Fault processing method and device for power grid equipment, equipment, medium and product

Abstract

The embodiment of the invention provides a fault processing method and device for power grid equipment, equipment, a medium and a product, and is applied to the technical field of power grid equipment. The method comprises the following steps: in response to a power grid equipment fault processing request, collecting multi-source data corresponding to power grid equipment in a preset time period; performing data processing based on the multi-source data to obtain fault tree data; inputting the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result; inputting the root cause analysis result into a pre-trained defect elimination model to obtain a defect processing strategy; performing resource allocation analysis based on the defect processing strategy and the multi-source data to obtain a defect processing scheme; and pushing the defect processing scheme to the client terminal, and indicating the client terminal to execute the defect processing scheme. The technical effect of improving the operation fault processing efficiency is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of power grid equipment, and in particular to a method, apparatus, equipment, medium, and product for handling faults of power grid equipment. Background Art

[0002] As modern power grids expand in scale and complexity, safe production management of power systems has become increasingly important. The operational stability and reliability of power grid equipment are directly related to the safety and efficiency of the entire power system. However, during operation, power grid equipment is inevitably affected by various factors such as equipment aging, external environmental interference, operational errors, and management loopholes, leading to failures and potential safety hazards. Therefore, timely resolution of power grid equipment failures is essential.

[0003] In the prior art, the fault handling method for power grid equipment is to collect operating data corresponding to the fault of the power grid equipment, determine the fault type of the power grid equipment by statistical analysis or machine learning analysis, and thus generate a corresponding solution strategy.

[0004] Since the prior art mainly implements fault detection and processing of power grid equipment through fault classification, the prior art has a technical problem of low fault processing efficiency. Summary of the Invention

[0005] The embodiments of the present application provide a fault handling method, apparatus, equipment, medium, and product for power grid equipment to achieve the technical effect of improving fault handling efficiency.

[0006] In a first aspect, an embodiment of the present application provides a method for handling a fault of a power grid device, comprising:

[0007] In response to a power grid equipment fault processing request, collecting multi-source data corresponding to the power grid equipment within a preset time period;

[0008] Perform data processing based on multi-source data to obtain fault tree data;

[0009] Input the fault tree data into the pre-trained root cause analysis model to obtain the root cause analysis results;

[0010] Input the root cause analysis results into the pre-trained defect elimination model to obtain the defect handling strategy;

[0011] Perform resource allocation analysis based on defect handling strategies and multi-source data to obtain defect handling solutions;

[0012] Push the defect handling solution to the client terminal and instruct the client terminal to execute the defect handling solution;

[0013] Among them, the pre-trained root cause analysis model and the pre-trained defect elimination model are both obtained by model training based on historical fault processing data of power grid equipment.

[0014] In one possible implementation, data processing is performed based on multi-source data to obtain fault tree data, including:

[0015] Extract features based on multi-source data to obtain a feature-enhanced data table;

[0016] Perform logical path conversion based on the feature enhancement data table and the device identification of the power grid equipment to obtain the logical path of the fault tree;

[0017] Based on the feature-enhanced data table, the device identification of the power grid equipment, and the topological structure data contained in the multi-source data, graph data is constructed to obtain the graph data of the fault tree; the topological structure data refers to the connection relationship between the power grid equipment and its related other power grid equipment;

[0018] Generate fault tree data based on the logical paths and diagram data.

[0019] In one possible implementation, the fault tree data is input into a pre-trained root cause analysis model to obtain a root cause analysis result, including:

[0020] Input the logical path in the fault tree data into the fault tree analysis module in the root cause analysis model to obtain the root cause probability distribution results and the path characteristics of the fault event;

[0021] The graph data in the fault tree data is input into the graph neural network module in the root cause analysis model to obtain the correlation characteristics of the power grid equipment and the propagation path of the fault event;

[0022] The root cause probability distribution results, the path characteristics of the fault event, the correlation characteristics of the power grid equipment and the propagation path of the fault event are input into the fusion module in the root cause analysis model to obtain the root cause analysis results.

[0023] In one possible implementation, the root cause analysis results are input into a pre-trained defect elimination model to obtain a defect handling strategy, including:

[0024] Obtain the real-time resource status of the power grid equipment, construct the state space based on the root cause analysis results and the real-time resource status, and obtain the state vector;

[0025] The state vector is input into the defect elimination model for reasoning to obtain multiple defect handling actions and the distribution probability of each defect handling action;

[0026] Based on real-time resource status, root cause analysis results, defect handling actions and their corresponding distribution probabilities, policy optimization is performed to obtain a defect handling strategy.

