characteristic spectrum to obtain a complex impedance sequence of the power battery pack, and performing a multi-scale discrete Fourier analysis on the complex impedance sequence to obtain a frequency domain impedance component; Performing a time-frequency joint decomposition on the frequency domain impedance component by a preset adaptive kernel function to obtain a time-frequency impedance matrix, and performing a bispectral analysis on the time-frequency impedance matrix to obtain a high-order impedance feature; Performing variational mode decomposition on the power battery pack based on the high-order impedance characteristics to obtain an intrinsic impedance function of the power battery pack, and performing Volterra series expansion on the intrinsic impedance function to obtain a nonlinear impedance kernel function of the power battery pack; The nonlinear impedance kernel function is eigen-decomposed by a preset covariance matrix to obtain an impedance eigenvector, and a spectrum is reconstructed based on the impedance eigenvector to obtain an impedance characteristic spectrum of the power battery pack.

[0035] Specifically, the electrochemical impedance spectrum of the power battery pack is analyzed based on the phase characteristic spectrum to obtain the impedance characteristic spectrum of the power battery pack. This process is achieved through a series of complex signal processing and data analysis technologies, aiming to extract key features that can reflect the changes in the internal state of the battery from the phase characteristic spectrum. First, the obtained phase characteristic spectrum will be input into a specially designed algorithm for processing. These algorithms are usually based on complex domain transformation, multi-scale discrete Fourier analysis, and high-order impedance feature extraction methods to reveal the behavioral characteristics of the battery pack at different frequencies. Specifically, in the application scenario of the power battery pack of an electric vehicle, it is assumed that we have obtained the phase characteristic spectrum of the battery pack for a period of time through the previous steps. In order to perform electrochemical impedance spectroscopy analysis, it is first necessary to perform a complex domain transformation on the phase characteristic spectrum, that is, to separate the real part and the imaginary part and map them to the complex plane to form a complex impedance sequence. For example, in a specific case, it is assumed that we collected phase characteristic spectrum data within 5 minutes, collected once per second, and collected a total of 300 data points. By performing a complex domain transformation on these data points, we can generate a complex impedance sequence containing 300 time points, each complex number representing the instantaneous impedance information at a certain moment. Next, a multiscale discrete Fourier analysis (MDFT) is performed on the complex impedance sequence to obtain the frequency domain impedance components. Multiscale discrete Fourier analysis is a method that decomposes a signal into multiple frequency components, which can reveal the dynamic characteristics of the battery pack at different frequencies. For example, in the above case, suppose we performed a multiscale discrete Fourier analysis on 300 complex impedance data points and obtained frequency domain impedance components of 128 frequency components. Each frequency component represents the impedance characteristics of the battery pack at a specific frequency, thus providing us with more detailed frequency response information. Subsequently, the frequency domain impedance components are decomposed in time and frequency by a preset adaptive kernel function to obtain a time-frequency impedance matrix, and a bispectral analysis is performed on the time-frequency impedance matrix to obtain high-order impedance features. Time-frequency joint decomposition is a method that combines time domain and frequency domain information to capture nonlinear interactions in the signal. For example, in the above case, suppose we used an adaptive kernel function to perform a time-frequency joint decomposition on 128 frequency domain impedance components to obtain a 128×128 time-frequency impedance matrix. By performing bispectral analysis on the matrix, we can identify the complex and nonlinear dynamic modes inside the battery pack, thereby extracting high-order impedance features. Furthermore, based on the high-order impedance features, variational mode decomposition (VMD) is performed on the power battery pack to obtain the intrinsic impedance function, and the intrinsic impedance function is expanded by Volterra series to obtain the nonlinear impedance kernel function. Variational mode decomposition is a method of decomposing a complex signal into several eigenmodes, which can separate different components in the signal. For example, in the above case, suppose we performed variational mode decomposition on the high-order impedance features and obtained three intrinsic impedance functions, corresponding to the dynamic characteristics of fast changes, medium changes, and slow changes, respectively.Then, by performing a Volterra series expansion on each intrinsic impedance function, we can obtain the corresponding nonlinear impedance kernel function, which describes the nonlinear behavior of the battery pack at different time scales. Next, the nonlinear impedance kernel function is eigendecomposed by a preset covariance matrix to obtain an impedance eigenvector, and the spectrum is reconstructed based on the impedance eigenvector to obtain the impedance characteristic spectrum of the final power battery pack. Covariance matrix eigendecomposition is a commonly used dimensionality reduction and feature extraction method that can retain the most important feature information. For example, in the above case, suppose we perform covariance matrix eigendecomposition on three nonlinear impedance kernel functions and obtain an impedance eigenvector containing 128 elements. Each element represents an important eigenvalue at a certain frequency, and these eigenvalues ​​together constitute the final impedance characteristic spectrum. In a specific example, suppose that the power battery pack of an electric bus has an abnormal drop in charging and discharging efficiency during a certain driving process. The technicians can conduct a detailed analysis through the above steps: first, use the phase characteristic spectrum obtained in the previous step; then, transform the phase characteristic spectrum into the complex domain to obtain a complex impedance sequence; then, perform multi-scale discrete Fourier analysis on the complex impedance sequence to obtain the frequency domain impedance component; then, perform a time-frequency joint decomposition of the frequency domain impedance component through a preset adaptive kernel function to obtain a time-frequency impedance matrix, and perform a bispectral analysis on it to obtain a high-order impedance feature; then perform a variational modal decomposition on the high-order impedance feature to obtain an intrinsic impedance function, and perform a Volterra series expansion on it to obtain a nonlinear impedance kernel function; finally, perform a characteristic decomposition of the nonlinear impedance kernel function through a preset covariance matrix to obtain an impedance characteristic vector, and reconstruct the spectrum based on the vector to obtain the final impedance characteristic spectrum. Through in-depth analysis of these impedance characteristic spectra, technicians can find out whether there are abnormal impedance fluctuations at a specific frequency, thereby providing reliable data support for subsequent fault diagnosis. For example, in the above case, suppose we find that the impedance value in a certain frequency range has increased significantly, while other frequency ranges are relatively stable, which may indicate that there are problems such as electrolyte decomposition or electrode material aging inside the battery pack. This change causes the battery to exhibit different impedance characteristics in the high-frequency and low-frequency regions, thus providing important clues for fault diagnosis. In addition, this method can be applied not only to fault diagnosis, but also to battery performance evaluation and optimization management. For example, by regularly collecting the electrical parameters of the battery pack and performing the above-mentioned processing, a long-term monitoring system for the battery health status can be established to detect potential problems in a timely manner and take preventive measures. In short, the process of electrochemical impedance spectroscopy analysis based on phase characteristic spectrum not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles.Throughout the process, the precise phase characteristic spectrum and effective electrochemical impedance spectroscopy analysis technology work together to capture and analyze even tiny changes in the battery state in a timely manner, providing a powerful tool for maintenance personnel. This approach greatly enhances the adaptability and response speed of the system, and contributes significantly to the development of new energy vehicles. For example, suppose we monitor the power battery pack of an electric bus for a month, collecting data once a day. After processing through the above steps, we found that the impedance characteristic spectrum on the 15th day showed that the impedance values ​​in certain frequency ranges were significantly higher than those in other time periods. Further analysis showed that this may be due to the aging of certain cells inside the battery pack. By replacing these aging cells in a timely manner, the overall performance of the battery pack can be effectively restored, its service life can be extended, and the overall performance and safety of the vehicle can be improved. This approach not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles.

