Fault diagnosis method, device and equipment for acquisition terminal of source load storage and charging system

Abstract

The invention provides a fault diagnosis method, device and equipment for an acquisition terminal of a source load storage and charging system, and relates to the technical field of fault diagnosis. The method comprises the following steps: based on a plurality of edge computing devices, performing data analysis on real-time operation data of each acquisition terminal in a target source load storage and charging system to obtain a plurality of pieces of first alarm information; according to a dominant fault identification model, performing fault analysis on the real-time acquisition data to obtain second alarm information of each acquisition terminal; and third alarm information is obtained based on the real-time collected data, the first alarm information and the hidden fault identification model, fault collection terminals are determined and fault diagnosis is carried out in combination with the second alarm information, and the fault type and the repair scheme of each fault collection terminal are obtained, so that each fault collection terminal is repaired according to the corresponding repair scheme. According to the method, the dominant fault and the hidden fault can be combined, the complex fault is accurately judged and automatically repaired, and the operation stability of the source load storage and charging system is improved.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and in particular to a fault diagnosis method, apparatus and equipment for a data acquisition terminal of a source-load-storage-charging system. Background Technology

[0002] The data acquisition terminal is the core hub for real-time interaction in the integrated energy source-load-storage-charging system, spanning the entire process from power supply and load consumption to energy storage regulation and charging services. It bears the heavy responsibility of collecting critical data such as voltage, current, power, and remaining power from distributed photovoltaic systems, energy storage battery clusters, charging piles, and various loads at millisecond levels. The operational stability of the data acquisition terminal affects the energy source-load-storage-charging system's ability to achieve peak shaving and valley filling, and local consumption of new energy, thus forming the foundation for maintaining the efficient operation of the entire energy ecosystem.

[0003] In existing technologies, fault diagnosis and recovery of data acquisition terminals mainly rely on a combination of manual inspection and offline testing. During routine maintenance, staff need to go to the site at fixed intervals to make a preliminary judgment on the terminal's status based on its operation. In addition, when the terminal issues a fault alarm or the system malfunctions, maintenance personnel need to bring specialized instruments such as multimeters and oscilloscopes to the site to offline test the operating parameters of components such as the CPU, memory, and communication modules, and to investigate the root cause of the problem by combining historical fault cases.

[0004] However, when using offline detection methods to detect faults in the power generation, storage, and charging system, it is impossible to capture operational anomalies in real time, resulting in a lag in detection results. In addition, offline detection methods rely on the experience of maintenance personnel to determine faults, making it difficult to accurately identify complex faults and affecting the stability of the power generation, storage, and charging system. Summary of the Invention

[0005] This invention provides a fault diagnosis method, apparatus, and equipment for a data acquisition terminal of a power-to-energy storage and charging system, in order to solve the problem that offline detection methods have a lag in detection results and are difficult to accurately identify complex faults, thus affecting the stability of the power-to-energy storage and charging system.

[0006] In a first aspect, embodiments of the present invention provide a fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system, comprising: Based on multiple pre-deployed edge computing devices, the real-time operating data of each acquisition terminal in the target source-load storage and charging system is analyzed to obtain the first alarm information of each acquisition terminal. Based on the pre-trained explicit fault identification model, fault analysis is performed on the real-time data collected by each acquisition terminal to obtain the second alarm information of each acquisition terminal. Based on the real-time data collected by each acquisition terminal, the first alarm information, and the pre-trained hidden fault identification model, the third alarm information of each acquisition terminal is obtained. Based on the second and third alarm information of each acquisition terminal, the faulty acquisition terminal is identified from all acquisition terminals. Fault diagnosis is performed on each faulty acquisition terminal to obtain the fault type and repair plan for each faulty acquisition terminal, so that each faulty acquisition terminal can be repaired according to the corresponding repair plan.

[0007] Secondly, embodiments of the present invention provide a fault diagnosis device for a data acquisition terminal of a source-load-storage-charging system, comprising: The analysis module is used to perform data analysis on the real-time operating data of each acquisition terminal in the target source-load storage and charging system based on multiple pre-deployed edge computing devices, and obtain the first alarm information of each acquisition terminal. The identification module is used to perform fault analysis on the real-time data collected by each acquisition terminal based on a pre-trained explicit fault identification model, and obtain the second alarm information of each acquisition terminal. The identification module is also used to obtain the third alarm information of each acquisition terminal based on the real-time acquisition data of each acquisition terminal, the first alarm information and the pre-trained hidden fault identification model. The diagnostic module is used to identify faulty acquisition terminals from all acquisition terminals based on the second and third alarm information of each acquisition terminal, perform fault diagnosis on each faulty acquisition terminal, obtain the fault type and repair plan for each faulty acquisition terminal, so that each faulty acquisition terminal can be repaired according to the corresponding repair plan.

[0008] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0009] In this embodiment of the invention, pre-deployed edge computing devices analyze the real-time operating data of each acquisition terminal within the target source-load-storage-charging system to obtain the first alarm information for each acquisition terminal. Then, based on a visible fault identification model, fault analysis is performed on the real-time acquisition data of each acquisition terminal to obtain the second alarm information. Furthermore, by combining the real-time acquisition data, the first alarm information, and the hidden fault identification model, a third alarm information for each acquisition terminal is obtained. This allows for detection based on real-time data, ensuring timely detection results. By combining real-time acquisition data, real-time operating data, and the hidden fault identification model, the third alarm information becomes more accurate and comprehensive. The second and third alarm information are used to determine the faulty acquisition terminal, fault type, and repair plan. The acquisition terminal is then controlled to automatically repair according to the fault repair plan. This approach, combining visible and hidden faults, accurately identifies complex faults and performs automatic repair, improving the stability of the source-load-storage-charging system. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the implementation of a fault diagnosis method for a data acquisition terminal in a source-load-storage-charging system provided in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the implementation of step S110 of the fault diagnosis method for the acquisition terminal of the source-load-storage-charging system provided in this embodiment of the invention. Figure 3 This is a flowchart illustrating the implementation of step S130 of the fault diagnosis method for the acquisition terminal of the source-load-storage-charging system provided in this embodiment of the invention. Figure 4 This is a schematic diagram of the structure of a fault diagnosis device for a data acquisition terminal of a source-load-storage-charging system provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0011] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0012] See Figure 1 The document illustrates a flowchart of the fault diagnosis method for a data acquisition terminal in a source-load-storage-charging system provided by an embodiment of the present invention, detailed below: Step S110: Based on multiple pre-deployed edge computing devices, perform data analysis on the real-time operating data of each acquisition terminal in the target source-load storage and charging system to obtain the first alarm information of each acquisition terminal.

