Remote fault diagnosis method and system for intelligent energy storage power supply

Abstract

The invention relates to a remote fault diagnosis method and system for an intelligent energy storage power supply, and the method comprises the following steps: carrying out the real-time monitoring of electrical parameters of the energy storage power supply, capturing abnormal fluctuation, and transmitting the abnormal fluctuation to a remote intelligent diagnosis center; and extracting a fault feature vector group through a deep learning algorithm, analyzing a fault propagation path, and generating a fault propagation link diagram. Causal relationship reasoning is carried out based on the graph, and a fault source and a fault type are positioned. And according to the fault source and the type evaluation influence range, obtaining a fault influence degree evaluation report. Finally, an intelligent fault processing suggestion scheme is generated according to the report, rapid and accurate fault diagnosis and processing are realized, and safe and stable operation of the energy storage system is ensured. The technical problems that a traditional fault diagnosis method often depends on the experience of field technicians and simple fault detection equipment, time and labor are consumed, and in the face of complex or hidden faults, the traditional fault diagnosis method is often out of effort are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage power supplies, and in particular to a remote fault diagnosis method and system for an intelligent energy storage power supply. Background Art

[0002] The development and application of energy storage power supply technology in current vehicle electrical systems, particularly the widespread adoption of intelligent energy storage power supplies in modern transportation vehicles like electric vehicles, has led to higher demands for the reliability and safety of these power systems. However, traditional fault diagnosis methods often rely on the experience of on-site technicians and simple fault detection equipment. This is not only time-consuming and labor-intensive, but also often falls short when faced with complex or hidden faults. Especially in emergency situations, it is crucial to promptly and accurately locate the source of the fault and take effective measures.

[0003] Traditional methods are also limited by their inability to achieve remote monitoring and diagnosis, which poses a significant challenge for managing fleets deployed with large-scale intelligent energy storage systems. Vehicles are often geographically dispersed, and when a fault occurs, technicians must be dispatched to the site for on-site troubleshooting, a process that is both inefficient and increases operating costs. Furthermore, due to the varying structures and operating principles of different energy storage systems, there is no universal method for troubleshooting in all situations, further complicating the issue.

[0004] To address these issues, it is crucial to research and develop a remote fault diagnosis method for intelligent energy storage power supplies based on deep learning algorithms. This method not only monitors the operating status of the energy storage power supply in real time but also automatically analyzes the cause of any anomalies detected and provides intelligent action suggestions, significantly improving the accuracy and efficiency of fault diagnosis. By establishing a fault propagation link diagram and reasoning about causal relationships, this method can more accurately locate the root cause of a fault and its impact, enabling vehicle maintenance teams to quickly respond and resolve issues, ensuring safe and stable vehicle operation. Summary of the Invention

[0005] The main purpose of the present invention is to provide a remote fault diagnosis method and system for an intelligent energy storage power supply, which solves the technical problem that traditional fault diagnosis methods often rely on the experience of on-site technicians and simple fault detection equipment, which is not only time-consuming and labor-intensive, but also often seems powerless when faced with complex or hidden faults.

[0006] To achieve the above objectives, the present invention provides a remote fault diagnosis method for an intelligent energy storage power supply, which is applied to a vehicle equipped with an energy storage power supply, and comprises the following steps: Performing real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operation status data set; When there is abnormal fluctuation in the electrical operation status data set, the electrical operation status data set is sent to a remote intelligent diagnosis center; Extracting fault features from the electrical operation status data set using a deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group; Performing fault propagation path analysis on the energy storage power supply based on the fault feature vector group to obtain a fault propagation link diagram; Performing causal reasoning on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result; Based on the fault root cause location result and the fault type, the fault impact range of the energy storage power supply is evaluated to obtain a fault impact assessment report; and based on the fault impact assessment report, an intelligent fault handling suggestion plan is generated.

[0007] Furthermore, the real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operation status data set includes: Performing multi-dimensional parameter acquisition on the battery pack of the energy storage power supply through a preset electrical acquisition sensor to obtain an original data set of the battery pack; preprocessing the original data set to obtain a preprocessed electrical parameter set; Performing time sequence status monitoring and analysis on the pre-processed electrical parameter set to obtain an electrical operation status data set.

[0008] Furthermore, the fault feature extraction is performed on the electrical operation status data set by the deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group, including: Performing time-frequency domain decomposition processing on the electrical operation status data set using the deep learning algorithm of the remote intelligent diagnosis center to obtain a multi-dimensional electrical feature matrix; Performing nonlinear dynamic characteristic analysis on the energy storage power supply based on the multidimensional electrical characteristic matrix to obtain a fault-related characteristic sequence, and performing adaptive quantization processing on the fault-related characteristic sequence to obtain a fault quantization characteristic set; Performing multi-scale fusion processing on the fault quantization feature set to obtain a fused feature vector, and performing fault correlation analysis based on the fused feature vector to obtain a fault correlation matrix; Performing fault feature dimensionality reduction processing on the fault correlation matrix to obtain a reduced-dimensionality feature group, and performing fault feature optimization clustering on the reduced-dimensionality feature group to obtain a fault feature vector group.

[0009] Furthermore, the fault propagation path analysis of the energy storage power supply is performed based on the fault feature vector group to obtain a fault propagation link diagram, including: Performing a spatiotemporal correlation analysis on the fault feature vector group to obtain a spatiotemporal feature matrix of fault propagation, and performing a dynamic topological structure analysis on the spatiotemporal feature matrix of fault propagation to obtain a fault propagation topology network; Performing fault propagation dynamics modeling on the energy storage power source based on the fault propagation topology network to obtain a fault propagation dynamics model, and performing multi-level coupling analysis on the fault propagation dynamics model to obtain a fault propagation coupling relationship diagram; performing fault propagation path tracing on the fault propagation coupling relationship graph to obtain a fault propagation path set, and performing propagation strength evaluation on the fault propagation path set to obtain a fault propagation strength matrix; A fault propagation link is constructed for the energy storage power source based on the fault propagation intensity matrix to obtain an initial fault propagation link, and the initial fault propagation link is optimized and reconstructed to obtain a fault propagation link graph.

[0010] Furthermore, the causal relationship reasoning is performed on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result, including: Performing temporal causal chain decomposition on the fault propagation link graph to obtain a multi-level causal relationship network, and performing dynamic weight calculation on the multi-level causal relationship network to obtain a causal strength matrix, wherein the causal strength matrix includes fault node impact weight, fault propagation temporal weight, fault correlation weight, and fault persistence weight; Performing causal link tracing analysis on the energy storage power supply based on the causal strength matrix to obtain a fault causal propagation sequence, and performing hierarchical decomposition processing on the fault causal propagation sequence to obtain a fault causal hierarchy diagram, wherein the fault causal hierarchy diagram includes first-level fault source characteristics, second-level fault source characteristics, fault transmission characteristics, and fault evolution characteristics; Performing causal correlation measurement on the fault causal hierarchy graph to obtain a causal correlation measurement matrix, and identifying fault source nodes based on the causal correlation measurement matrix to obtain a fault source candidate set; Performing a fault source credibility assessment on the energy storage power supply based on the fault source candidate set to obtain a fault source credibility assessment matrix, and performing a multi-dimensional cross-validation on the fault source credibility assessment matrix to obtain a fault root cause location result; Perform fault type matching analysis on the fault root cause location result to obtain the fault type, wherein the fault type includes battery pack performance degradation fault, battery pack internal short circuit fault, battery pack abnormal temperature fault and battery pack abnormal charging and discharging fault.