[0036] In a specific embodiment, performing fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack includes: Based on the impedance characteristic spectrum, a spectral domain hierarchical analysis is performed on the power battery pack to obtain a multi-layer impedance characteristic set of the power battery pack, and based on the multi-layer impedance characteristic set, a high-dimensional feature transformation is performed on the power battery pack to obtain a transformation characteristic space of the power battery pack, and a fault correlation analysis is performed on the transformation characteristic space to obtain a fault correlation spectrum of the power battery pack; Decoupling the fault mode of the power battery pack based on the fault correlation spectrum to obtain a decoupling feature matrix, and performing fuzzy clustering analysis on the decoupling feature matrix to obtain a fault clustering spectrum of the power battery pack; The fault cluster spectrum is subjected to feature fusion through adaptive fusion to obtain a fused feature vector, and the fused feature vector is used as a fault feature vector of the power battery pack.

[0037] Specifically, the process of mapping the fault characteristics of the power battery pack based on the impedance characteristic spectrum of the battery pack to obtain the fault characteristic vector of the power battery pack is realized through a series of complex signal processing and data analysis technologies, aiming to extract key features that can reflect the changes in the internal state of the battery from the impedance characteristic spectrum and convert it into a feature representation form that is easy to understand and process. First, the obtained impedance characteristic spectrum will be input into a specially designed algorithm for processing. These algorithms are usually based on spectral domain hierarchical analysis, high-dimensional feature transformation, fault correlation analysis, and fuzzy clustering analysis to reveal the specific fault mode inside the battery pack. Specifically, in the application scenario of the power battery pack of an electric vehicle, it is assumed that we have obtained the impedance characteristic spectrum of the battery pack for a period of time through the previous steps. In order to perform fault characteristic mapping, it is first necessary to perform spectral domain hierarchical analysis on the impedance characteristic spectrum, that is, to decompose it into multiple levels of impedance feature sets. For example, in a specific case, it is assumed that we collected impedance characteristic spectrum data within 5 minutes, collected once per second, and collected a total of 300 data points. By performing spectral domain hierarchical analysis on these data points, we can generate multiple levels of impedance feature sets, each level representing the impedance characteristics under different frequency ranges or time scales. Suppose we have obtained three levels of impedance feature sets, corresponding to the high frequency (1kHz-10kHz), medium frequency (100Hz-1kHz) and low frequency (10Hz-100Hz) ranges. Next, based on the multi-layer impedance feature set, a high-dimensional feature transformation is performed on the power battery pack to obtain a transformed feature space. High-dimensional feature transformation is a method of mapping raw data to a higher-dimensional space, which can reveal more hidden feature information. For example, in the above case, suppose we performed principal component analysis (PCA) or linear discriminant analysis (LDA) on the three-level impedance feature set to obtain a transformed feature space containing 128 feature dimensions. Each feature dimension represents a specific change in electrochemical or physical properties, and these features together constitute a complete feature descriptor. Subsequently, a fault correlation analysis is performed on the transformed feature space to obtain a fault correlation spectrum. Fault correlation analysis aims to identify which features are significantly associated with known fault modes. For example, in the above case, suppose we performed a fault correlation analysis on 128 feature dimensions and found that 10 of them were highly correlated with common battery aging phenomena (such as electrolyte decomposition, electrode material shedding, etc.). By further analyzing these feature dimensions, we can generate a fault correlation spectrum, which shows the correlation strength between each feature dimension and the fault mode. Then, based on the fault correlation spectrum, the fault mode of the power battery pack is decoupled to obtain a decoupled feature matrix. Fault mode decoupling is a method of separating complex fault features into multiple independent modes, which can more clearly identify different fault types.For example, in the above case, suppose we perform independent component analysis (ICA) on the fault correlation spectrum and obtain a set of 5 decoupled feature matrices. Each decoupled feature matrix represents a specific fault mode, such as electrolyte decomposition, electrode material aging, etc. Then, fuzzy clustering analysis is performed on the decoupled feature matrix to obtain a fault cluster spectrum. Fuzzy clustering analysis is a method that classifies similar features into one category, which can reveal the intrinsic connection between fault modes. For example, in the above case, suppose we perform fuzzy C-means clustering (FCM) on 5 decoupled feature matrices and obtain a spectrum containing 3 fault clusters. Each fault cluster represents a group of fault modes with similar characteristics, which provides us with more detailed fault classification information. Finally, the fault cluster spectrum is feature fused through adaptive fusion to obtain a fused feature vector, which is used as the fault feature vector of the power battery pack. Adaptive fusion is a method that combines multiple feature information to retain the most important feature information and remove redundant information. For example, in the above case, suppose we perform adaptive fusion on 3 fault cluster spectra to obtain a fused feature vector containing 64 elements. Each element represents an important eigenvalue of a certain fault mode, and these eigenvalues ​​together constitute the final fault feature vector. In a specific example, suppose that the power battery pack of an electric bus has an abnormal decrease in charging and discharging efficiency during a certain driving process. Technicians can perform detailed analysis through the above steps: first, use the impedance characteristic spectrum obtained in the previous step; then, perform spectrum domain layered analysis on the impedance characteristic spectrum to obtain a multi-layer impedance feature set; then, perform high-dimensional feature transformation based on the multi-layer impedance feature set to obtain a transformed feature space; then, perform fault correlation analysis on the transformed feature space to obtain a fault correlation spectrum; then decouple the fault correlation spectrum from the fault mode to obtain a decoupled feature matrix; then, perform fuzzy clustering analysis on the decoupled feature matrix to obtain a fault clustering spectrum; finally, perform feature fusion on the fault clustering spectrum through adaptive fusion to obtain the final fault feature vector. Through in-depth analysis of these fault feature vectors, technicians can find out whether there is a specific fault mode and its corresponding fault location, thereby providing clear guidance for subsequent maintenance and repair work. For example, in the above case, suppose we find that a specific fault cluster is closely related to the aging of certain units inside the battery pack. By further analyzing the characteristic vectors in the fault cluster, the specific location of the faulty unit can be determined, and these aging units can be replaced in time to restore the overall performance of the battery pack. In addition, this method can not only be applied to fault diagnosis, but also to battery performance evaluation and optimization management. For example, by regularly collecting the electrical parameters of the battery pack and performing the above processing, a long-term monitoring system for the battery health status can be established to detect potential problems in time and take preventive measures.In summary, the process of fault feature mapping based on impedance characteristic spectrum not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Throughout the process, the precise impedance characteristic spectrum and effective fault feature mapping technology work together to capture and analyze even tiny changes in battery status in a timely manner, providing a powerful tool for maintenance personnel. This method greatly enhances the adaptability and response speed of the system and contributes significantly to the development of new energy vehicles. For example, suppose we monitor the power battery pack of an electric bus for a month and collect data once a day. After processing through the above steps, we find that the fault feature vector on the 15th day shows that the values ​​of certain feature dimensions are significantly higher than those of other time periods. Further analysis shows that this may be due to the aging of certain cells inside the battery pack. By replacing these aging cells in a timely manner, the overall performance of the battery pack can be effectively restored, its service life can be extended, and the overall performance and safety of the vehicle can be improved. This method not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles.