[0013] In some embodiments, edge computing devices are small computing devices deployed near data acquisition terminals. Their core function is to process data locally, reducing data transmission latency. For example, in a park-based energy source-load-storage-charging system, the local computing module near the photovoltaic panel data acquisition sensors and charging pile data terminals is a typical edge computing device, capable of quickly processing operational data from surrounding data acquisition terminals. A target energy source-load-storage-charging system refers to an integrated energy system encompassing energy supply (source), energy consumption (load), energy storage regulation (storage), and charging services (charging), and is the target of fault diagnosis. For example, a system composed of distributed photovoltaic power supply devices, residential electricity loads, centralized energy storage battery packs, and public charging piles in a residential community constitutes a target energy source-load-storage-charging system. Data acquisition terminals are core devices that run throughout the entire energy source-load-storage-charging system, responsible for real-time acquisition of key energy data, and are the basic units for data acquisition. For example, sensors monitoring the voltage of energy storage battery clusters, terminals recording the charging power of charging piles, and devices collecting load current all belong to data acquisition terminals.

[0014] It's important to note that real-time operational data refers to various data generated by the acquisition terminal during its own operation, reflecting its working status. This excludes system energy-related data and focuses solely on the terminal itself. For example, the terminal's operating temperature, communication signal strength, data transmission success rate, and internal component operating voltage are all examples of real-time operational data. Data analysis involves filtering, processing, and analyzing the collected data to uncover anomalies, forming the basis for generating alarm information. For instance, comparing the acquisition terminal's real-time operational data with normal operating standards to filter out data exceeding reasonable ranges and determine if an anomaly exists is a typical data analysis operation. The first alarm message is generated by the edge computing device after analyzing the acquisition terminal's real-time operational data, indicating a potential anomaly in the terminal's operation. For example, when the edge computing device detects that a acquisition terminal's operating temperature is consistently higher than the normal range, the generated terminal temperature anomaly warning is the first alarm message.

[0015] See Figure 2 The specific processing method of the above step S110 includes steps S1102-S1102, the specific content of which is as follows: Step S1101: Calculate the geographical distance between each acquisition terminal in the target source-load storage and charging system and each pre-deployed edge computing device.

[0016] In some embodiments, geographical distance refers to the actual straight-line distance between a data acquisition terminal within the target source-load-storage-charging system and each pre-deployed edge computing device in physical space, directly reflecting the degree of spatial correlation between the two. For example, in a source-load-storage-charging system in an industrial park, a data acquisition terminal is installed in the central area of ​​a photovoltaic array, and an edge computing device is installed in a control box next to the photovoltaic array. The straight-line distance (e.g., 20 meters) between the two from the installation point of the data acquisition terminal to the control box of the edge device is their geographical distance.

[0017] Step S1102: For each data acquisition terminal, the real-time operating data of the data acquisition terminal is analyzed using the edge computing device that is geographically closest to the data acquisition terminal to obtain the first alarm information for each data acquisition terminal.

[0018] In some embodiments, the geographically closest edge computing device refers to the edge computing device with the smallest geographical distance to a specific data acquisition terminal among all pre-deployed edge computing devices. Its core advantage is that it can receive and process the real-time operating data of the data acquisition terminal nearby, minimizing data transmission latency and loss. For example, a source-load-storage-charging system in a commercial park deploys a total of 5 edge computing devices (distributed at the east, south, west, and north entrances of the park and the central computer room). The data acquisition terminal next to the energy storage battery on the west side of the park is closer to the edge computing device at the west entrance (15 meters) than the other 4 devices (all more than 50 meters). This edge computing device at the west entrance is the geographically closest edge computing device to the data acquisition terminal.

[0019] In one possible implementation, step S1102 is specifically processed as follows: using the edge computing device geographically closest to the acquisition terminal, alarm operation data exceeding the operation threshold is extracted from the real-time operation data of the acquisition terminal; for each alarm operation data, the minimum difference between the alarm operation data and the corresponding operation threshold is calculated, and the alarm event and alarm level corresponding to the alarm operation data are determined based on the minimum difference.

[0020] In some embodiments, the operating threshold is a preset reasonable range boundary value for the real-time operating data of the acquisition terminal. It is used to determine whether the data is in a normal operating state and is the core standard for distinguishing normal data from abnormal data. These thresholds are usually determined based on the hardware performance, working environment requirements, and system operating specifications of the acquisition terminal. Different types of real-time operating data correspond to different operating thresholds. Alarm operating data is abnormal data that exceeds the corresponding operating threshold and is filtered out from the real-time operating data of the acquisition terminal. This type of data directly reflects that the acquisition terminal may have an operational anomaly and is the basic data source for generating alarm information. For example, if the operating temperature threshold of a certain acquisition terminal is 5℃-40℃, when a temperature data of 45℃ appears in its real-time operating data, 45℃ is alarm operating data. In addition, when the communication signal strength is lower than the minimum effective communication signal strength value, the corresponding signal strength data also belongs to alarm operating data. The minimum difference refers to the minimum numerical difference between a single alarm operating data and the operating threshold corresponding to that data. It is used to quantify the degree of abnormality of the alarm operating data. The larger the difference, the further the data deviates from the normal range. It's important to note that if the operating threshold includes both an upper and lower limit, it's necessary to first determine whether the alarm operating data exceeds the upper limit or falls below the lower limit before calculating the difference between that data and the corresponding boundary threshold. This difference is the minimum difference. For example, if the operating threshold for the acquisition terminal voltage is 18V-24V, and a certain alarm operating data is 26V, then the minimum difference is 26V-24V. An alarm event is a specific abnormal event name determined based on the type of alarm operating data and its corresponding abnormal behavior. It clearly identifies the type of problem existing in the acquisition terminal, allowing maintenance personnel to quickly understand the core content of the anomaly. Different types of alarm operating data correspond to different alarm events, and different abnormal directions of the same type of data (such as exceeding the upper limit or falling below the lower limit) may also correspond to different alarm events. The alarm level is a severity level of the anomaly classified according to the size of the minimum difference between the alarm operating data and the operating threshold. It is used to distinguish the urgency of the alarm event and provide a basis for subsequent processing priorities. Generally, the larger the minimum difference, the higher the alarm level, and the higher the priority it needs to be handled. Common alarm levels can be divided into minor alarms, general alarms, and serious alarms. The specific classification criteria need to be determined based on the system's tolerance for anomalies and the scope of the fault's impact.