[0011] Furthermore, the fault impact range of the energy storage power supply is evaluated based on the fault root location result and the fault type to obtain a fault impact assessment report, including: Performing fault diffusion dynamic modeling analysis on the fault root cause location result to obtain a fault diffusion feature matrix, and performing multi-dimensional quantization processing on the fault diffusion feature matrix to obtain a fault diffusion quantization set; Tracing the fault-affected links of the energy storage power source based on the fault diffusion quantization set to obtain a fault-affected propagation network, and performing hierarchical decomposition processing on the fault-affected propagation network to obtain a fault-affected hierarchical map; Performing a fault loss assessment analysis on the fault impact hierarchy map to obtain a fault loss assessment matrix, and performing multi-dimensional risk quantification based on the fault loss assessment matrix to obtain a fault risk quantification set; Based on the fault risk quantification set, a fault impact assessment is performed on the energy storage power supply to obtain a fault impact assessment report.

[0012] Furthermore, the fault impact assessment of the energy storage power source is performed based on the fault risk quantification set to obtain a fault impact assessment report, including: Performing multi-dimensional feature decomposition on the fault risk quantification set to obtain a fault risk feature vector group, and performing time series correlation analysis on the fault risk feature vector group to obtain a fault risk time series correlation matrix; quantifying the degree of fault impact of the energy storage power supply based on the fault risk time series correlation matrix to obtain a quantified set of fault impact degrees, and performing multi-level decomposition processing on the quantified set of fault impact degrees to obtain a fault impact level feature graph; Tracing the fault impact propagation path of the fault impact hierarchical feature graph to obtain a fault impact propagation link set, and performing fault impact intensity assessment based on the fault impact propagation link set to obtain a fault impact intensity assessment matrix; Defining the fault impact range of the energy storage power supply based on the fault impact intensity assessment matrix to obtain a fault impact range boundary map; Performing a comprehensive assessment of the degree of fault impact of the energy storage power supply based on the fault impact range boundary map to obtain a fault impact degree assessment set, and performing fault risk level classification based on the fault impact degree assessment set to obtain a fault risk level matrix; A fault impact assessment report is generated for the energy storage power supply based on the fault risk level matrix to obtain a fault impact assessment report.

[0013] The present invention also provides a remote fault diagnosis system for an intelligent energy storage power supply, which is applied to a vehicle and includes: A monitoring module, configured to monitor electrical parameters of the energy storage power supply in real time and obtain an electrical operation status data set; a judgment module, configured to send the electrical operation status data set to a remote intelligent diagnosis center when abnormal fluctuations occur in the electrical operation status data set; An extraction module, configured to extract fault features from the electrical operation status data set using a deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group; an analysis module, configured to perform fault propagation path analysis on the energy storage power supply based on the fault feature vector group to obtain a fault propagation link diagram; An inference module, configured to perform causal reasoning on the energy storage power supply based on the fault propagation link diagram, and obtain a fault root cause location result and a fault type corresponding to the fault root cause location result; An evaluation module is used to evaluate the fault impact range of the energy storage power supply based on the fault root cause location result and the fault type to obtain a fault impact assessment report; and to generate an intelligent fault handling suggestion plan based on the fault impact assessment report.

[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 having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0016] The remote fault diagnosis method for an intelligent energy storage power supply provided by the present invention comprises the following steps: real-time monitoring of the electrical parameters of the energy storage power supply, capturing abnormal fluctuations and sending them to a remote intelligent diagnosis center. A fault feature vector group is extracted through a deep learning algorithm, the fault propagation path is analyzed, and a fault propagation link diagram is generated. Causal reasoning is performed based on this diagram to locate the root cause of the fault and its type. The impact range is evaluated based on the root cause and type of the fault, and a fault impact assessment report is obtained. Finally, an intelligent fault handling recommendation plan is generated based on the report to achieve rapid and accurate fault diagnosis and processing, ensuring the safe and stable operation of the energy storage system. This solves the technical problem that traditional fault diagnosis methods often rely on the experience of on-site technicians and simple fault detection equipment, which is not only time-consuming and labor-intensive, but also often seems powerless when faced with complex or hidden faults. It realizes the location of the root cause of the fault through causal reasoning and determines the corresponding fault type. This method provides a scientific and systematic means to analyze problems, ensures the accuracy of fault cause judgment, and provides a clear direction for subsequent maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 1 is a schematic diagram of the steps of a remote fault diagnosis method for an intelligent energy storage power supply according to an embodiment of the present invention; Figure 2This is a structural block diagram of a remote fault diagnosis system for an intelligent energy storage power supply according to an 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 purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to 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 remote fault diagnosis method for an intelligent energy storage power supply in one embodiment of the present invention; In one embodiment of the present invention, a remote fault diagnosis method for an intelligent energy storage power supply is provided, which is applied to a vehicle provided with an energy storage power supply, and includes the following steps: Step S1: Real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operation status data set.

[0021] Specifically, real-time monitoring of the electrical parameters of the energy storage power supply is performed to generate an electrical operating status dataset. This process is the fundamental component of the entire intelligent energy storage power supply remote fault diagnosis method. Its core lies in continuously acquiring various electrical parameters generated by the energy storage power supply during operation through high-precision sensors and data acquisition equipment. These parameters include, but are not limited to, key indicators such as voltage, current, temperature, power factor, and charge-discharge efficiency, which together constitute the core content of the electrical operating status dataset. To achieve this goal, it is first necessary to deploy multiple types of sensors at key nodes of the energy storage power supply, such as current transformers to monitor current changes and thermistors to detect temperature fluctuations. Furthermore, it is necessary to utilize high-precision data acquisition modules to integrate these dispersed signals into a unified digital format, thereby forming a complete electrical operating status dataset. In practical applications, for example, during the operation of an electric vehicle, its energy storage power supply will continuously generate real-time changes in electrical parameters. When the vehicle is at high speed or accelerating rapidly, the battery output current may suddenly increase. At this time, if the temperature of a battery cell rises abnormally, the sensor will capture this change and convert it into data recorded in the electrical operating status dataset. This real-time monitoring method not only fully reflects the operating status of the energy storage power supply but also provides reliable data support for subsequent fault diagnosis. This is because only based on accurate and comprehensive operating status data can the remote intelligent diagnosis center's deep learning algorithm effectively extract fault feature vectors, thereby achieving accurate fault analysis and location. Therefore, this link plays a connecting role in the entire method, guaranteeing both the data source and the foundation for subsequent analysis.

[0022] Step S2: When there is abnormal fluctuation in the electrical operation status data set, the electrical operation status data set is sent to a remote intelligent diagnosis center.

[0023] Specifically, when the electrical operating status dataset exhibits abnormal fluctuations, it is transmitted to a remote intelligent diagnostic center. This process relies on an intelligent data analysis and transmission mechanism. First, the system dynamically analyzes the real-time monitored electrical operating status dataset, using a pre-set algorithm or model to determine whether the data exhibits abnormal fluctuations. For example, if the voltage of the energy storage power supply suddenly exceeds the normal range, or if the current fluctuates dramatically within a short period of time, these abnormalities are identified by the system as potential fault signals. To ensure the accuracy of detecting abnormal fluctuations, a comparative analysis is typically conducted between historical and current data, while statistical methods or machine learning models are used to set reasonable thresholds to avoid false positives or false negatives. For example, in an electric vehicle, if the temperature of the energy storage power supply suddenly rises while the vehicle is driving smoothly, and the current output also exhibits erratic fluctuations, the system will identify this as abnormal fluctuation and trigger subsequent data transmission. Once the abnormal fluctuations in the electrical operating status dataset are confirmed, the system will immediately transmit the dataset to the remote intelligent diagnostic center via wireless communication technology. This transmission process requires high-speed, stable communication networks, such as 5G or dedicated IoT communication protocols, to ensure that data can be delivered to the diagnostic center quickly and completely. At the same time, in order to ensure the security and integrity of the data, encryption technology and verification mechanisms are also used during the transmission process to prevent data tampering or loss during transmission. For example, in the above-mentioned electric vehicle application scenario, when the vehicle's energy storage power supply experiences abnormal fluctuations, the on-board terminal will automatically package the relevant electrical operating status data set and upload it to the remote intelligent diagnostic center through the Internet of Vehicles for further analysis by subsequent deep learning algorithms. This timely data transmission mechanism not only provides real-time data support for remote diagnosis, but also buys valuable time for quickly locating the root cause of the fault, thereby effectively improving the response speed and reliability of the entire fault diagnosis system.