[0027] In one possible implementation, a resource allocation analysis is performed based on the defect handling strategy and multi-source data to obtain a defect handling solution, including:

[0028] Determine the resource allocation strategy for each defect handling action in the defect handling strategy based on multi-source data, real-time resource status, and demand information of power grid equipment;

[0029] Based on the resource allocation strategy corresponding to each defect handling action, the defect handling strategy is updated to obtain the optimized defect handling strategy;

[0030] Based on the optimized defect handling strategy, environmental data from multi-source data, and root cause analysis results, the defect handling process is designed to obtain a defect handling solution;

[0031] Among them, environmental data refers to the environmental data corresponding to power grid equipment in multi-source data.

[0032] In one possible implementation, the method further includes:

[0033] Obtain the execution results of the defect handling solution fed back by the customer terminal;

[0034] Determine the defect handling plan, execution results, and multi-source data corresponding to the defect handling plan as historical processing records, and add the historical processing records to a preset incremental database;

[0035] When the number of accumulated historical processing records in the preset incremental database reaches a preset threshold, the root cause analysis model and the defect elimination model are updated based on the accumulated historical processing records.

[0036] In a second aspect, an embodiment of the present application provides a fault handling device for a power grid device, comprising:

[0037] An acquisition module, configured to collect multi-source data corresponding to power grid equipment within a preset time period in response to a power grid equipment fault processing request;

[0038] A first processing module is used to process data based on multi-source data to obtain fault tree data;

[0039] The second processing module is used to input the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result;

[0040] The third processing module is used to input the root cause analysis results into the pre-trained defect elimination model to obtain the defect treatment strategy;

[0041] The fourth processing module is used to perform resource allocation analysis based on the defect handling strategy and multi-source data to obtain a defect handling solution;

[0042] a fifth processing module, configured to push the defect handling solution to the client terminal and instruct the client terminal to execute the defect handling solution;

[0043] Among them, the pre-trained root cause analysis model and the pre-trained defect elimination model are both obtained by model training based on historical fault processing data of power grid equipment.

[0044] In a possible implementation, the first processing module is further configured to:

[0045] Extract features based on multi-source data to obtain a feature-enhanced data table;

[0046] Perform logical path conversion based on the feature enhancement data table and the device identification of the power grid equipment to obtain the logical path of the fault tree;

[0047] Based on the feature-enhanced data table, the device identification of the power grid equipment, and the topological structure data contained in the multi-source data, graph data is constructed to obtain the graph data of the fault tree; the topological structure data refers to the connection relationship between the power grid equipment and its related other power grid equipment;

[0048] Generate fault tree data based on the logical paths and diagram data.

[0049] In a possible implementation, the second processing module is further configured to:

[0050] Input the logical path in the fault tree data into the fault tree analysis module in the root cause analysis model to obtain the root cause probability distribution results and the path characteristics of the fault event;

[0051] The graph data in the fault tree data is input into the graph neural network module in the root cause analysis model to obtain the correlation characteristics of the power grid equipment and the propagation path of the fault event;

[0052] The root cause probability distribution results, the path characteristics of the fault event, the correlation characteristics of the power grid equipment and the propagation path of the fault event are input into the fusion module in the root cause analysis model to obtain the root cause analysis results.

[0053] In a possible implementation, the third processing module is further configured to:

[0054] Obtain the real-time resource status of the power grid equipment, construct the state space based on the root cause analysis results and the real-time resource status, and obtain the state vector;

[0055] The state vector is input into the defect elimination model for reasoning to obtain multiple defect handling actions and the distribution probability of each defect handling action;

[0056] Based on real-time resource status, root cause analysis results, defect handling actions and their corresponding distribution probabilities, policy optimization is performed to obtain a defect handling strategy.

[0057] In a possible implementation, the fourth processing module is further configured to:

[0058] Determine the resource allocation strategy for each defect handling action in the defect handling strategy based on multi-source data, real-time resource status, and demand information of power grid equipment;

[0059] Based on the resource allocation strategy corresponding to each defect handling action, the defect handling strategy is updated to obtain the optimized defect handling strategy;

[0060] Based on the optimized defect handling strategy, environmental data from multi-source data, and root cause analysis results, the defect handling process is designed to obtain a defect handling solution;

[0061] Among them, environmental data refers to the environmental data corresponding to power grid equipment in multi-source data.