[0021] Step S120: Based on the pre-trained explicit fault identification model, perform fault analysis on the real-time data collected by each acquisition terminal to obtain the second alarm information for each acquisition terminal.

[0022] In some embodiments, an explicit fault identification model is an algorithmic model specifically designed to identify obvious and easily perceptible faults in a source-load-storage-charging system. These faults are typically accompanied by significant data anomalies. For example, a model that can directly identify intuitive faults such as data acquisition interruptions, sudden drops in charging power, and severe voltage exceedances is an explicit fault identification model. Real-time acquired data refers to the core energy data related to the source-load-storage-charging system collected in real time by the acquisition terminal, focusing on the system's energy operating status rather than the terminal itself. For example, the real-time output power of photovoltaic panels, the remaining capacity of energy storage batteries, the real-time charging current of charging piles, and the power consumption of the load are all examples of real-time acquired data. Fault analysis involves targeted processing of real-time acquired data to determine whether the data reflects a fault in the system or equipment, and the approximate scope of the fault. For example, analyzing the drastic fluctuations in the output voltage of charging piles in real-time acquired data to determine that there may be an energy transmission fault in the charging pile is a process of fault analysis. The second alarm message is an alarm prompt about obvious system faults generated by the explicit fault identification model after analyzing the real-time data collected by the acquisition terminal. For example, when the explicit fault identification model finds that the remaining power of the energy storage battery drops sharply in a short period of time, exceeding the normal range, the generated prompt of abnormal drop in energy storage battery power is the second alarm message.

[0023] In one possible implementation, step S120 is specifically processed as follows: for each acquisition terminal, the following steps are performed: extract abnormal acquisition data from the real-time acquisition data of the acquisition terminal; input the abnormal acquisition data into a pre-trained explicit fault identification model to obtain the second alarm information of the acquisition terminal.

[0024] In some embodiments, abnormal acquisition data refers to energy-related data that is filtered from the real-time acquisition data of the acquisition terminal and does not conform to the normal operating rules or preset reasonable range of the source-load-storage-charging system. This type of data directly reflects potential explicit problems in the system's energy supply, load consumption, energy storage regulation, or charging service, and is the core basis for subsequent identification of explicit faults. For example, if the real-time output power collected by the photovoltaic panel acquisition terminal suddenly drops to near zero during a period of sufficient sunlight, while the power should normally remain at a high level during that period, this power data is abnormal acquisition data. When training a pre-trained explicit fault identification model, the first step is to collect a large amount of real-time data from the acquisition terminals in the source-load-storage-charging system. This data covers data under normal operating conditions and abnormal data corresponding to various explicit faults, and these data are labeled with clear fault type tags. Next, the collected data is preprocessed, including cleaning noisy data, correcting missing values, and standardizing the data format. The processed data is then divided into training and validation sets. Subsequently, the training set data is input into the deep learning model for iterative training. During this process, the model parameters are continuously adjusted to optimize the model's ability to capture explicit fault features. Finally, the model performance is tested using the validation set data to determine whether the model's accuracy and false positive rate in identifying explicit faults meet the preset standards. If the standards are not met, the steps of data supplementation, parameter adjustment, or model structure optimization are repeated until the model performance meets the fault diagnosis requirements, and the training is completed.

[0025] Step S130: Based on the real-time data collected by each acquisition terminal, the first alarm information, and the pre-trained hidden fault identification model, the third alarm information of each acquisition terminal is obtained.

[0026] In some embodiments, the latent fault identification model is an algorithmic model specifically designed to identify hidden, difficult-to-detect potential faults in a source-load-storage-charging system. These faults do not exhibit significant, obvious anomalies and require multi-dimensional data analysis for judgment. For example, a model capable of identifying potential problems such as a slight increase in data transmission latency, slow accumulation of component wear, and a subtle decline in data processing efficiency is a latent fault identification model. The third alarm information is generated by combining real-time data collected from multiple acquisition terminals and the first alarm information, after analysis by the latent fault identification model, to provide alarm information about potential and hidden faults.

[0027] See Figure 3 The specific processing method of step S130 above includes steps S1301-S1304, and the specific content is as follows: Step S1301: Using a preset correlation calculation formula, calculate the correlation coefficient between the real-time data collected by each acquisition terminal and the real-time data collected by other acquisition terminals, and determine the correlation coefficient as the first correlation between each acquisition terminal and other acquisition terminals.

[0028] In some embodiments, the preset correlation calculation formula is a pre-defined mathematical formula used to quantify the degree of correlation between real-time data collected by two acquisition terminals. Its core function is to derive a numerical value reflecting the correlation between the changes in the two data points through data calculation. Common examples include the Pearson correlation coefficient formula and the Spearman rank correlation formula. The correlation coefficient is a specific value obtained by calculating the real-time data collected by the two acquisition terminals using the preset correlation calculation formula. The magnitude of this value directly reflects the tightness of the correlation between the two sets of data. The closer the value is to 1, the stronger the correlation; the closer the value is to 0, the weaker the correlation. For example, the correlation coefficient between the power data of the photovoltaic acquisition terminal and the charging power data of the energy storage acquisition terminal is calculated to be 0.8, indicating that the changes in the two data points are highly positively correlated, that is, when the photovoltaic power increases, the energy storage charging power also tends to increase.

[0029] In some embodiments, the data items collected by each acquisition terminal and other acquisition terminals may be the same or different. When different data items exist in every two acquisition terminals, some of the different data items may be related, while others may not be related. When determining the correlation calculation formula, the correlation of different data items that are not related to each acquisition terminal and other acquisition terminals is 0. However, for the same data items and different data items that are related, their differences should be taken into account to determine the correlation calculation formula. The preset correlation calculation formula is:

[0030] in, For the first The first data acquisition terminal and the first The correlation coefficient of each data acquisition terminal. For the first The first data acquisition terminal and the first The number of identical data collection items from each data acquisition terminal. For the first The first data acquisition terminal The average value of the data collected for each data item within one hour. For the first The first data acquisition terminal The real-time value of each data collection item within one hour. For the first The first data acquisition terminal The average value of the data collected for each data item within one hour. For the first The first data acquisition terminal The real-time value of each data collection item within one hour. For the first The first data acquisition terminal and the first The number of different but related data collection items in each collection terminal for and The sum of, For the first The first data acquisition terminal and the first The consistency of the first data acquisition item of each acquisition terminal; if consistent, then... If the value is 1, then... It is 0.