[0024] Step S3: extracting fault features from the electrical operation status data set using the deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group.

[0025] Specifically, the remote intelligent diagnostic center uses a deep learning algorithm to extract fault features from the electrical operating status dataset, generating a set of fault feature vectors. This process is the core technical component of the entire fault diagnosis method, crucial for leveraging advanced deep learning models to mine potential fault features from complex datasets. Specifically, after the electrical operating status dataset is transmitted to the remote intelligent diagnostic center, the system first preprocesses the data, including noise removal, missing value filling, and normalization, to ensure data quality and consistency. This processed data is then fed into a pre-trained deep learning model, typically based on architectures such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), or autoencoders. This model automatically learns patterns in the data and extracts fault-related features. For example, in an electric vehicle application scenario, if the voltage and temperature data of the energy storage power supply exhibit abnormal fluctuations over a certain period of time, the deep learning algorithm analyzes the time series characteristics of this data and the correlations between multiple parameters to identify possible fault features, such as battery cell overheating or internal short circuits. Based on this, a deep learning algorithm converts the extracted fault features into a set of high-dimensional fault feature vectors. These vectors not only capture the specific manifestations of the fault, but also reflect the likely cause and severity of the fault. To achieve this, the model must be trained with extensive historical data to ensure it can accurately distinguish between normal and abnormal operating conditions and exhibit strong generalization capabilities. For example, in the aforementioned electric vehicle example, the remote intelligent diagnostic center might train a deep learning model based on historical data from thousands of past energy storage power supply failures to accurately identify different types of faults. When a new dataset of electrical operating conditions is input, the model can quickly extract fault feature vectors corresponding to the current abnormal fluctuations, providing a reliable basis for subsequent fault propagation path analysis. This deep learning-based fault feature extraction method not only significantly improves diagnostic accuracy but also significantly reduces fault analysis time, thereby effectively ensuring the safe operation of the vehicle.

[0026] Step S4: performing fault propagation path analysis on the energy storage power supply based on the fault feature vector group to obtain a fault propagation link diagram.

[0027] Specifically, the fault propagation path analysis of the energy storage power supply is performed based on the fault feature vector group, generating a fault propagation chain diagram. This process aims to deeply analyze the potential fault propagation mechanisms within the energy storage power supply. First, using the fault feature vector group extracted by the deep learning algorithm in the previous step, the system can identify specific abnormal conditions and their possible causes during the operation of the energy storage power supply. Next, to further understand how these faults propagate from one component to another and ultimately affect the entire energy storage system, a detailed fault propagation chain analysis model is constructed. This model typically uses graph theory and network analysis methods, treating each key component of the energy storage power supply (such as battery cells, circuit boards, sensors, etc.) as nodes and the physical connections or functional dependencies between them as edges. For example, in an electric vehicle application scenario, if the diagnostic center discovers an abnormally high temperature in a battery cell and unstable current output, the fault feature vector group may indicate battery overheating and the resulting activation of circuit protection mechanisms. Using this information, the system then constructs a fault propagation chain diagram. During this process, each fault feature is mapped to a corresponding component, and the potential fault propagation path is determined based on the interactions between components. For example, in the aforementioned case, if a battery cell overheats, causing a nearby temperature sensor to trigger an alarm, which in turn prompts the battery management system to take measures to limit current input to avoid greater risk, this chain reaction is recorded and forms a clear fault propagation chain diagram. This not only helps technicians intuitively understand the development of the fault but also provides an important basis for subsequent causal reasoning. Furthermore, by continuously updating and optimizing this fault propagation chain diagram, the accuracy of future predictions of similar faults can be improved, making maintenance work more efficient and targeted. Therefore, this link is crucial to improving the reliability and safety of energy storage power supplies and lays a solid foundation for the safe and stable operation of vehicles.

[0028] Step S5: performing causal reasoning on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result.

[0029] Specifically, causal reasoning is performed on the energy storage power supply based on the fault propagation chain diagram to obtain a fault root cause location result and the corresponding fault type. This process aims to deeply analyze the fundamental cause and specific type of the fault. First, after obtaining a detailed fault propagation chain diagram, the system uses causal reasoning algorithms to analyze the interactions and impact paths between components. These algorithms may include Bayesian networks, decision trees, or other forms of logical reasoning models, which can infer the most likely fault source based on known fault characteristics and propagation paths. For example, in an electric vehicle application scenario, if the fault propagation chain diagram shows an abnormal increase in battery cell temperature, which further leads to unstable current output and triggers the battery management system's protection mechanism, causal reasoning can gradually eliminate factors that are unlikely to have caused the initial problem, such as external environmental factors or failures of non-critical components. During this process, the system carefully examines the correlation between each potential fault point and the observed phenomenon, and how these phenomena spread along the fault propagation chain diagram. Taking battery overheating as an example, if analysis indicates that the overheating is not caused by external high temperatures but rather by an internal short circuit in the battery, the system will use this finding as one of the results of fault root cause location and further determine that the fault is an electrical fault. To ensure the accuracy of the reasoning, the system also verifies it by combining historical data and learning results from similar cases to ensure that the proposed fault root cause location is not only reasonable but also has a high degree of confidence. Ultimately, through a comprehensive analysis of the entire fault propagation path, the system can not only clearly indicate the specific fault location of the energy storage power supply, such as a specific battery module or a specific circuit in the management system, but also accurately determine the fault type, such as electrical, mechanical, or software. This analysis provides a solid theoretical basis for the subsequent development of effective repair strategies, while also greatly improving the efficiency and targetedness of maintenance work, ensuring the safe and stable operation of the vehicle.

[0030] Step S6: Based on the fault root location result and the fault type, the fault impact range of the energy storage power supply is evaluated to obtain a fault impact assessment report; and based on the fault impact assessment report, an intelligent fault handling suggestion plan is generated.

[0031] Specifically, based on the fault root cause location results and the fault type, the system assesses the scope of the fault impact on the energy storage power supply, generating a fault impact assessment report. Based on this fault impact assessment report, it generates intelligent fault handling recommendations. This process aims to comprehensively understand the fault's impact on the energy storage power supply and the entire vehicle system, and provide targeted remediation guidance. First, after determining the fault's root cause and type, the system uses this information to assess the potential impact of the fault. For example, in an electric vehicle application scenario, if an internal short circuit in a battery module is confirmed to be the root cause of an abnormal temperature rise and is classified as an electrical fault, the system then analyzes the specific impact of the fault on other components, such as the battery management system, charging circuit, and vehicle power output. This includes calculating the reduction in driving range due to decreased battery efficiency and assessing the potential for further damage to other battery cells or triggering broader system protection mechanisms. To achieve this, the system comprehensively considers factors such as the energy storage power supply's design parameters, operating environment conditions, and historical fault data. Through simulation and analysis, the system quantifies the fault's impact, ultimately generating a detailed fault impact assessment report. This report not only describes the direct impact of the fault on the current system, but also predicts its potential long-term consequences, helping decision makers to fully grasp the situation. After clarifying the scope of the fault's impact, the system automatically generates an intelligent fault handling recommendation plan. This plan is based on the various indicators in the fault impact assessment report, combined with industry best practices and technical manuals for specific vehicle models, to propose a series of practical maintenance measures and prevention strategies. For example, in response to the above-mentioned internal short circuit problem of the battery module, the intelligent fault handling recommendation plan may recommend the immediate replacement of the faulty battery cell, while checking whether there is a risk of overheating in adjacent cells, and even recommend optimizing the battery cooling system to prevent similar problems from recurring. This intelligent recommendation based on detailed data analysis can not only effectively solve the current fault problem, but also provide scientific guidance for future maintenance work, ensuring the continued safety and reliability of the vehicle.