[0062] In a possible implementation, the device further includes a sixth processing module, which is configured to:

[0063] Obtain the execution results of the defect handling solution fed back by the customer terminal;

[0064] Determine the defect handling plan, execution results, and multi-source data corresponding to the defect handling plan as historical processing records, and add the historical processing records to a preset incremental database;

[0065] When the number of accumulated historical processing records in the preset incremental database reaches a preset threshold, the root cause analysis model and the defect elimination model are updated based on the accumulated historical processing records.

[0066] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory, a processor;

[0067] Memory stores computer-executable instructions;

[0068] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and various possible implementations of the first aspect.

[0069] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the above-mentioned first aspect and various possible implementation methods of the first aspect.

[0070] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and various possible implementation methods of the first aspect.

[0071] The embodiments of the present application provide a fault handling method, device, equipment, medium, and product for power grid equipment. When the power grid equipment initiates a fault handling request, the method obtains multi-source data of the power grid equipment within a preset time period; performs data processing on the multi-source data to obtain fault tree data; inputs the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result for the power grid equipment fault; inputs the root cause analysis result into a pre-trained defect elimination model to obtain a defect handling strategy for the power grid equipment fault; combines the multi-source data and the defect handling strategy to perform resource allocation, obtains a final defect handling solution, and prompts the client terminal to execute the defect handling solution. Compared with the prior art, the present application uses a root cause analysis model to analyze the deep-seated causes of power grid equipment faults, providing more accurate root cause location of the fault; and uses a defect handling model to generate a defect handling strategy for fault handling, combines multi-source data to reasonably allocate resources, obtains a defect handling solution, improves the efficiency of defect elimination, and thus achieves the technical effect of improving fault handling efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0073] Figure 1 Schematic diagram of the process of troubleshooting the power grid equipment provided in this application Figure 1 ;

[0074] Figure 2 Schematic diagram of the process of troubleshooting the power grid equipment provided in this application Figure 2 ;

[0075] Figure 3 Schematic diagram of the process of troubleshooting the power grid equipment provided in this application Figure 3 ;

[0076] Figure 4 A schematic diagram of the structure of the fault handling device for power grid equipment provided in this application;

[0077] Figure 5 This is a schematic diagram of the structure of the electronic device provided in this application.

[0078] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0079] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0080] In the existing technology, when fault processing is performed on power grid equipment, data analysis is required for the operating data of the power grid equipment. The data analysis method is mainly to determine the fault type based on the data characteristics, and determine the processing strategy for solving the conceptual fault based on the mapping relationship between the fault type and the processing strategy.

[0081] However, due to the statistical analysis and machine learning methods in the existing technology, only the fault types of power grid equipment are analyzed, the specific causes of power grid equipment failures are ignored, and there is a lack of in-depth analysis of the faults. As a result, the fault resolution strategies in the existing technology cannot accurately solve the power grid equipment failures, resulting in the technical problem of low fault handling efficiency in the existing technology.

[0082] To address the above technical issues, this application proposes the following technical concept: upon detecting a fault handling request initiated by a power grid device, it collects multi-source data generated by the power grid device within a preset time period, conducts an in-depth analysis of the power grid device fault, obtains a root cause analysis result of the fault, combines the root cause analysis result with a defect handling model to obtain a defect handling strategy, and uses multi-source data to perform resource allocation analysis to obtain a defect handling solution suitable for the current fault. Compared with existing technologies, this application utilizes a root cause analysis model and a defect elimination model to perform a hierarchical fault cause analysis of power grid device faults and generate a detailed defect handling strategy, thereby achieving the technical effect of improving fault handling efficiency.

[0083] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0084] Figure 1Schematic diagram of the process of troubleshooting the power grid equipment provided in this application Figure 1 ,like Figure 1 As shown, the method includes:

[0085] S101 : In response to a power grid device fault processing request, collect multi-source data corresponding to power grid devices within a preset time period.