[0038] In a specific embodiment, the performing fault location analysis on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack includes: Performing a fault sensitivity analysis on the power battery pack based on the fault feature vector to obtain a fault sensitivity matrix of the power battery pack, and performing a singular value decomposition on the fault sensitivity matrix to obtain a fault feature space of the power battery pack; Performing envelope extraction on the fault feature space by Hilbert transform to obtain a fault envelope spectrum of the power battery pack, and performing bidirectional recursive analysis on the fault envelope spectrum to obtain a fault propagation link of the power battery pack; Performing fault delimitation mapping on the power battery pack based on the fault propagation link to obtain a fault boundary sequence, and topologically reconstructing the fault boundary sequence to obtain a fault topology network; Calculating the fault probability of the fault topology network by kernel density estimation to obtain a fault probability distribution, and performing maximum likelihood positioning on the fault probability distribution to obtain a set of fault source points of the power battery pack; Performing a fault influence domain analysis on the power battery pack based on the fault source point set to obtain a fault diffusion matrix of the power battery pack, and performing a spectral clustering analysis on the fault diffusion matrix to obtain a fault area spectrum of the power battery pack; A fault location analysis is performed on the power battery pack using the fault area spectrum to obtain a fault location result of the power battery pack.

[0039] Specifically, the process of performing fault location analysis on the power battery pack based on the fault feature vector to obtain the fault location result of the power battery pack is realized through a series of complex signal processing and data analysis technologies, aiming to extract key information that can reflect the specific fault location inside the battery from the fault feature vector and convert it into a feature representation that is easy to understand and process. First, the obtained fault feature vector will be input into a specially designed algorithm for processing, which is usually based on methods such as fault sensitivity analysis, singular value decomposition, Hilbert transform, bidirectional recursive analysis, topological reconstruction, kernel density estimation, and spectral clustering analysis to reveal the specific fault source point inside the battery pack and its impact range. Specifically, in the application scenario of the power battery pack of an electric vehicle, it is assumed that we have obtained the fault feature vector of the battery pack for a period of time through the previous steps. In order to perform fault location analysis, it is first necessary to perform fault sensitivity analysis on the power battery pack based on the fault feature vector to obtain a fault sensitivity matrix. Fault sensitivity analysis aims to identify which feature dimensions are most sensitive to specific fault modes. For example, in a specific case, it is assumed that we collected fault feature vector data within 5 minutes, collected once per second, and collected a total of 300 data points. By performing fault sensitivity analysis on these data points, we can generate a fault sensitivity matrix containing 128 feature dimensions. Each feature dimension represents a specific change in electrochemical or physical properties, and these features together constitute a complete feature descriptor. Next, singular value decomposition (SVD) is performed on the fault sensitivity matrix to obtain a fault feature space. Singular value decomposition is a method of decomposing complex data into multiple main components, which can extract the most important feature information. For example, in the above case, suppose we perform singular value decomposition on 128 feature dimensions and obtain a fault feature space containing 50 main components. Each main component represents an important fault feature, which provides us with more detailed fault mode information. Subsequently, the fault feature space is envelope extracted by Hilbert transform to obtain the fault envelope spectrum of the battery pack. Hilbert transform is a technique for extracting signal envelopes, which can reveal instantaneous amplitude changes in signals. For example, in the above case, suppose we perform Hilbert transform on 50 main components and obtain a fault envelope spectrum containing 50 frequency components. Each frequency component represents an instantaneous amplitude change at a certain moment, thus providing us with more detailed fault propagation information. Next, we perform a bidirectional recursive analysis on the fault envelope spectrum to obtain the fault propagation link of the power battery pack. Bidirectional recursive analysis is a method that combines forward and reverse time series information to capture the fault propagation path. For example, in the above case, suppose we perform a bidirectional recursive analysis on 50 frequency components and obtain a set of 10 fault propagation links.Each link represents a possible fault propagation path, which provides us with more detailed fault propagation information. Then, based on the fault propagation link, the power battery pack is fault delimited and mapped to obtain a fault boundary sequence, and the fault boundary sequence is topologically reconstructed to obtain a fault topology network. Fault delimitation mapping is a method of mapping fault propagation paths to actual physical locations, which can reveal the specific location of the fault. For example, in the above case, suppose we perform fault delimitation mapping on 10 fault propagation links and obtain a sequence of 50 fault boundaries. By topologically reconstructing these fault boundaries, we can generate a fault topology network containing 10 nodes. Each