[0032] In a specific embodiment, the real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operation status data set includes: Collect multi-dimensional parameters of the battery pack of the energy storage power supply through a preset electrical collection sensor to obtain an original data set of the battery pack; Preprocessing the original data set to obtain a preprocessed electrical parameter set; Performing time sequence status monitoring and analysis on the pre-processed electrical parameter set to obtain an electrical operation status data set.

[0033] Specifically, the real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operating status dataset involves three main steps: first, multi-dimensional parameter acquisition is performed on the energy storage power supply's battery pack using pre-set electrical acquisition sensors to obtain a raw battery pack dataset; second, the raw dataset is pre-processed to obtain a pre-processed electrical parameter set; and finally, the pre-processed electrical parameter set is subjected to time-series state monitoring and analysis to obtain an electrical operating status dataset. To implement these steps, the entire process requires precise design and efficient execution. In the first step, multi-dimensional parameter acquisition is performed on the energy storage power supply's battery pack using pre-set electrical acquisition sensors. This is the foundation for obtaining accurate data. For example, in electric vehicle applications, the energy storage power supply is the vehicle's power core, and its performance directly affects vehicle safety and range. To fully understand the battery pack's operating status, various types of sensors, such as voltage sensors, current sensors, and temperature sensors, are typically installed at key locations in the battery pack. These sensors can capture various key parameters of the battery pack in real time during operation, including but not limited to the voltage value, charge and discharge current, and temperature changes of each battery cell. Take an electric car traveling on a highway as an example. When the vehicle accelerates or climbs a hill, the battery pack undergoes frequent charge and discharge cycles. During this process, the voltage and current of the battery cells fluctuate rapidly, and the temperature may rise due to the high load. This information is recorded by the corresponding sensors, forming a raw data set for the battery pack, providing a rich foundation for subsequent data analysis. The next step is to preprocess the raw data set to obtain a preprocessed electrical parameter set. This step is crucial because it directly affects the accuracy of subsequent analysis. The main goals of preprocessing are to remove noise, fill missing values, and perform data normalization. Given the potential for data loss due to signal interference, sensor failure, or environmental factors in real applications, effective measures must be taken to ensure data quality. For example, in the electric car example mentioned above, the data collected by the sensors may contain noise or be intermittent due to bumps or electromagnetic interference during driving. In this case, filtering techniques can be used to remove noise, and interpolation methods can be used to fill missing data points. Furthermore, to ensure that data from different sources can be compared on a consistent scale, data normalization is also required. This results in a preprocessed electrical parameter set that is not only more accurate but also more suitable for further analysis. The final step involves time-series state monitoring and analysis of the preprocessed electrical parameter set to generate an electrical operating status dataset. This step emphasizes the importance of time series, as the state of energy storage power sources changes dynamically over time. Time-series analysis of preprocessed data provides a deeper understanding of the battery pack's operating mode and potential issues.For example, when an electric vehicle operates for extended periods, the overall health of the battery pack can be assessed by analyzing the voltage fluctuations, current trends, and temperature fluctuations of the battery cells over a period of time. If a battery cell's voltage drops sharply within a short period of time, accompanied by an abnormally high temperature, this could be a warning sign of impending failure. Based on these time-series status monitoring results, the system can promptly identify abnormalities and incorporate them into the electrical operating status dataset, providing a basis for further fault diagnosis. This provides strong support for both routine maintenance and rapid response in emergencies, ensuring the safe and stable operation of the energy storage power supply and the entire vehicle system. In this way, the entire process not only achieves comprehensive monitoring of the energy storage power supply but also provides a solid data foundation for subsequent fault diagnosis and resolution.

[0034] In a specific embodiment, the fault feature extraction of the electrical operation status dataset using the deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group includes: Performing time-frequency domain decomposition processing on the electrical operation status data set using the deep learning algorithm of the remote intelligent diagnosis center to obtain a multi-dimensional electrical feature matrix; Performing nonlinear dynamic characteristic analysis on the energy storage power supply based on the multidimensional electrical characteristic matrix to obtain a fault-related characteristic sequence, and performing adaptive quantization processing on the fault-related characteristic sequence to obtain a fault quantization characteristic set; Performing multi-scale fusion processing on the fault quantization feature set to obtain a fused feature vector, and performing fault correlation analysis based on the fused feature vector to obtain a fault correlation matrix; Performing fault feature dimensionality reduction processing on the fault correlation matrix to obtain a reduced-dimensionality feature group, and performing fault feature optimization clustering on the reduced-dimensionality feature group to obtain a fault feature vector group.

[0035] Specifically, the remote intelligent diagnostic center's deep learning algorithm extracts fault features from the electrical operating status dataset to generate a set of fault feature vectors. This process is the core technical component of the entire intelligent energy storage power supply remote fault diagnosis method. First, the process involves using the remote intelligent diagnostic center's deep learning algorithm to perform time-frequency domain decomposition on the electrical operating status dataset to generate a multidimensional electrical feature matrix. In practical applications, such as electric vehicles, when driving under complex road conditions or in extreme weather conditions, their energy storage power supplies (such as battery packs) experience a series of complex dynamic changes. To capture the details of these changes, the system first needs to conduct in-depth analysis of the electrical operating status dataset collected from the sensors. This is where time-frequency domain decomposition technology becomes crucial, as it converts time series data into a multidimensional feature representation that combines time and frequency information. For example, by performing a fast Fourier transform (FFT) on the current and voltage time series, it can reveal periodic components and non-periodic disturbances hidden within the signals, thereby constructing a multidimensional electrical feature matrix. Next, the energy storage system performs nonlinear dynamic feature analysis based on the multidimensional electrical feature matrix to obtain a fault-related feature sequence. This feature sequence is then adaptively quantized to produce a fault quantization feature set. At this stage, the system further explores the complex interactions between the energy storage system's internal components. For example, in an electric vehicle, if the voltage fluctuation of a battery cell differs significantly from that of other cells, this may indicate a potential fault point. Nonlinear dynamic analysis, such as using a recurrent neural network (RNN) or a long short-term memory network (LSTM), can track the evolution of this fluctuation pattern over time and identify feature sequences associated with the fault. These feature sequences are then adaptively quantized to convert the continuous numerical range into a limited number of discrete levels for easier processing. This approach not only reduces the amount of data but also enhances the model's ability to distinguish between different fault types. The fault quantization feature set is then fused at multiple scales to produce a fused feature vector. Fault correlation analysis is then performed based on this fused feature vector to produce a fault correlation matrix. This process emphasizes the importance of comprehensively assessing the health of the energy storage system from multiple perspectives. Using electric vehicles as an example, in addition to considering the state of each battery cell individually, it is also necessary to understand how they interact as a whole. Multi-scale fusion processing allows information from different levels, such as local cell-level features and global system-level features, to be combined to form a more comprehensive fused feature vector. Fault correlation analysis based on these fused feature vectors can then help determine which components have direct or indirect relationships and how these relationships evolve over time and with changing operating conditions.The resulting fault correlation matrix provides a clear perspective on the interdependencies between components and potential fault propagation paths. The final step is to perform fault feature dimensionality reduction on the fault correlation matrix to obtain a reduced feature group. This reduced feature group is then subjected to fault feature optimization clustering to obtain a set of fault feature vectors. Dimensionality reduction aims to reduce data dimensionality while retaining the most important fault-related information. Common dimensionality reduction techniques include principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE). In the context of electric vehicles, this means selecting the most representative fault features from a multitude of possible fault features to accurately describe the overall health of the energy storage system. After dimensionality reduction, clustering algorithms such as K-means or DBSCAN are then used to optimize clustering of the reduced feature group, grouping fault features based on similarity. This not only helps more accurately locate the specific fault type but also provides a scientific basis for developing targeted maintenance strategies. For example, this approach can quickly determine whether a problem with the battery management system or damage to an individual battery cell is causing damage, allowing appropriate remedial measures to ensure safe and reliable vehicle operation. Through the powerful functions of deep learning algorithms, the entire process achieves the goal of accurately extracting fault characteristics from massive data, providing a solid foundation for subsequent fault root cause location and processing.