[0031] It should be noted that the first correlation is an indicator used to describe the degree of correlation between the real-time data collected by the two acquisition terminals, which is directly determined by the correlation coefficient between them. It is based solely on the correlation calculation result of the raw data and has not yet been corrected by incorporating any fault alarm information. It serves as the basis for subsequent correlation adjustments. For example, if the correlation coefficient between the real-time charging current data of the charging pile acquisition terminal and the real-time electricity consumption data of the load acquisition terminal is calculated to be 0.6, this correlation of 0.6 is directly defined as the first correlation between the two, without considering whether there is any second alarm information between the two terminals.

[0032] Step S1302: Based on the second alarm information of each acquisition terminal, adjust the first correlation between the acquisition terminal and other acquisition terminals to obtain the second correlation between the acquisition terminal and other acquisition terminals.

[0033] In some embodiments, adjusting the first correlation refers to correcting the determined first correlation based on the second alarm information of the acquisition terminal. The purpose is to make the corrected correlation more closely reflect the actual fault scenario. If a terminal has a significant fault, its data correlation may be affected by the fault, causing deviations. Adjustment can reduce the interference of such deviations on correlation judgment. For example, if an energy storage acquisition terminal has a second alarm message indicating an abnormal drop in power, it means that its real-time acquired data may be distorted due to the fault. In this case, the first correlation between the terminal and the photovoltaic acquisition terminal needs to be appropriately lowered based on this alarm message to avoid inaccurate correlation judgment due to distorted data. The second correlation is a more accurate data correlation index between acquisition terminals obtained after adjusting the first correlation. It integrates the first correlation of the original data and the second alarm information, and can more realistically reflect the actual correlation between the two terminals under fault scenarios. For example, if the first correlation between the photovoltaic acquisition terminal and the energy storage acquisition terminal is 0.8, and the energy storage terminal has a second alarm message, the correlation is corrected to 0.5 after adjustment. This 0.5 corresponds to the second correlation between the two, which better reflects the actual data correlation state when the energy storage terminal fails.

[0034] In one possible implementation, step S1302 specifically includes: extracting first alarm events and corresponding alarm levels for each first alarm event from the first alarm information of the first acquisition terminal; wherein the first acquisition terminal is any acquisition terminal; extracting second alarm events and corresponding alarm levels for each second alarm event from the first alarm information of the second acquisition terminal; wherein the second acquisition terminal is any acquisition terminal other than the first acquisition terminal; if there is a first alarm event with an alarm level exceeding a preset level in the first acquisition terminal, or a second alarm event with an alarm level exceeding a preset level in the second acquisition terminal, then the connection between the first acquisition terminal and the second acquisition terminal is established. The credibility of the first correlation is determined as the first credibility. If there is no first alarm event with an alarm level exceeding the preset level in the first acquisition terminal, and there is no second alarm event with an alarm level exceeding the preset level in the second acquisition terminal, then the credibility between the first acquisition terminal and the second acquisition terminal is calculated based on the first alarm event of the first acquisition terminal, the alarm level corresponding to each first alarm event, the second alarm event of the second acquisition terminal, the alarm level corresponding to each second alarm event, and the preset credibility calculation formula. The product of the credibility between the first acquisition terminal and the second acquisition terminal and the first correlation between the first acquisition terminal and the second acquisition terminal is determined as the second correlation between the first acquisition terminal and the second acquisition terminal.

[0035] In some embodiments, the first acquisition terminal is any one of the selected acquisition terminals when analyzing the correlation between two acquisition terminals, serving as one of the benchmark objects for correlation analysis. It can be arbitrarily selected from all acquisition terminals in the source-load-storage-charging system according to actual analysis needs. The first alarm event is the event content extracted from the first alarm information of the first acquisition terminal, specifically reflecting the abnormal operation of that terminal. Each first alarm information contains one or more such events, each corresponding to a specific operational anomaly point of the terminal. For example, if the first alarm information of the first acquisition terminal is "slightly high charging module temperature" and "extended data upload interval," then "slightly high charging module temperature" and "extended data upload interval" are the two first alarm events of that terminal. The second acquisition terminal is any other acquisition terminal selected besides the first acquisition terminal when analyzing the correlation between two acquisition terminals. It forms a paired analysis relationship with the first acquisition terminal to explore the impact of alarms on the correlation between different terminals. The second alarm event is the event content extracted from the first alarm information of the second acquisition terminal, specifically reflecting the abnormal operation of that terminal. Both the second and first alarm events are essentially specific manifestations of terminal operational anomalies. The preset severity level is a pre-set threshold for the alarm level, used to determine whether an alarm will significantly interfere with the reliability of the primary correlation between the data collection terminals. It is typically set to an alarm level that may cause data distortion at the terminals. The reliability of the primary correlation refers to the reliability of the primary correlation between the first and second data collection terminals in a real-world scenario. It measures whether the original correlated data is distorted due to terminal alarms; the more severe the alarm, the lower the reliability. For example, if the primary correlation between the first and second data collection terminals is 0.8, but the first data collection terminal has a high-level alarm, it indicates that the correlation may be unreliable due to data distortion, and its reliability needs to be reassessed based on the alarm situation. The primary reliability is a fixed reliability value set for the primary correlation between the two. This value is determined in advance based on the general rules of data distortion in this type of scenario and does not change with specific alarm details. For example, it can be pre-set that if either terminal has an alarm exceeding the preset level, the reliability of the primary correlation is 0.3. Therefore, when either the first or second data collection terminal has an alarm exceeding the preset level, the reliability of the primary correlation between the two is determined to be 0.3, i.e., the primary reliability.

[0036] It should be noted that the preset confidence calculation formula is a pre-designed mathematical formula used to quantify the confidence of the primary correlation when neither of the two acquisition terminals has alarms exceeding the preset level, by combining the first alarm events and corresponding alarm levels of both terminals. The formula typically assigns different weights to alarm levels of varying severity to reflect the different impacts of alarms on correlation. The preset confidence calculation formula is: (Weighted sum of alarm levels for the first acquisition terminal + Weighted sum of alarm levels for the second acquisition terminal) / Total weight limit. Confidence is the numerical value of the reliability of the primary correlation calculated using the preset confidence calculation formula when neither the first nor the second acquisition terminal has alarms exceeding the preset level. This value varies with the number and severity of alarm events on both terminals; the milder and fewer the alarms, the higher the confidence. For example, if the first acquisition terminal has only slight data delay and the second acquisition terminal has only general signal fluctuations, the confidence of the primary correlation calculated using the preset formula is 0.6, and 0.6 is the confidence level in this case.