node represents a possible fault source point, which provides us with more detailed fault location information. Further, the fault probability of the fault topology network is calculated by kernel density estimation to obtain the fault probability distribution, and the fault probability distribution is located by maximum likelihood to obtain the fault source point set of the power battery pack. Kernel density estimation is a method for estimating probability density functions, which can reveal the possibility of fault occurrence. For example, in the above case, suppose we perform kernel density estimation on 10 nodes and obtain a fault source point set containing 5 high-probability fault source points. Each source point represents a location where a fault is most likely to occur, and maximum likelihood localization helps us determine the specific locations of these source points. Next, based on the set of fault source points, the fault impact domain analysis of the power battery pack is performed to obtain the fault diffusion matrix of the power battery pack, and the fault diffusion matrix is ​​subjected to spectral clustering analysis to obtain the fault area spectrum of the power battery pack. Fault impact domain analysis aims to evaluate the impact range of each fault source point. For example, in the above case, suppose we performed fault impact domain analysis on 5 fault source points and obtained a fault diffusion matrix containing 20 nodes. By performing spectral clustering analysis on the matrix, we can generate a fault area spectrum containing 4 fault areas. Each area represents an area with similar fault modes, which provides us with more detailed fault classification information. Finally, the fault location analysis of the power battery pack is performed through the fault area spectrum to obtain the final fault location result. For example, in the above case, suppose we find that a specific fault area is closely related to the aging of certain cells inside the battery pack. By further analyzing the characteristic information in the fault area, the specific location of the faulty cell can be determined, and these aging cells can be replaced in time to restore the overall performance of the battery pack. In a specific example, suppose that the power battery pack of an electric bus has an abnormal decrease in charging and discharging efficiency during a certain driving process.The technicians can conduct a detailed analysis through the above steps: first, use the fault feature vector obtained in the previous step; then, perform fault sensitivity analysis on the fault feature vector to obtain a fault sensitivity matrix; then, perform singular value decomposition on the fault sensitivity matrix to obtain a fault feature space; then, perform envelope extraction on the fault feature space through Hilbert transform to obtain a fault envelope spectrum; then, perform bidirectional recursive analysis on the fault envelope spectrum to obtain a fault propagation link; then, perform fault delimitation mapping based on the fault propagation link to obtain a fault boundary sequence, and perform topological reconstruction on it to obtain a fault topology network; then, calculate the fault probability of the fault topology network through kernel density estimation to obtain a fault probability distribution, and perform maximum likelihood location on it to obtain a fault source point set; then, perform fault influence domain analysis based on the fault source point set to obtain a fault diffusion matrix, and perform spectral clustering analysis on it to obtain a fault area spectrum; finally, perform fault location analysis on the power battery pack through the fault area spectrum to obtain the final fault location result. Through in-depth analysis of these fault location results, technicians can find out whether there is a specific fault mode and its corresponding fault location, thereby providing clear guidance for subsequent maintenance and repair work. For example, in the above case, suppose we find that a specific fault area is closely related to the aging of certain cells inside the battery pack. By further analyzing the characteristic information in the fault area, the specific location of the faulty cells can be determined, and these aging cells can be replaced in time to restore the overall performance of the battery pack. In addition, this method can be applied not only to fault diagnosis, but also to battery performance evaluation and optimization management. For example, by regularly collecting the electrical parameters of the battery pack and performing the above processing, a long-term monitoring system for the battery health status can be established to detect potential problems in time and take preventive measures. In short, the process of fault location analysis based on fault feature vectors not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Throughout the process, the precise fault feature vectors and effective fault location technology work together to capture and analyze even tiny changes in battery status in a timely manner, providing a powerful tool for maintenance personnel. This method greatly enhances the adaptability and response speed of the system, and contributes significantly to the development of new energy vehicles. For example, suppose we monitor the power battery pack of an electric bus for a month, collecting data once a day. After the above steps, we found that the fault location results on the 15th day showed that the fault probability of certain fault areas was significantly higher than that of other time periods. Further analysis showed that this may be due to the aging of certain cells inside the battery pack. By replacing these aging cells in a timely manner, the overall performance of the battery pack can be effectively restored, its service life can be extended, and the overall performance and safety of the vehicle can be improved.This approach not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles.