[0036] In a specific embodiment, the performing of fault propagation path analysis on the energy storage power source based on the fault feature vector group to obtain a fault propagation link diagram includes: Performing a spatiotemporal correlation analysis on the fault feature vector group to obtain a spatiotemporal feature matrix of fault propagation, and performing a dynamic topological structure analysis on the spatiotemporal feature matrix of fault propagation to obtain a fault propagation topology network; Performing fault propagation dynamics modeling on the energy storage power source based on the fault propagation topology network to obtain a fault propagation dynamics model, and performing multi-level coupling analysis on the fault propagation dynamics model to obtain a fault propagation coupling relationship diagram; performing fault propagation path tracing on the fault propagation coupling relationship graph to obtain a fault propagation path set, and performing propagation strength evaluation on the fault propagation path set to obtain a fault propagation strength matrix; A fault propagation link is constructed for the energy storage power source based on the fault propagation intensity matrix to obtain an initial fault propagation link, and the initial fault propagation link is optimized and reconstructed to obtain a fault propagation link graph.

[0037] Specifically, the fault propagation path analysis of the energy storage power supply based on the fault feature vector group to generate a fault propagation link diagram is a key step in deeply analyzing the potential fault propagation mechanism within the energy storage power supply. First, the process involves performing spatiotemporal correlation analysis on the fault feature vector group to obtain a spatiotemporal fault propagation feature matrix, and then performing dynamic topological analysis on this spatiotemporal fault propagation feature matrix to obtain a fault propagation topological network. In electric vehicle applications, when an abnormality occurs in the vehicle's energy storage power supply (such as a battery pack), such as a voltage fluctuation or temperature rise exceeding the normal range in a battery cell, these abnormalities are not limited to a single point in time but evolve over time and spatial location. Therefore, through spatiotemporal correlation analysis, the system can identify the temporal and spatial distribution patterns of these anomalies, forming a spatiotemporal fault propagation feature matrix containing both time series and spatial location information. Furthermore, using dynamic topological analysis methods, such as constructing a network model that describes the connectivity and changes between different components, it is possible to reveal the interactions and dependencies between the various components within the energy storage power supply, ultimately generating a fault propagation topological network. Based on the fault propagation topology network, the fault propagation dynamics of the energy storage power supply are modeled to obtain a fault propagation dynamics model. This model is then subjected to a multi-level coupling analysis to generate a fault propagation coupling diagram. This step simulates how a fault, starting from an initial point, gradually affects the rest of the energy storage system. In the electric vehicle example described above, if a battery cell degrades due to overheating, this localized failure could trigger a chain reaction, affecting adjacent battery cells and their connected circuit protection devices. To accurately simulate this complex propagation process, a detailed fault propagation dynamics model is required that accounts for multiple factors, including the speed and intensity of the fault, and the response time of the affected components. Subsequently, through multi-level coupling analysis, the system can explore the interactions between components at different levels, such as direct physical connections between battery cells and indirect effects through the power management system. This allows for the generation of a fault propagation coupling diagram, clearly illustrating the fault's propagation path and impact range within the system. Next, the fault propagation paths are traced within the fault propagation coupling diagram to obtain a set of fault propagation paths. The propagation strength of this set of fault propagation paths is then assessed to generate a fault propagation strength matrix. During this process, the system tracks every possible fault propagation path in detail, identifying its starting point, nodes it passes through, and its endpoint. For electric vehicles, this means not only focusing on issues with individual battery cells, but also considering the overall health of the entire battery pack and even the entire vehicle's electrical system. By comprehensively evaluating all possible propagation paths, the fault propagation intensity along each path can be quantified, namely the degree and likelihood of impact to each component along that path.The result of this step is a fault propagation intensity matrix, which provides a quantitative basis for understanding the specific propagation of faults throughout the system. The final step involves constructing a fault propagation chain for the energy storage power supply based on the fault propagation intensity matrix, obtaining an initial fault propagation chain. This initial chain is then optimized and reconstructed to form a fault propagation chain diagram. In this step, the system first constructs a preliminary fault propagation chain based on the fault propagation intensity matrix, clearly showing which components are most susceptible to the fault and how these effects are transmitted. However, this preliminary chain may not be accurate or complete, and therefore requires further optimization and reconstruction. By incorporating more actual operating data and historical fault cases as reference, the initial chain can be adjusted and improved, ensuring that the final fault propagation chain diagram reflects both actual conditions and provides predictive value. For example, in electric vehicle applications, continuously updating and refining the fault propagation chain diagram can help technicians locate problems more quickly and accurately, and take effective preventive measures to ensure safe and stable vehicle operation. In summary, this series of steps not only deepens our understanding of the internal fault propagation mechanisms of the energy storage power supply but also provides a scientific basis for developing targeted maintenance strategies.

[0038] In a specific embodiment, the causal relationship reasoning is performed on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result, including: Performing temporal causal chain decomposition on the fault propagation link graph to obtain a multi-level causal relationship network, and performing dynamic weight calculation on the multi-level causal relationship network to obtain a causal strength matrix, wherein the causal strength matrix includes fault node impact weight, fault propagation temporal weight, fault correlation weight, and fault persistence weight; Performing causal link tracing analysis on the energy storage power supply based on the causal strength matrix to obtain a fault causal propagation sequence, and performing hierarchical decomposition processing on the fault causal propagation sequence to obtain a fault causal hierarchy diagram, wherein the fault causal hierarchy diagram includes first-level fault source characteristics, second-level fault source characteristics, fault transmission characteristics, and fault evolution characteristics; Performing causal correlation measurement on the fault causal hierarchy graph to obtain a causal correlation measurement matrix, and identifying fault source nodes based on the causal correlation measurement matrix to obtain a fault source candidate set; Performing a fault source credibility assessment on the energy storage power supply based on the fault source candidate set to obtain a fault source credibility assessment matrix, and performing a multi-dimensional cross-validation on the fault source credibility assessment matrix to obtain a fault root cause location result; Perform fault type matching analysis on the fault root cause location result to obtain the fault type, wherein the fault type includes battery pack performance degradation fault, battery pack internal short circuit fault, battery pack abnormal temperature fault and battery pack abnormal charging and discharging fault.