[0037] Step S1303: Using each acquisition terminal as a node and the second correlation between each acquisition terminal and other acquisition terminals as edges, construct a graph structure to obtain the acquisition terminal graph structure of the source-load-storage-charge system.

[0038] In some embodiments, a node is a basic unit in a graph structure representing a single data acquisition terminal within a source-load-storage-charging system. Each data acquisition terminal corresponds to an independent node in the graph structure, and different terminals can be distinguished by labels (such as terminal number or terminal type). For example, when constructing the graph structure, photovoltaic panel data acquisition terminal 1, energy storage battery data acquisition terminal A, and charging pile data acquisition terminal 3, etc., in a community, each correspond to a node in the graph, clearly identifying the identity of each data acquisition terminal through the nodes. An edge is an element in the graph structure used to connect two nodes (data acquisition terminals) and carry secondary correlation information between them. The core function of an edge is to intuitively reflect the correlation strength between different data acquisition terminals. Usually, the magnitude of the secondary correlation can be reflected by the attributes of the edge, such as numerical labels and line thickness. The higher the correlation, the larger the label value of the edge or the thicker the line. For example, in the graph structure, the line segment connecting the photovoltaic data acquisition terminal node and the load data acquisition terminal node is an edge, and the label 0.7 on the edge represents the secondary correlation between the two, intuitively showing the degree of correlation between the two terminals. Graph structure construction refers to the process of building a network structure that intuitively displays the relationships between all data acquisition terminals, using data acquisition terminals as nodes and secondary correlations between terminals as edges. The constructed graph structure clearly presents the relationship network between terminals within the system, providing structured data support for subsequent latent fault identification. For example, if 10 data acquisition terminals in an industrial park's energy storage and charging system are each treated as 10 nodes, and each node is connected to the other with edges of corresponding attributes based on its secondary correlation with other nodes, the resulting network containing nodes, edges, and correlation information is the result of graph structure construction. The data acquisition terminal graph structure of an energy storage and charging system is a graph structure constructed for a specific target energy storage and charging system that fully reflects all data acquisition terminals and their relationships within the system. Its nodes cover all data acquisition terminals within the system, and its edges cover the secondary correlations between all terminals. It is one of the core input data for latent fault identification models. For example, a commercial complex's energy storage and charging system contains 15 data acquisition terminals across four categories: photovoltaic, energy storage, charging piles, and commercial loads. The network structure built using these 15 terminals as nodes and the secondary correlations between each terminal as edges is the system's dedicated data acquisition terminal graph structure.

[0039] Step S1304: Input the data acquisition terminal diagram structure of the source-load-storage-charging system into the latent fault identification model to obtain the third alarm information of each data acquisition terminal.

[0040] It should be noted that the training process of the latent fault identification model includes: First, real-time data from multiple acquisition terminals within the source-load-storage-charging system needs to be collected. This real-time data needs to cover the correlation data of each terminal under normal operating conditions, as well as the correlation data of terminals under latent fault scenarios, such as abnormal correlation data between terminals caused by component latent aging or slight communication delays. The data is then labeled with the corresponding latent fault type or normal state label. Next, the data is preprocessed. In addition to routine noise removal and missing value completion, the correlation coefficients between terminals need to be calculated. Based on this correlation data, a large number of acquisition terminal graph structure samples containing both normal and latent fault scenarios are constructed, and the samples are divided into training and validation sets. Finally, the appropriate graph structure data is selected for computation. The model framework can be built by selecting graph neural networks (GNNs) and their derivatives such as graph neural networks (GCNs) and graph analytic networks (GATs) that can effectively capture the correlation features between terminal nodes. Then, graph structure samples from the training set are input into the model for iterative training. During this process, model parameters such as the learning rate and graph convolutional layer weights are continuously adjusted to optimize the model's ability to capture abnormal graph structure features such as sudden drops in local terminal correlation and deviations in node correlation patterns corresponding to latent faults. Finally, graph structure samples from the validation set are used to test the model's performance, determining whether the model's accuracy and false negative rate in identifying latent faults meet the preset standards. If the standards are not met, latent fault data is supplemented, the graph structure construction method is optimized, or the model structure is adjusted until the model can stably and accurately identify latent faults, thus completing the training.

[0041] Step S140: Based on the second and third alarm information of each acquisition terminal, identify the faulty acquisition terminal from all acquisition terminals, perform fault diagnosis on each faulty acquisition terminal, obtain the fault type and repair plan for each faulty acquisition terminal, so that each faulty acquisition terminal can be repaired according to the corresponding repair plan.

[0042] It should be noted that after receiving the first alarm information from each data acquisition terminal, this information needs to be sent to the cloud platform. The cloud platform then retrieves the second and third alarm information from each terminal, performs fault diagnosis, and determines a repair plan. After obtaining the repair plan, it needs to be sent from the cloud platform to pre-deployed edge computing devices, and then distributed to each data acquisition terminal via these devices.

[0043] In some embodiments, a fault acquisition terminal is a terminal whose existence or potential fault is determined after comprehensively analyzing the contents of the second and third alarm information. Fault diagnosis is a process of detailed analysis of the faulty acquisition terminal to clarify the specific cause and manifestation of the fault, providing a basis for subsequent determination of the fault type and repair plan. For example, determining that the fault is caused by poor contact of the communication module by analyzing the alarm information and operation records of the faulty acquisition terminal is the process of fault diagnosis. Fault type is a specific classification of the fault of the faulty acquisition terminal according to its cause, manifestation, and scope of impact, clarifying the core attributes of the fault. For example, faults may be classified into different types such as communication faults, power supply faults, data acquisition faults, and internal component aging faults. Repair plan is an executable solution developed for a specific fault type, used to guide the faulty acquisition terminal to restore normal operation. For example, a solution for communication faults, such as restarting the communication module and checking the connection lines, and a solution for power supply faults, such as replacing aging power supply components, are both repair plans.