[0086] In this step, the multi-source data collected includes: operating status data of power grid equipment, defect records of power grid equipment, environmental parameters corresponding to the power grid equipment, power grid topology data corresponding to the power grid equipment, identification of the power grid equipment, and the time when the fault handling request for the power grid equipment was initiated. Multi-source data collection can be accomplished by using multiple IoT sensors deployed on the power grid equipment to collect operating status data and environmental parameters of the power grid equipment within a preset time period; and by accessing the distributed database corresponding to the power grid equipment to obtain defect record data. IoT sensors include, but are not limited to, current sensors, voltage sensors, temperature sensors, and various environmental sensors.

[0087] Exemplarily, the preset time period is set to the last 5 hours; the IoT sensors installed on the power grid equipment are used to collect the following data corresponding to the power grid equipment: current change data within 5 hours, voltage change data within 5 hours, equipment temperature change data within 5 hours, ambient temperature, humidity, air pressure, wind speed and gas concentration change data within 5 hours; the collected data are integrated into multi-source data and uploaded to the cloud data management platform based on the IoT gateway, and unified data storage and data management are performed using a distributed database.

[0088] S102: Process the data based on the multi-source data to obtain fault tree data.

[0089] In this step, data processing is performed based on multi-source data. The resulting fault tree data includes the fault tree's logical path and graph data. The fault tree's logical path includes: a time series feature matrix for power grid equipment, an event code sequence corresponding to the power grid equipment's faults, and an environmental feature vector corresponding to the power grid equipment's environmental parameters. The event corresponding to the power grid equipment's fault refers to the event type corresponding to the current fault: for example, whether the current fault is a voltage event, a vibration event, or a dual event.

[0090] Optionally, a possible implementation method of processing multi-source data to obtain fault tree data is:

[0091] S1021. Perform feature extraction based on multi-source data to obtain a feature enhanced data table.

[0092] In this step, before extracting features from multi-source data, it is necessary to clean and align the multi-source data. The data alignment and cleaning process is implemented as follows:

[0093] a1. Perform cleaning processing based on multi-source data, fill in missing values ​​and unify data formats to obtain standardized data.

[0094] In this step, the methods of filling missing values ​​can be: using interpolation to fill missing values, using mode filling or predicting the corresponding filling value of the missing value based on context prediction. Mr. Hu uses moving average or regression prediction to predict and fill according to the data trend before and after the time point.

[0095] It should be noted that when filling missing values, it is also possible to replace or delete outliers in multi-source data and convert data from different sources into a unified format. The unified format can be achieved by using standardized units, unified timestamps, and feature engineering encoding.

[0096] a2. Perform time series alignment on the time series data in the standardized data. Align the standardized data according to the timestamp and fill in the data at the exact time point using linear interpolation to obtain aligned data. The time series data refers to the time series data of the power grid equipment's current, voltage, equipment temperature, ambient temperature, ambient humidity, and ambient gas concentration.

[0097] In this step, feature extraction is performed using the aligned and normalized data to generate a feature-enhanced data table. Specifically, feature extraction is performed based on the aligned and normalized data to obtain multiple types of data features, which are then interactively integrated to generate the feature-enhanced data table. The multiple types of data features include statistical features, temporal features, environmental features, and semantic features.

[0098] S1022: Perform logical path conversion based on the feature enhancement data table and the device identification of the power grid device to obtain a logical path of the fault tree.

[0099] In this step, the logical path can be obtained by: extracting the timing characteristics of the power grid equipment from the feature enhancement data table to obtain a timing characteristic matrix; performing event detection based on the feature enhancement data table, determining the event type corresponding to each time node according to the feature enhancement data of different time nodes, combining multiple event types to obtain an event coding sequence; extracting environmental characteristics based on the feature enhancement data table to obtain an environmental feature vector; combining the timing characteristic matrix, the event coding sequence, and the environmental feature vector to obtain sequence data; mapping the sequence data to a preset fault tree template to obtain a logical path of the fault tree.

[0100] Exemplarily, the event code sequence may be obtained by:

[0101] b1. Threshold detection: Threshold detection is performed at each time node. When it is determined that the threshold range is exceeded, the corresponding event type is determined. For example, if the current is detected to be greater than 120% of the rated current value at time node 1, an overload event is determined at time node 1.

[0102] b2. Pattern recognition: Use the time series prediction model to predict time series data and determine the event type corresponding to the time series data.

[0103] b3. According to the mapping relationship between event type and code, the code corresponding to each event is obtained, and multiple codes are combined to obtain an event code sequence.