[0040] In a specific embodiment, the performing of a fault influence domain analysis on the power battery pack based on the fault source point set to obtain a fault diffusion matrix of the power battery pack includes: Based on the fault source point set, the power battery pack is grid-segmented to obtain a multi-layer fault grid map, and a topological connectivity analysis is performed on the multi-layer fault grid map to obtain a fault conduction path set; Performing fault diffusion dynamics modeling based on the fault conduction path set to obtain a fault diffusion dynamics equation, and performing fault evolution analysis on the fault diffusion dynamics equation to obtain a fault evolution characteristic diagram; Calculating the fault influence intensity of the power battery pack based on the fault evolution characteristic diagram to obtain a fault weight vector, and performing nonlinear coupling analysis on the fault weight vector to obtain a fault coupling network; A causal association analysis is performed on the fault coupling network to obtain a fault transmission chain, and the fault impact of the power battery pack is quantified based on the fault transmission chain to obtain a fault impact matrix, and a spectral clustering analysis is performed on the fault impact matrix to obtain a fault impact area set of the power battery pack, and a fault diffusion matrix of the power battery pack is obtained based on the fault impact area set.

[0041] Specifically, the fault influence domain analysis of the power battery pack is performed based on the fault source point set to obtain the fault diffusion matrix of the power battery pack. This process is achieved through a series of complex signal processing and data analysis technologies, aiming to extract key information that can reflect the specific fault propagation path and its impact range inside the battery from the fault source point set, and convert it into a feature representation that is easy to understand and process. First, the obtained fault source point set will be input into a specially designed algorithm for processing. These algorithms are usually based on grid segmentation, topological connectivity analysis, fault diffusion dynamics modeling, fault evolution analysis, nonlinear coupling analysis, and pedigree clustering analysis to reveal the specific fault propagation path and its impact range inside the battery pack. Specifically, in the application scenario of the power battery pack of electric vehicles, it is assumed that we have obtained the fault source point set of the battery pack in a certain period of time through the previous steps. In order to perform fault influence domain analysis, it is first necessary to grid segment the power battery pack based on the fault source point set to obtain a multi-layer fault grid map. Grid segmentation is a method of dividing a complex system into multiple small areas, which can more accurately locate the fault location and its impact range. For example, in a specific case, suppose we collected data on the fault source point set within 5 minutes, once every second, and collected a total of 300 data points. By gridding these data points, we can generate a multi-layer fault grid map, where each grid represents a possible fault location. Suppose we divide the entire battery pack into 100 grids, each of which represents an independent unit or component. Next, perform topological connectivity analysis on the multi-layer fault grid map to obtain a set of fault conduction paths. Topological connectivity analysis aims to identify which grids have physical connections or functional associations, thereby determining possible fault propagation paths. For example, in the above case, suppose we performed topological connectivity analysis on 100 grids and obtained a set of 50 fault conduction paths. Each path represents a possible fault propagation path, which provides us with more detailed fault propagation information. Subsequently, based on the set of fault conduction paths, fault diffusion dynamics modeling is performed to obtain the fault diffusion dynamics equation, and fault evolution analysis is performed on the fault diffusion dynamics equation to obtain the fault evolution feature map. Fault diffusion dynamics modeling is a method to simulate how faults propagate in a system, which can reveal the law of fault change over time. For example, in the above case, suppose we modeled the fault diffusion dynamics of 50 fault conduction paths and obtained a fault diffusion dynamics equation that describes the fault diffusion process. By performing fault evolution analysis on this equation, we can generate a fault evolution feature map containing 10 time points. Each time point represents the fault state at a specific moment, thus providing us with more detailed fault propagation dynamics information.Next, the fault impact intensity of the power battery pack is calculated based on the fault evolution characteristic graph to obtain the fault weight vector, and the fault weight vector is subjected to nonlinear coupling analysis to obtain the fault coupling network. The fault impact intensity calculation is intended to evaluate the impact of each fault node, while the nonlinear coupling analysis is used to reveal the interaction between different fault nodes. For example, in the above case, suppose we calculated the fault impact intensity of the fault evolution characteristic graph at 10 time points and obtained a fault weight vector containing 100 elements. Each element represents the impact intensity of a certain fault node. By performing nonlinear coupling analysis on the vector, we can generate a fault coupling network containing 50 nodes. Each node represents a fault node, and the edges between nodes