[0044] In one possible implementation, step S140 is specifically processed as follows: For each data acquisition terminal, the following steps are performed: Extract the third alarm event and the alarm level of each third alarm event from the second alarm information and the third alarm information of the data acquisition terminal; Divide all the third alarm events of the data acquisition terminal according to the alarm type, and calculate the first sum of the alarm levels of all the third alarm events in each alarm type; When the first sum of the first alarm types of the data acquisition terminal exceeds a preset value, the data acquisition terminal is identified as a faulty data acquisition terminal; wherein, the first alarm type is any alarm type of the data acquisition terminal.

[0045] In some embodiments, a third alarm event is an event extracted from both the second and third alarm information of the acquisition terminal, reflecting a potential fault-related event of the terminal. It encompasses specific abnormal manifestations corresponding to both explicit and implicit faults, and forms the basis for subsequent alarm classification by type. If the second alarm information of a acquisition terminal includes a sudden drop in charging power and the third alarm information includes a latent aging of the charging module, then these two specific abnormal manifestations constitute the terminal's third alarm events. The alarm level of each third alarm event refers to its severity. Alarm type is a classification label for all third alarm events based on their problem attributes or impact scope. The purpose is to categorize scattered alarm events by commonalities, facilitating centralized assessment of the severity of a particular type of fault. Common alarm types include communication, power, temperature, and component aging, with multiple third alarm events sharing the same attributes within the same type. The first sum is the sum of the alarm levels of all third-level alarm events under a specific alarm type. It is used to quantify the overall severity of alarms of that type. The larger the sum, the more significant the cumulative impact of alarms under that type. For example, a temperature alarm on a data acquisition terminal contains two third-level alarm events: the terminal casing temperature is too high (alarm level 2) and the internal chip temperature is slightly high (alarm level 1). The sum of these two alarm levels is the first sum for the temperature alarm.

[0046] It should be noted that the preset value is a threshold pre-set for the first sum of each alarm type. It is used to determine whether the cumulative severity of the alarm of that type has reached the standard requiring the data acquisition terminal to be classified as a faulty terminal. Different preset values ​​can be set for different alarm types according to their impact on the system. For example, considering that power alarms have a significant impact on energy transmission in the source-load-storage-charging system, a higher preset value is set for their first sum; temperature alarms have a relatively mild impact, so a lower preset value is set. When the first sum of a certain type exceeds the corresponding preset value, it indicates that the alarm of that type is sufficiently severe. The first alarm type is any one of all alarm types included in the data acquisition terminal. When determining a faulty data acquisition terminal, each alarm type must be checked one by one. If the first sum of any alarm type exceeds the preset value, the terminal will be determined as a faulty data acquisition terminal. For example, a certain data acquisition terminal has three types of alarms: communication alarms, power alarms, and temperature alarms. During the evaluation, the communication alarm is first checked as the first alarm type to see if its first sum value exceeds the preset value. Then, the power alarm and temperature alarm are checked as the first alarm types in turn. If any type exceeds the standard, a fault can be determined.

[0047] In one possible implementation, step S140 further includes the following steps for each fault acquisition terminal: inputting the second alarm information and the third alarm information of the fault acquisition terminal into a pre-trained fault diagnosis model to obtain the fault type of the fault acquisition terminal; extracting the corresponding repair scheme from the repair scheme database, and determining the repair scheme as the repair scheme for the fault acquisition terminal, so that the fault acquisition terminal can be repaired according to the repair scheme.

[0048] In some embodiments, the fault diagnosis model is a pre-trained algorithm model specifically designed to receive second and third alarm information from the fault acquisition terminal and, by analyzing the characteristics of these alarms, output a model indicating the specific fault cause category of the terminal. Its core function is to transform scattered alarm information into a clear fault root cause judgment. For example, if the model is input with a sudden drop in charging power and latent aging of the charging module of a fault acquisition terminal, the model can analyze and output a fault caused by charging module aging leading to abnormal power output; this model is the fault diagnosis model. The fault type refers to the category of the specific fault cause of the fault acquisition terminal output by the fault diagnosis model. For example, fault types include communication module poor contact leading to data transmission interruption, loose power supply lines leading to unstable power supply, and battery cell degradation leading to abnormal energy storage. The repair solution database pre-stores the association between various fault types and corresponding executable repair measures; that is, each fault type is matched with verified specific operational steps that can resolve the fault, facilitating quick searching for suitable repair solutions. For example, the database might store repair steps for a data transmission interruption fault caused by poor communication module contact: turning off the terminal power → re-plugging and unplugging the communication interface → turning on the power and checking the data transmission status. For a power supply instability fault caused by a loose power supply line: checking the power supply line connectors after power off → retightening the loose connectors → testing the power supply stability. It should be noted that each fault type may have multiple repair solutions. When multiple repair solutions exist for each fault type, one can be randomly selected, or the lowest-cost repair solution can be chosen based on the cost of each option.

[0049] By using pre-deployed edge computing devices, real-time operational data from each acquisition terminal within the target power-energy storage and charging system is analyzed to obtain the first alarm information for each acquisition terminal. This allows for rapid analysis of real-time operational data, improving data processing efficiency and reducing the data processing load on the cloud platform. Then, based on a visible fault identification model, fault analysis is performed on the real-time acquired data from each acquisition terminal to obtain the second alarm information. Combining real-time acquired data, the first alarm information, and the hidden fault identification model, a third alarm information for each acquisition terminal is obtained. This enables detection based on real-time data, ensuring timely detection results. By combining real-time acquired data, real-time operational data, and the hidden fault identification model, the third alarm information becomes more accurate and comprehensive. The second and third alarm information are used to determine the faulty acquisition terminal and its repair plan, and the acquisition terminal is controlled to automatically repair according to the repair plan. This approach, combining visible and hidden faults, accurately identifies complex faults and performs automatic repairs, improving the stability of the power-energy storage and charging system.

[0050] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0051] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0052] Figure 4 The diagram shows a fault diagnosis device for a data acquisition terminal of a source-load-storage-charging system according to an embodiment of the present invention. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below: like Figure 4 As shown, the fault diagnosis device 4 for the data acquisition terminal of the source-load-storage-charging system includes: Analysis module 41 is used to perform data analysis on the real-time operating data of each acquisition terminal in the target source-load storage and charging system based on multiple pre-deployed edge computing devices, and obtain the first alarm information of each acquisition terminal; The identification module 42 is used to perform fault analysis on the real-time data collected by each acquisition terminal based on the pre-trained explicit fault identification model, and obtain the second alarm information of each acquisition terminal. The identification module 42 is also used to obtain the third alarm information of each acquisition terminal based on the real-time acquisition data of each acquisition terminal, the first alarm information and the pre-trained hidden fault identification model. The diagnostic module 43 is used to identify faulty acquisition terminals from all acquisition terminals based on the second and third alarm information of each acquisition terminal, perform fault diagnosis on each faulty acquisition terminal, obtain the fault type and repair plan for each faulty acquisition terminal, so that each faulty acquisition terminal can be repaired according to the corresponding repair plan.