[0039] Specifically, causal reasoning is performed on the energy storage power supply based on the fault propagation link graph to obtain the fault root cause location results and the corresponding fault type. This process is a key step in achieving accurate fault diagnosis. First, the fault propagation link graph is decomposed into a temporal causal chain to obtain a multi-level causal network. Dynamic weights are then calculated on this multi-level causal network to generate a causal strength matrix. In electric vehicle applications, when an abnormality occurs in the energy storage power supply, such as an abnormally high battery pack temperature or voltage fluctuations outside the normal range, these phenomena do not exist in isolation but are interconnected through a series of complex causal relationships. Therefore, by performing temporal causal chain decomposition on the fault propagation link graph, the entire fault propagation process can be divided into multiple stages and levels, forming a multi-level causal network. Next, to quantify the importance of each node and path, dynamic weights are calculated, including the fault node impact weight, the fault propagation temporal weight, the fault correlation weight, and the fault persistence weight, ultimately generating a causal strength matrix. This step helps identify the key factors in the fault propagation process. Based on the causal strength matrix, causal chain traceability analysis is performed on the energy storage power supply to obtain a fault causal propagation sequence. This fault causal propagation sequence is then hierarchically decomposed to produce a fault causal hierarchy diagram. This process not only focuses on how the fault propagates from one point to another, but also attempts to understand the logical relationships and temporal order between each fault point. For example, in the electric vehicle example mentioned above, if a battery cell overheating is found to be caused by excessive charging current, which may be due to improper battery management system (BMS) settings, causal chain traceability analysis can be used to identify the specific propagation path from the BMS configuration issue to the battery cell overheating. Furthermore, by performing hierarchical decomposition on the fault causal propagation sequence, a fault causal hierarchy diagram is constructed, which includes first-level fault source characteristics, second-level fault source characteristics, fault transmission characteristics, and fault evolution characteristics. This graphical tool provides technicians with a clear perspective, helping them better understand the fault mechanism and its evolution process. Next, the fault causal hierarchy diagram is causally quantified to obtain a causal correlation quantification matrix. Based on this causal correlation quantification matrix, fault source nodes are identified to obtain a candidate set of fault sources. In this step, the system further evaluates the correlations between each fault node based on the previously obtained causal strength matrix and fault causal hierarchy diagram. By quantifying these correlations, it is possible to more accurately determine which nodes are most likely to be the source of the fault. For example, in the above example, if a specific battery management module is found to frequently appear on multiple fault propagation paths and has a high correlation with other components, then this module is likely the primary cause of the overall system failure. The resulting set of candidate fault sources provides the basis for subsequent in-depth analysis.Based on the candidate set of fault sources, the system performs a fault source credibility assessment on the energy storage power supply to generate a fault source credibility assessment matrix. This credibility assessment matrix is then cross-validated across multiple dimensions to determine the root cause of the fault. During this phase, the system comprehensively evaluates each node in the candidate set, considering its likelihood of being the root cause of the fault. This includes reviewing historical data, simulating responses under different scenarios, and analyzing interactions with other components. This multi-dimensional cross-validation ensures that the final root cause of the fault is highly reliable. For example, in the actual application of electric vehicles, if a detailed analysis confirms that a battery management module has a design flaw, and this flaw is closely related to repeatedly observed battery overheating, it can be identified as the true root cause of the fault. The final step is to perform a fault type matching analysis on the root source location results to determine the specific fault type, such as battery pack performance degradation, internal short circuit, abnormal battery pack temperature, or abnormal battery pack charge and discharge. Once the root cause of the fault is determined, it is classified according to its specific manifestations. For example, if a battery cell is found to have degraded performance due to long-term use and is unable to maintain normal charge and discharge efficiency, this can be classified as a battery pack performance degradation failure; if a manufacturing defect causes an internal short circuit in the battery, this should be considered an internal battery pack short circuit failure. In this way, not only can the root cause of the fault be accurately located, but its specific type can also be determined, providing a scientific basis for developing targeted repair plans. In short, this series of steps together constitutes a comprehensive and systematic fault diagnosis process, effectively improving the efficiency and accuracy of energy storage power supply maintenance work.

[0040] In a specific embodiment, the fault impact range assessment of the energy storage power supply is performed based on the fault root location result and the fault type to obtain a fault impact assessment report, including: Performing fault diffusion dynamic modeling analysis on the fault root cause location result to obtain a fault diffusion feature matrix, and performing multi-dimensional quantization processing on the fault diffusion feature matrix to obtain a fault diffusion quantization set; Tracing the fault-affected links of the energy storage power source based on the fault diffusion quantization set to obtain a fault-affected propagation network, and performing hierarchical decomposition processing on the fault-affected propagation network to obtain a fault-affected hierarchical map; Performing a fault loss assessment analysis on the fault impact hierarchy map to obtain a fault loss assessment matrix, and performing multi-dimensional risk quantification based on the fault loss assessment matrix to obtain a fault risk quantification set; Based on the fault risk quantification set, a fault impact assessment is performed on the energy storage power supply to obtain a fault impact assessment report.

[0041] Specifically, based on the fault root location results and the fault type, the fault impact scope of the energy storage power supply is assessed to generate a fault impact assessment report. This process is crucial for ensuring a comprehensive understanding of the potential impact of the fault on the energy storage system and the entire vehicle system, and for providing a scientific basis for subsequent maintenance and repair work. First, the fault root location results are subjected to a dynamic fault diffusion modeling analysis to generate a fault diffusion characteristic matrix. This fault diffusion characteristic matrix is then subjected to multi-dimensional quantization to produce a fault diffusion quantization set. In electric vehicle applications, once a battery cell overheating due to an internal short circuit is identified as the root cause of the fault and classified as an internal short circuit fault within the battery pack, the next step is to simulate how this fault propagates from the source to surrounding components. By establishing a dynamic fault diffusion model, the potential fault impact scope and evolutionary trends under different conditions can be predicted. For example, in a high-temperature environment or when the vehicle is traveling at high speed for extended periods, this short circuit could cause adjacent battery cells to overheat, potentially triggering the battery management system (BMS) to implement protective measures and limit current input. To more accurately describe these diffusion patterns, the model output must be converted into a fault diffusion feature matrix containing information in multiple dimensions, such as time, space, and intensity. These features are then quantified to form a fault diffusion quantization set. Based on this fault diffusion quantization set, the fault impact links of the energy storage power supply are traced to obtain a fault impact propagation network. This network is then hierarchically decomposed to produce a fault impact hierarchy map. This step emphasizes the importance of understanding the fault propagation path and its hierarchical structure throughout the system from a holistic perspective. In the above example, if an internal short circuit occurs in a battery cell, it not only affects its own performance but can also trigger a chain reaction, affecting other connected battery cells, power electronics components, and the entire vehicle control system. By tracing the fault impact links, a detailed fault impact propagation network can be constructed, demonstrating the interactions between various components. Next, by hierarchically decomposing the propagation paths in this network, the complex impact relationships can be simplified into multiple levels, each representing a specific type of fault propagation mode or degree of impact, ultimately forming a fault impact hierarchy map. The benefit of this approach is that it not only clearly demonstrates how a failure propagates from one point to the entire system, but also identifies which components are most vulnerable to the failure and the dependencies between them. This is followed by a failure loss assessment analysis of the failure impact hierarchy map to generate a failure loss assessment matrix. Based on this failure loss assessment matrix, multi-dimensional risk quantification is performed to obtain a failure risk quantification set. In this step, the system considers not only the extent of physical damage but also a comprehensive assessment of factors such as the resulting economic losses, safety risks, and the impact on user experience.For electric vehicles, if a critical battery cell fails due to a short circuit, it could directly lead to a significant reduction in driving range or even, in extreme cases, cause serious consequences such as fire. Therefore, a detailed assessment of each possible failure consequence is necessary, quantifying it into specific numerical indicators to form a failure loss assessment matrix. By introducing multi-dimensional risk quantification methods, such as probability analysis and cost-benefit analysis, these loss assessment results can be further refined to generate a failure risk quantification set encompassing multiple risk factors. This step provides quantitative support for developing effective response strategies. The final step is to assess the impact of the failure on the energy storage power supply based on this failure risk quantification set, generating a failure impact assessment report. This step integrates the results of all previous analyses to provide a comprehensive report that comprehensively reflects the scope and extent of the failure impact. In the case of electric vehicles, this report not only identifies the battery cell experiencing an internal short circuit but also details the potential subsequent issues that may arise from the failure, such as the impact on other battery cells, interference with the battery management system, and threats to vehicle safety. The report also includes preventative measures and remediation recommendations for these issues, such as replacing the damaged battery cell, optimizing the battery cooling system design, or upgrading the BMS software. In short, this series of meticulous analysis steps can not only help technicians quickly and accurately locate faults and assess their impact, but also provide strong technical support to ensure the safe and stable operation of vehicles.