represent the interaction relationship between them. Then, the fault coupling network is subjected to causal association analysis to obtain the fault transmission chain, and the fault impact quantification of the power battery pack is performed based on the fault transmission chain to obtain the fault impact matrix. The causal association analysis is intended to identify which fault nodes have the strongest causal relationship, while the fault impact quantification is used to evaluate the impact of each fault node on the overall system. For example, in the above case, suppose we perform causal association analysis on the fault coupling network of 50 nodes and obtain a set of 20 fault transmission chains. Each link represents a possible fault propagation path. By quantifying the fault impact of these fault transmission chains, we can generate a fault impact matrix containing 100 elements. Each element represents the degree of influence of a certain fault node on the overall system. Further, the fault impact matrix is ​​subjected to spectral clustering analysis to obtain the set of fault impact regions of the battery pack, and the fault diffusion matrix of the power battery pack is obtained based on the set of fault impact regions. Pedigree clustering analysis is a method of classifying similar nodes into one category, which can reveal the intrinsic connection between fault nodes. For example, in the above case, suppose we perform spectral clustering analysis on the fault impact matrix of 100 elements and obtain a set of fault impact regions containing 4 fault impact regions. Each region represents an area with similar fault modes, which provides us with more detailed fault classification information. Finally, based on the set of fault impact regions, we can generate a fault diffusion matrix containing 100 elements, each element representing the degree of influence of a certain fault node on the overall system. In a specific example, suppose that the power battery pack of an electric bus has an abnormal decrease in charging and discharging efficiency during a certain driving process.The technicians can conduct a detailed analysis through the above steps: first, use the fault source point set obtained in the previous step; then, mesh the fault source point set to obtain a multi-layer fault grid map; then, perform topological connectivity analysis on the multi-layer fault grid map to obtain a fault conduction path set; then, perform fault diffusion dynamics modeling based on the fault conduction path set to obtain the fault diffusion dynamics equation, and perform fault evolution analysis on it to obtain a fault evolution feature map; then, calculate the fault influence intensity based on the fault evolution feature map to obtain the fault weight vector, and perform nonlinear coupling analysis on it to obtain a fault coupling network; then, perform causal association analysis on the fault coupling network to obtain the fault transmission chain, and quantify the fault influence based on the fault transmission chain to obtain the fault influence degree matrix; finally, perform pedigree clustering analysis on the fault influence degree matrix to obtain the fault impact area set, and obtain the final fault diffusion matrix based on the fault impact area set. Through in-depth analysis of these fault diffusion matrices, technicians can find out whether there is a specific fault mode and its corresponding fault propagation path and impact range, thereby providing clear guidance for subsequent maintenance and repair work. For example, in the above case, suppose we find that a specific fault area is closely related to the aging of certain cells inside the battery pack. By further analyzing the characteristic information in the fault area, the specific location of the faulty cells can be determined, and these aging cells can be replaced in time to restore the overall performance of the battery pack. In addition, this method can be applied not only to fault diagnosis, but also to battery performance evaluation and optimization management. For example, by regularly collecting the electrical parameters of the battery pack and performing the above processing, a long-term monitoring system for the battery health status can be established to detect potential problems in time and take preventive measures. In summary, the process of fault impact domain analysis based on the fault source point set not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles. Throughout the process, the precise fault source point set and the effective fault impact domain analysis technology work together, so that even small changes in the battery state can be captured and analyzed in time, providing a powerful tool for maintenance personnel. This method greatly enhances the adaptability and response speed of the system, and contributes to the development of new energy vehicles. For example, suppose we monitor the power battery pack of an electric bus for a month, collecting data once a day. After processing through the above steps, we found that the fault diffusion matrix on the 15th day showed that the failure probability of certain fault areas was significantly higher than that of other time periods. Further analysis showed that this may be due to the aging of certain cells inside the battery pack. By replacing these aging cells in a timely manner, the overall performance of the battery pack can be effectively restored, its service life can be extended, and the overall performance and safety of the vehicle can be improved. This method not only improves the accuracy and reliability of fault detection, but also provides a scientific basis for optimizing the battery management system, ensuring the safety and efficient operation of electric vehicles.