[0053] In one possible implementation, the analysis module 41 is specifically used to: calculate the geographical distance between each acquisition terminal in the target source-load storage and charging system and each pre-deployed edge computing device; for each acquisition terminal, use the edge computing device with the closest geographical distance to the acquisition terminal to perform data analysis on the real-time operating data of the acquisition terminal, and obtain the first alarm information of each acquisition terminal.

[0054] In one possible implementation, the analysis module 41 is further configured to: extract alarm operation data exceeding the operation threshold from the real-time operation data of the acquisition terminal using the edge computing device geographically closest to the acquisition terminal; calculate the minimum difference between the alarm operation data and the corresponding operation threshold for each alarm operation data, and determine the alarm event and alarm level corresponding to the alarm operation data based on the minimum difference.

[0055] In one possible implementation, the identification module 42 is specifically used to: for each acquisition terminal, perform the following steps: extract abnormal acquisition data from the real-time acquisition data of the acquisition terminal; input the abnormal acquisition data into a pre-trained explicit fault identification model to obtain the second alarm information of the acquisition terminal.

[0056] In one possible implementation, the identification module 42 is further configured to: calculate the correlation coefficient between the real-time data collected by each acquisition terminal and the real-time data collected by other acquisition terminals using a preset correlation calculation formula, and determine the correlation coefficient as the first correlation between each acquisition terminal and other acquisition terminals; adjust the first correlation between the acquisition terminal and other acquisition terminals based on the second alarm information of each acquisition terminal to obtain the second correlation between the acquisition terminal and other acquisition terminals; construct a graph structure with each acquisition terminal as a node and the second correlation between each acquisition terminal and other acquisition terminals as an edge to obtain the acquisition terminal graph structure of the source-load-storage-charging system; and input the acquisition terminal graph structure of the source-load-storage-charging system into the latent fault identification model to obtain the third alarm information of each acquisition terminal.

[0057] In one possible implementation, the identification module 42 is further configured to: extract first alarm events and corresponding alarm levels of each first alarm event from the first alarm information of the first acquisition terminal; wherein the first acquisition terminal is any acquisition terminal; extract second alarm events and corresponding alarm levels of each second alarm event from the first alarm information of the second acquisition terminal; wherein the second acquisition terminal is any acquisition terminal other than the first acquisition terminal; if there is a first alarm event in the first acquisition terminal with an alarm level exceeding a preset level, or there is a second alarm event in the second acquisition terminal with an alarm level exceeding a preset level, then the credibility of the first correlation between the first acquisition terminal and the second acquisition terminal is determined as the first credibility; If there are no first alarm events with alarm levels exceeding the preset level in the first data acquisition terminal, and there are no second alarm events with alarm levels exceeding the preset level in the second data acquisition terminal, then the reliability between the first data acquisition terminal and the second data acquisition terminal is calculated based on the first alarm events of the first data acquisition terminal, the alarm level corresponding to each first alarm event, the second alarm events of the second data acquisition terminal, the alarm level corresponding to each second alarm event, and the preset reliability calculation formula; the product of the reliability between the first data acquisition terminal and the second data acquisition terminal and the first correlation between the first data acquisition terminal and the second data acquisition terminal is determined as the second correlation between the first data acquisition terminal and the second data acquisition terminal.

[0058] In one possible implementation, the diagnostic module 43 is specifically used to: for each acquisition terminal, perform the following steps: extract the third alarm event and the alarm level of each third alarm event from the second alarm information and the third alarm information of the acquisition terminal; divide all the third alarm events of the acquisition terminal according to the alarm type, and calculate the first sum value of the alarm levels of all the third alarm events in each alarm type; when the first sum value of the first alarm type of the acquisition terminal exceeds the preset value, the acquisition terminal is identified as a faulty acquisition terminal; wherein, the first alarm type is any alarm type of the acquisition terminal.

[0059] In one possible implementation, the diagnostic module 43 is further configured to: for each fault acquisition terminal, perform the following steps: input the second alarm information and the third alarm information of the fault acquisition terminal into a pre-trained fault diagnosis model to obtain the fault type of the fault acquisition terminal; extract the corresponding repair scheme from the repair scheme database, and determine the repair scheme as the repair scheme for the fault acquisition terminal, so that the fault acquisition terminal can be repaired according to the repair scheme.

[0060] Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. For example... Figure 5As shown, the electronic device 5 of this embodiment includes a processor 50 and a memory 51. The memory 51 stores a computer program 52. When the processor 50 executes the computer program 52, it implements the steps in the various method embodiments described above. Alternatively, when the processor 50 executes the computer program 52, it implements the functions of each module / unit in the various device embodiments described above.

[0061] For example, computer program 52 may be divided into one or more modules / units, which are stored in memory 51 and executed by processor 50 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 52 in electronic device 5.

[0062] Electronic device 5 may include, but is not limited to, processor 50 and memory 51. Those skilled in the art will understand that... Figure 5 This is merely an example of electronic device 5 and does not constitute a limitation on electronic device 5. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 5 may also include input / output devices, network access devices, buses, etc.

[0063] For the sake of simplicity and clarity, only the above-described functional modules / units are used as examples. In practical applications, the functions described above can be assigned to different functional modules / units as needed. These modules / units can be implemented in hardware, software, or a combination of both.

[0064] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0065] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A fault diagnosis method for a data acquisition terminal in a source-load-storage-charging system, characterized in that, include: Based on multiple pre-deployed edge computing devices, the real-time operating data of each acquisition terminal in the target source-load storage and charging system is analyzed to obtain the first alarm information of each acquisition terminal. Based on the pre-trained explicit fault identification model, fault analysis is performed on the real-time data collected by each acquisition terminal to obtain the second alarm information of each acquisition terminal. Based on the real-time data collected by each acquisition terminal, the first alarm information, and the pre-trained hidden fault identification model, the third alarm information of each acquisition terminal is obtained. Based on the second and third alarm information of each acquisition terminal, the faulty acquisition terminal is identified from all acquisition terminals. Fault diagnosis is performed on each faulty acquisition terminal to obtain the fault type and repair plan for each faulty acquisition terminal, so that each faulty acquisition terminal can be repaired according to the corresponding repair plan.