[0042] In a specific embodiment, the fault impact assessment of the energy storage power source is performed based on the fault risk quantification set to obtain a fault impact assessment report, including: Performing multi-dimensional feature decomposition on the fault risk quantification set to obtain a fault risk feature vector group, and performing time series correlation analysis on the fault risk feature vector group to obtain a fault risk time series correlation matrix; quantifying the degree of fault impact of the energy storage power supply based on the fault risk time series correlation matrix to obtain a quantified set of fault impact degrees, and performing multi-level decomposition processing on the quantified set of fault impact degrees to obtain a fault impact level feature graph; Tracing the fault impact propagation path of the fault impact hierarchical feature graph to obtain a fault impact propagation link set, and performing fault impact intensity assessment based on the fault impact propagation link set to obtain a fault impact intensity assessment matrix; Defining the fault impact range of the energy storage power supply based on the fault impact intensity assessment matrix to obtain a fault impact range boundary map; Performing a comprehensive assessment of the degree of fault impact of the energy storage power supply based on the fault impact range boundary map to obtain a fault impact degree assessment set, and performing fault risk level classification based on the fault impact degree assessment set to obtain a fault risk level matrix; A fault impact assessment report is generated for the energy storage power supply based on the fault risk level matrix to obtain a fault impact assessment report.

[0043] Specifically, the fault impact assessment of the energy storage power supply is performed based on the fault risk quantification set, resulting in a fault impact assessment report. This process aims to comprehensively assess the scope and extent of the fault's impact on the energy storage system and provide a scientific basis for subsequent maintenance and repair work. First, the fault risk quantification set is subjected to multi-dimensional feature decomposition to obtain a set of fault risk feature vectors. These fault risk feature vectors are then subjected to time series correlation analysis to generate a fault risk time series correlation matrix. In an electric vehicle application scenario, suppose a battery cell overheats due to an internal short circuit. This failure could affect not only the cell itself but also the entire battery pack and even the safe operation of the vehicle. To better understand these impacts, the data in the fault risk quantification set must be decomposed according to multiple dimensions (such as time, space, and intensity) to form a set of fault risk feature vectors that can describe different aspects of the impact. Next, through time series correlation analysis, the temporal trends of these features and their interrelationships can be identified, thereby constructing a fault risk time series correlation matrix containing time series information. For example, if an abnormal temperature rise is observed in a battery cell during a specific time period, followed by similar events in adjacent cells, temporal correlation analysis can be used to determine the temporal order and speed of this propagation. Based on the fault risk temporal correlation matrix, the fault impact of the energy storage power supply is quantified to generate a quantified set of fault impacts. This set is then subjected to a multi-level decomposition process to create a fault impact hierarchy feature map. During this process, the system not only focuses on the speed of fault propagation but also quantifies the impact at each stage. Continuing with the example of an electric vehicle, after confirming that a battery cell is overheating due to an internal short circuit, further analysis is needed to determine how this fault gradually spreads to other components. By quantifying the fault risk temporal correlation matrix, the specific extent of damage to each affected component can be determined and summarized as a quantified set of fault impacts. Next, through multi-level decomposition, these quantified results are categorized and organized according to impact levels, forming a fault impact hierarchy feature map that clearly illustrates the scope and extent of the fault's impact. This step helps technicians intuitively understand how the fault propagates from its initial point to each layer and the extent of the impact at each layer. Next, the fault impact propagation path is traced within the fault impact hierarchy feature graph to obtain a set of fault impact propagation links. Based on this set of links, the fault impact intensity is assessed to produce a fault impact intensity assessment matrix. In this step, the system traces each possible fault propagation path in detail, determining its starting point, the nodes it passes through, and its end point, and assessing the fault impact intensity along each path.For example, in the aforementioned case, if an overheating problem in a battery cell is found to not only affect adjacent cells but also trigger the battery management system (BMS)'s protection mechanism, limiting current input to avoid further risk, fault propagation path tracing can be used to identify the specific propagation paths from the initial fault point to the final affected component. Based on these paths, the impact intensity is further assessed, namely, the degree and likelihood of impact on each component, generating a fault impact intensity assessment matrix. The results of this step provide important quantitative support for formulating effective response strategies. Next, the fault impact scope of the energy storage power supply is defined based on the fault impact intensity assessment matrix to generate a fault impact scope boundary map. Having determined the primary fault propagation paths and impact intensity through the previous analysis steps, the precise boundaries of the fault impact need to be determined. For example, in the electric vehicle example, if an internal short circuit in a battery cell is confirmed to primarily affect several directly connected cells without affecting other components, a fault impact scope boundary map can be constructed based on this information. This graphical tool helps technicians quickly locate the problem area and take targeted measures for repair or isolation. The final step is to conduct a comprehensive fault impact assessment of the energy storage power supply based on the fault impact range boundary map, generating a fault impact assessment set. This set is then used to categorize fault risks, resulting in a fault risk matrix. A fault impact assessment report is then generated for the energy storage power supply based on the fault risk matrix, ultimately resulting in a fault impact assessment report. In this step, the system integrates the results of all previous analyses to perform a comprehensive fault impact assessment. By combining the various indicators in the fault impact assessment set, a quantitative score can be assigned to the overall fault impact, and the fault risk can be classified into different levels based on the score. For example, in the above example, if the internal short circuit in the battery cell is confirmed to be localized and has not caused serious consequences, it can be classified as a low risk level. Conversely, if the fault could potentially trigger a chain reaction, threatening the safety of the entire vehicle, it should be classified as a high risk level. Based on these risk levels, a detailed fault impact assessment report is generated, which not only identifies the specific location and type of the fault but also provides corresponding repair recommendations and preventive measures to ensure the safe and stable operation of the energy storage system. Together, these steps form a systematic fault impact assessment process, laying a solid foundation for ensuring the safety and reliability of electric vehicles and other equipment.

[0044] The above describes the remote fault diagnosis method of the intelligent energy storage power supply in the embodiment of the present invention. The following describes the remote fault diagnosis system of the intelligent energy storage power supply in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a remote fault diagnosis system for an intelligent energy storage power supply includes: A monitoring module 21 is used to monitor the electrical parameters of the energy storage power supply in real time to obtain an electrical operation status data set; The judgment module 22 is configured to send the electrical operation status data set to a remote intelligent diagnosis center when abnormal fluctuations occur in the electrical operation status data set; An extraction module 23 is configured to extract fault features from the electrical operation status data set using a deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group; An analysis module 24 is configured to perform a fault propagation path analysis on the energy storage power source based on the fault feature vector group to obtain a fault propagation link diagram; An inference module 25 is configured to perform causal relationship inference on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result; The evaluation module 26 is configured to evaluate the fault impact range of the energy storage power supply based on the fault root cause location result and the fault type to obtain a fault impact assessment report; and generate an intelligent fault handling suggestion plan based on the fault impact assessment report.

[0045] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0046] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, memory, display screen, input device, network interface and database connected via a system bus. The processor of the computer design 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 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 via a network connection. When the computer program is executed by the processor, the above method is implemented.

[0047] 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.

[0048] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-described method when executed by a processor. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0049] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media provided herein 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. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM.

[0050] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0051] 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 description 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 remote fault diagnosis method for an intelligent energy storage power supply, characterized in that: The method is applied to a vehicle provided with an energy storage power supply, and comprises the following steps: Performing real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operation status data set; When there is abnormal fluctuation in the electrical operation status data set, the electrical operation status data set is sent to a remote intelligent diagnosis center; Extracting fault features from the electrical operation status data set using a deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group; Performing fault propagation path analysis on the energy storage power supply based on the fault feature vector group to obtain a fault propagation link diagram; Performing causal reasoning on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result; Based on the fault root cause location result and the fault type, the fault impact range of the energy storage power supply is evaluated to obtain a fault impact assessment report; and based on the fault impact assessment report, an intelligent fault handling suggestion plan is generated.