[0042] The above describes the fault diagnosis method of the power battery pack in the embodiment of the present invention. The following describes the fault diagnosis device of the power battery pack in the embodiment of the present invention. Figure 2 , an embodiment of a fault diagnosis device for a power battery pack in an embodiment of the present invention includes: The acquisition module 21 is used to acquire electrical parameters of the power battery pack through a preset data acquisition device to obtain electrical acquisition parameters of the power battery pack; A first analysis module 22, configured to perform phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack; A second analysis module 23 is used to perform electrochemical impedance spectrum analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnose whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack; A mapping module 24, configured to perform fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack, if any; The positioning module 25 is used to perform fault positioning analysis on the power battery pack based on the fault feature vector to obtain a fault positioning result of the power battery pack.

[0043] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the above method embodiment, which will not be repeated here.

[0044] Reference Figure 3 The present invention also provides a computer device in an embodiment, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0045] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0046] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0047] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0048] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0049] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for diagnosing a fault of a power battery pack, characterized in that: The following steps are involved: The electrical parameters of the power battery pack are collected by a preset data collection device to obtain the electrical collection parameters of the power battery pack; Performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack; Performing electrochemical impedance spectroscopy analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnosing whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack; If so, performing fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack; A fault location analysis is performed on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack.

2. The fault diagnosis method of the power battery pack according to claim 1, characterized in that: The electrical parameters of the power battery pack are collected by a preset data collection device to obtain the electrical collection parameters of the power battery pack, including: Collecting electrical parameters of the power battery pack through a preset data acquisition device to obtain preliminary collected electrical parameters of the power battery pack; Performing discrete wavelet transform on the preliminary collected electrical parameters to obtain multi-resolution electrical parameter data of the power battery pack; wherein the preliminary collected electrical parameters include voltage parameters and current collection parameters; Performing singular value decomposition on the power battery pack based on the multi-resolution electrical parameter data to obtain electrical singular value data; The electrical singular value data are denoised and subjected to principal component analysis to obtain electrical acquisition parameters of the power battery pack.

3. The fault diagnosis method of the power battery pack according to claim 1, characterized in that: The performing phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack includes: Performing time-frequency domain conversion on the power battery pack based on the electrical acquisition parameters to obtain time-frequency characteristic data of the power battery pack, and performing complex domain mapping on the time-frequency characteristic data to obtain a complex domain characteristic sequence of the power battery pack; Performing phase angle extraction on the complex domain characteristic sequence to obtain an instantaneous phase sequence of the power battery pack; and performing phase unwrapping processing on the instantaneous phase sequence to obtain a continuous phase sequence; Performing phase difference calculation on the continuous phase sequence to obtain a phase difference feature of the power battery pack, and performing multi-scale decomposition based on the phase difference feature to obtain a multi-scale phase component; Performing energy distribution calculation on the multi-scale phase component to obtain a phase energy spectrum of the power battery pack, and performing nonlinear feature extraction on the phase energy spectrum to obtain a nonlinear phase feature; Phase reconstruction is performed based on the nonlinear phase characteristics to obtain a phase characteristic spectrum of the power battery pack; wherein the phase characteristic spectrum is used to reflect the spectrum data of the overall phase change characteristics in the power battery pack.