2. The fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system according to claim 1, characterized in that, The third alarm information for each acquisition terminal is obtained based on the real-time data collected by each acquisition terminal, the second alarm information, and the pre-trained latent fault identification model, including: Using a preset correlation calculation formula, the correlation coefficient between the real-time data collected by each acquisition terminal and the real-time data collected by other acquisition terminals is calculated, and the correlation coefficient is determined as the first correlation between each acquisition terminal and other acquisition terminals. Based on the second alarm information of each data acquisition terminal, the first correlation between the data acquisition terminal and other data acquisition terminals is adjusted to obtain the second correlation between the data acquisition terminal and other data acquisition terminals. Using each acquisition terminal as a node and the second correlation between each acquisition terminal and other acquisition terminals as edges, a graph structure is constructed to obtain the acquisition terminal graph structure of the source-load-storage-charging system. The data acquisition terminal diagram of the source-load-storage-charging system is input into the latent fault identification model to obtain the third alarm information of each data acquisition terminal.

3. The fault diagnosis method for the acquisition terminal of the source-load-storage-charging system according to claim 2, characterized in that, The step of adjusting the first correlation between the acquisition terminal and other acquisition terminals based on the first alarm information of each acquisition terminal to obtain the second correlation between the acquisition terminal and other acquisition terminals includes: From the first alarm information of the first acquisition terminal, extract the first alarm event of the first acquisition terminal and the alarm level corresponding to each first alarm event; wherein, the first acquisition terminal is any acquisition terminal; Extract the second alarm events and the corresponding alarm levels of the second alarm events from the first alarm information of the second acquisition terminal; wherein, the second acquisition terminal is any acquisition terminal other than the first acquisition terminal; If there is a first alarm event in the first acquisition terminal with an alarm level exceeding a preset level, or a second alarm event in the second acquisition terminal with an alarm level exceeding a preset level, then the credibility of the first correlation between the first acquisition terminal and the second acquisition terminal is determined as the first credibility. If there is no first alarm event in the first acquisition terminal with an alarm level exceeding the preset level, and there is no second alarm event in the second acquisition terminal with an alarm level exceeding the preset level, then the credibility between the first acquisition terminal and the second acquisition terminal is calculated based on the first alarm event of the first acquisition terminal, the alarm level corresponding to each first alarm event, the second alarm event of the second acquisition terminal, the alarm level corresponding to each second alarm event, and the preset credibility calculation formula. The product of the reliability between the first acquisition terminal and the second acquisition terminal and the first correlation between the first acquisition terminal and the second acquisition terminal is determined as the second correlation between the first acquisition terminal and the second acquisition terminal.

4. The fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system according to claim 1, characterized in that, The method of identifying faulty data acquisition terminals from all data acquisition terminals based on the second and third alarm information of each data acquisition terminal includes: For each data acquisition terminal, perform the following steps: Extract the third alarm event and the alarm level of each third alarm event from the second and third alarm information of the data acquisition terminal. All third alarm events of the acquisition terminal are divided according to alarm type, and the first sum of alarm levels of all third alarm events in each alarm type is calculated; When the sum of the first alarm types of the data acquisition terminal exceeds a preset value, the data acquisition terminal is identified as a faulty data acquisition terminal; wherein, the first alarm type is any alarm type of the data acquisition terminal.

5. The fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system according to claim 4, characterized in that, The process of diagnosing faults in each fault acquisition terminal to obtain the fault type and repair plan for each terminal includes: For each fault acquisition terminal, perform the following steps: The second and third alarm information of the fault acquisition terminal are input into the pre-trained fault diagnosis model to obtain the fault type of the fault acquisition terminal. The corresponding repair scheme for the fault type is extracted from the repair scheme database, and the repair scheme is determined as the repair scheme for the fault acquisition terminal, so that the fault acquisition terminal can be repaired according to the repair scheme.

6. The fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system according to claim 1, characterized in that, The method involves analyzing the real-time operational data of each acquisition terminal within the target source-load storage and charging system based on multiple pre-deployed edge computing devices to obtain the first alarm information for each acquisition terminal, including: Calculate the geographical distance between each acquisition terminal in the target source-load storage and charging system and each pre-deployed edge computing device; For each data acquisition terminal, the real-time operating data of the terminal is analyzed using the edge computing device that is geographically closest to the terminal to obtain the first alarm information for each terminal.

7. The fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system according to claim 6, characterized in that, For each data acquisition terminal, the edge computing device geographically closest to that terminal is used to analyze its real-time operating data to obtain the first alarm information for each terminal, including: By using the edge computing device that is geographically closest to the acquisition terminal, alarm operation data that exceeds the operation threshold is extracted from the real-time operation data of the acquisition terminal; For each alarm operation data, calculate the minimum difference between the alarm operation data and the corresponding operation threshold, and determine the alarm event and alarm level corresponding to the alarm operation data based on the minimum difference.

8. The fault diagnosis method for a data acquisition terminal of a source-load-storage-charging system according to claim 1, characterized in that, The method involves performing fault analysis on the real-time data collected by each acquisition terminal based on a pre-trained explicit fault identification model to obtain the second alarm information for each acquisition terminal, including: For each data acquisition terminal, perform the following steps: Extract abnormal data from the real-time data collected by the acquisition terminal; The abnormal data is input into a pre-trained explicit fault identification model to obtain the second alarm information of the acquisition terminal.

9. A fault diagnosis device for a data acquisition terminal in a source-load-storage-charging system, characterized in that, include: The analysis module is used to perform data analysis on the real-time operating data of each acquisition terminal in the target source-load storage and charging system based on multiple pre-deployed edge computing devices, and obtain the first alarm information of each acquisition terminal. The identification module is used to perform fault analysis on the real-time data collected by each acquisition terminal based on a pre-trained explicit fault identification model, and obtain the second alarm information of each acquisition terminal. The identification module is also used to obtain the third alarm information of each acquisition terminal based on the real-time acquisition data of each acquisition terminal, the first alarm information and the pre-trained hidden fault identification model. The diagnostic module is used to identify faulty acquisition terminals from all acquisition terminals based on the second and third alarm information of each acquisition terminal, perform fault diagnosis on each faulty acquisition terminal, obtain the fault type and repair plan for each faulty acquisition terminal, so that each faulty acquisition terminal can be repaired according to the corresponding repair plan.

10. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 8.

Images