2. The remote fault diagnosis method of the intelligent energy storage power supply according to claim 1, characterized in that: The real-time monitoring of electrical parameters of the energy storage power supply to obtain an electrical operation status data set includes: Collect multi-dimensional parameters of the battery pack of the energy storage power supply through a preset electrical collection sensor to obtain an original data set of the battery pack; Preprocessing the original data set to obtain a preprocessed electrical parameter set; Performing time sequence status monitoring and analysis on the pre-processed electrical parameter set to obtain an electrical operation status data set.

3. The remote fault diagnosis method of the intelligent energy storage power supply according to claim 1, characterized in that: The extracting fault features from the electrical operation status data set by the deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group includes: Performing time-frequency domain decomposition processing on the electrical operation status data set using the deep learning algorithm of the remote intelligent diagnosis center to obtain a multi-dimensional electrical feature matrix; Performing nonlinear dynamic characteristic analysis on the energy storage power supply based on the multidimensional electrical characteristic matrix to obtain a fault-related characteristic sequence, and performing adaptive quantization processing on the fault-related characteristic sequence to obtain a fault quantization characteristic set; Performing multi-scale fusion processing on the fault quantization feature set to obtain a fused feature vector, and performing fault correlation analysis based on the fused feature vector to obtain a fault correlation matrix; Performing fault feature dimensionality reduction processing on the fault correlation matrix to obtain a reduced-dimensionality feature group, and performing fault feature optimization clustering on the reduced-dimensionality feature group to obtain a fault feature vector group.

4. The remote fault diagnosis method of the intelligent energy storage power supply according to claim 1, characterized in that: The performing fault propagation path analysis on the energy storage power supply based on the fault feature vector group to obtain a fault propagation link diagram includes: Performing a spatiotemporal correlation analysis on the fault feature vector group to obtain a spatiotemporal feature matrix of fault propagation, and performing a dynamic topological structure analysis on the spatiotemporal feature matrix of fault propagation to obtain a fault propagation topology network; Performing fault propagation dynamics modeling on the energy storage power source based on the fault propagation topology network to obtain a fault propagation dynamics model, and performing multi-level coupling analysis on the fault propagation dynamics model to obtain a fault propagation coupling relationship diagram; performing fault propagation path tracing on the fault propagation coupling relationship graph to obtain a fault propagation path set, and performing propagation strength evaluation on the fault propagation path set to obtain a fault propagation strength matrix; A fault propagation link is constructed for the energy storage power source based on the fault propagation intensity matrix to obtain an initial fault propagation link, and the initial fault propagation link is optimized and reconstructed to obtain a fault propagation link graph.

5. The remote fault diagnosis method of the intelligent energy storage power supply according to claim 1, characterized in that: The performing causal reasoning on the energy storage power supply based on the fault propagation link diagram to obtain a fault root cause location result and a fault type corresponding to the fault root cause location result includes: Performing temporal causal chain decomposition on the fault propagation link graph to obtain a multi-level causal relationship network, and performing dynamic weight calculation on the multi-level causal relationship network to obtain a causal strength matrix, wherein the causal strength matrix includes fault node impact weight, fault propagation temporal weight, fault correlation weight, and fault persistence weight; Performing causal link tracing analysis on the energy storage power supply based on the causal strength matrix to obtain a fault causal propagation sequence, and performing hierarchical decomposition processing on the fault causal propagation sequence to obtain a fault causal hierarchy diagram, wherein the fault causal hierarchy diagram includes first-level fault source characteristics, second-level fault source characteristics, fault transmission characteristics, and fault evolution characteristics; Performing causal correlation measurement on the fault causal hierarchy graph to obtain a causal correlation measurement matrix, and identifying fault source nodes based on the causal correlation measurement matrix to obtain a fault source candidate set; Performing a fault source credibility assessment on the energy storage power supply based on the fault source candidate set to obtain a fault source credibility assessment matrix, and performing a multi-dimensional cross-validation on the fault source credibility assessment matrix to obtain a fault root cause location result; Perform fault type matching analysis on the fault root cause location result to obtain the fault type, wherein the fault type includes battery pack performance degradation fault, battery pack internal short circuit fault, battery pack abnormal temperature fault and battery pack abnormal charging and discharging fault.

6. The remote fault diagnosis method of the intelligent energy storage power supply according to claim 1, characterized in that: The fault impact range assessment of the energy storage power supply is performed based on the fault root source location result and the fault type to obtain a fault impact assessment report, including: Performing fault diffusion dynamic modeling analysis on the fault root cause location result to obtain a fault diffusion feature matrix, and performing multi-dimensional quantization processing on the fault diffusion feature matrix to obtain a fault diffusion quantization set; Tracing the fault-affected links of the energy storage power source based on the fault diffusion quantization set to obtain a fault-affected propagation network, and performing hierarchical decomposition processing on the fault-affected propagation network to obtain a fault-affected hierarchical map; Performing a fault loss assessment analysis on the fault impact hierarchy map to obtain a fault loss assessment matrix, and performing multi-dimensional risk quantification based on the fault loss assessment matrix to obtain a fault risk quantification set; Based on the fault risk quantification set, a fault impact assessment is performed on the energy storage power supply to obtain a fault impact assessment report.

7. The remote fault diagnosis method of the intelligent energy storage power supply according to claim 6, characterized in that: The fault impact assessment of the energy storage power supply is performed based on the fault risk quantification set to obtain a fault impact assessment report, including: Performing multi-dimensional feature decomposition on the fault risk quantification set to obtain a fault risk feature vector group, and performing time series correlation analysis on the fault risk feature vector group to obtain a fault risk time series correlation matrix; quantifying the degree of fault impact of the energy storage power supply based on the fault risk time series correlation matrix to obtain a quantified set of fault impact degrees, and performing multi-level decomposition processing on the quantified set of fault impact degrees to obtain a fault impact level feature graph; Tracing the fault impact propagation path of the fault impact hierarchical feature graph to obtain a fault impact propagation link set, and performing fault impact intensity assessment based on the fault impact propagation link set to obtain a fault impact intensity assessment matrix; Defining the fault impact range of the energy storage power supply based on the fault impact intensity assessment matrix to obtain a fault impact range boundary map; Performing a comprehensive assessment of the degree of fault impact of the energy storage power supply based on the fault impact range boundary map to obtain a fault impact degree assessment set, and performing fault risk level classification based on the fault impact degree assessment set to obtain a fault risk level matrix; A fault impact assessment report is generated for the energy storage power supply based on the fault risk level matrix to obtain a fault impact assessment report.

8. A remote fault diagnosis system for an intelligent energy storage power supply, characterized in that: Applied to vehicles, including: A monitoring module, configured to monitor electrical parameters of the energy storage power supply in real time and obtain an electrical operation status data set; a judgment module, configured to send the electrical operation status data set to a remote intelligent diagnosis center when abnormal fluctuations occur in the electrical operation status data set; An extraction module, configured to extract fault features from the electrical operation status data set using a deep learning algorithm of the remote intelligent diagnosis center to obtain a fault feature vector group; an analysis module, configured to perform fault propagation path analysis on the energy storage power supply based on the fault feature vector group to obtain a fault propagation link diagram; An inference module, configured to perform causal reasoning on the energy storage power supply based on the fault propagation link diagram, and obtain a fault root cause location result and a fault type corresponding to the fault root cause location result; An evaluation module is used to evaluate the fault impact range of the energy storage power supply based on the fault root cause location result and the fault type to obtain a fault impact assessment report; and to generate an intelligent fault handling suggestion plan based on the fault impact assessment report.

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