4. The fault diagnosis method of the power battery pack according to claim 1, characterized in that: The performing electrochemical impedance spectrum analysis on the power battery pack based on the phase characteristic spectrum to obtain the impedance characteristic spectrum of the power battery pack includes: Performing a complex domain transformation on the power battery pack based on the phase characteristic spectrum to obtain a complex impedance sequence of the power battery pack, and performing a multi-scale discrete Fourier analysis on the complex impedance sequence to obtain a frequency domain impedance component; Performing a time-frequency joint decomposition on the frequency domain impedance component by a preset adaptive kernel function to obtain a time-frequency impedance matrix, and performing a bispectral analysis on the time-frequency impedance matrix to obtain a high-order impedance feature; Performing variational mode decomposition on the power battery pack based on the high-order impedance characteristics to obtain an intrinsic impedance function of the power battery pack, and performing Volterra series expansion on the intrinsic impedance function to obtain a nonlinear impedance kernel function of the power battery pack; The nonlinear impedance kernel function is eigen-decomposed by a preset covariance matrix to obtain an impedance eigenvector, and a spectrum is reconstructed based on the impedance eigenvector to obtain an impedance characteristic spectrum of the power battery pack.

5. The fault diagnosis method of the power battery pack according to claim 1, characterized in that: The performing fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack includes: Based on the impedance characteristic spectrum, a spectral domain hierarchical analysis is performed on the power battery pack to obtain a multi-layer impedance characteristic set of the power battery pack, and based on the multi-layer impedance characteristic set, a high-dimensional feature transformation is performed on the power battery pack to obtain a transformation characteristic space of the power battery pack, and a fault correlation analysis is performed on the transformation characteristic space to obtain a fault correlation spectrum of the power battery pack; Decoupling the fault mode of the power battery pack based on the fault correlation spectrum to obtain a decoupling feature matrix, and performing fuzzy clustering analysis on the decoupling feature matrix to obtain a fault clustering spectrum of the power battery pack; The fault cluster spectrum is subjected to feature fusion through adaptive fusion to obtain a fused feature vector, and the fused feature vector is used as a fault feature vector of the power battery pack.

6. The fault diagnosis method of the power battery pack according to claim 1, characterized in that: The performing fault location analysis on the power battery pack based on the fault feature vector to obtain a fault location result of the power battery pack includes: Performing a fault sensitivity analysis on the power battery pack based on the fault feature vector to obtain a fault sensitivity matrix of the power battery pack, and performing a singular value decomposition on the fault sensitivity matrix to obtain a fault feature space of the power battery pack; Performing envelope extraction on the fault feature space by Hilbert transform to obtain a fault envelope spectrum of the power battery pack, and performing bidirectional recursive analysis on the fault envelope spectrum to obtain a fault propagation link of the power battery pack; Performing fault delimitation mapping on the power battery pack based on the fault propagation link to obtain a fault boundary sequence, and topologically reconstructing the fault boundary sequence to obtain a fault topology network; Calculating the fault probability of the fault topology network by kernel density estimation to obtain a fault probability distribution, and performing maximum likelihood positioning on the fault probability distribution to obtain a set of fault source points of the power battery pack; Performing a fault influence domain analysis on the power battery pack based on the fault source point set to obtain a fault diffusion matrix of the power battery pack, and performing a spectral clustering analysis on the fault diffusion matrix to obtain a fault area spectrum of the power battery pack; A fault location analysis is performed on the power battery pack using the fault area spectrum to obtain a fault location result of the power battery pack.

7. The fault diagnosis method of the power battery pack according to claim 6, characterized in that: The performing a fault influence domain analysis on the power battery pack based on the fault source point set to obtain a fault diffusion matrix of the power battery pack includes: Based on the fault source point set, the power battery pack is grid-segmented to obtain a multi-layer fault grid map, and a topological connectivity analysis is performed on the multi-layer fault grid map to obtain a fault conduction path set; Performing fault diffusion dynamics modeling based on the fault conduction path set to obtain a fault diffusion dynamics equation, and performing fault evolution analysis on the fault diffusion dynamics equation to obtain a fault evolution characteristic diagram; Calculating the fault influence intensity of the power battery pack based on the fault evolution characteristic diagram to obtain a fault weight vector, and performing nonlinear coupling analysis on the fault weight vector to obtain a fault coupling network; A causal association analysis is performed on the fault coupling network to obtain a fault transmission chain, and the fault impact of the power battery pack is quantified based on the fault transmission chain to obtain a fault impact matrix, and a spectral clustering analysis is performed on the fault impact matrix to obtain a fault impact area set of the power battery pack, and a fault diffusion matrix of the power battery pack is obtained based on the fault impact area set.

8. A fault diagnosis device for a power battery pack, characterized in that: include: The acquisition module is used to acquire electrical parameters of the power battery pack through a preset data acquisition device to obtain electrical acquisition parameters of the power battery pack; A first analysis module, configured to perform phase spectrum analysis on the power battery pack based on the electrical acquisition parameters to obtain a phase characteristic spectrum of the power battery pack; A second analysis module is used to perform electrochemical impedance spectroscopy analysis on the power battery pack based on the phase characteristic spectrum to obtain an impedance characteristic spectrum of the power battery pack, and diagnose whether the power battery pack has a fault based on the impedance characteristic spectrum of the power battery pack; A mapping module, configured to perform fault feature mapping on the power battery pack based on the impedance characteristic spectrum to obtain a fault feature vector of the power battery pack, if any; The positioning module is used to perform fault positioning analysis on the power battery pack based on the fault feature vector to obtain a fault positioning result of the power battery pack.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images