[0104] Optionally, a possible implementation method of mapping the sequence data into a preset fault tree template to obtain the logical path of the fault tree is:

[0105] c1. Based on the sequence data, calculate the correlation between each event in the event coding sequence and each fault tree node in the preset fault tree template.

[0106] c2. Obtain the logical path of the fault tree based on the degree of correlation.

[0107] S1023: Based on the feature enhancement data table, the device identification of the power grid device, and the topology structure data included in the multi-source data, graph data is constructed to obtain graph data of the fault tree.

[0108] In this step, topological data refers to the connections between grid devices and their associated grid devices. The purpose of constructing graph data is to capture the implicit connections between grid devices and other grid devices, quantify the interactions between devices, and provide topological input for graph neural networks.

[0109] Optionally, a possible implementation process for obtaining the fault tree diagram data is:

[0110] d1. Define a graph structure based on topology data, using power grid devices as nodes, physical connections and logical dependencies between power grid devices as edges, and historical fault propagation probabilities between power grid devices as edge weights.

[0111] Among them, the historical fault propagation probability is obtained based on the analysis of defect record data in multi-source data.

[0112] d2. Build graph data based on the feature-enhanced data table, the device identification of the power grid equipment, and the graph structure.

[0113] In this step, the graph data includes the node feature matrix and the adjacency matrix.

[0114] S1024: Generate fault tree data based on the logical path and the graph data.

[0115] In this step, the logical path and graph data are spliced ​​to obtain structured data. The storage form of the structured data can be a dictionary form, with the logical path as the keyword and the corresponding specific logical path data as the keyword value, and the graph data as the keyword and the corresponding specific graph data as the keyword value.

[0116] S103: Input the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result.

[0117] In this step, the fault tree data is input into the root cause analysis model. The fault tree analysis module, based on the root cause analysis model, processes the logical paths in the fault tree data to obtain the root cause probability distribution. The graph neural network module in the root cause analysis model processes the graph data to obtain a correlation characteristic matrix representing the correlation characteristics of power grid equipment. The results of the two modules are combined to obtain the root cause analysis results. The root cause analysis results include the root cause node distribution of power grid equipment faults, the chain impact path of power grid equipment faults, and the correlation map between power grid equipment and other related equipment.

[0118] It should be noted that the root cause analysis results in this step are obtained in the following Figure 2 Further explanation is given in the illustrated embodiment and no redundant description is given here.

[0119] S104: Input the root cause analysis results into a pre-trained defect elimination model to obtain a defect handling strategy.

[0120] In this step, the process of processing the root cause analysis results using the defect elimination model is as follows:

[0121] S1041. Create an action space based on the root cause analysis results.

[0122] S1042: Search for an optimal strategy based on the reward mechanism and the action space to obtain a defect handling strategy corresponding to the root cause analysis result.

[0123] It should be noted that the purpose of this step is to convert the root cause results into executable strategies. Figure 3 Further explanation is given in the embodiment shown and no redundant description is given here.

[0124] S105: Perform resource allocation analysis based on the defect handling strategy and multi-source data to obtain a defect handling solution.

[0125] In this step, resource allocation aims to optimize strategies within the actual constraints of power grid equipment and generate a feasible defect resolution plan. The defect resolution plan may include task sequences, resource allocation, and safety conditions.

[0126] Optionally, a possible implementation method for obtaining a defect handling solution is:

[0127] S1051. Determine a resource allocation strategy corresponding to each defect handling action in the defect handling strategy based on multi-source data, real-time resource status, and demand information of power grid equipment.

[0128] In this step, real-time resource status includes human resources and equipment resources; grid demand refers to the scope of impact of grid equipment failures. The defect handling strategy includes at least one defect, at least one defect handling action, the priority of each defect handling action, and the resource allocation plan corresponding to each defect. A single defect can correspond to multiple defect handling actions, and a single defect handling action can be used to eliminate multiple defects. The purpose of defect elimination is to resolve the grid equipment failure.

[0129] Exemplarily, the implementation process of determining the resource allocation strategy may be:

[0130] e1. Extract defect data based on multi-source data of power grid equipment to obtain defect data associated with faults of power grid equipment.

[0131] e2. Analyze the threat level of each defect in the defect handling strategy to the stability and security of the power grid based on defect data.

[0132] e3. Obtain historical fault handling data of power grid equipment based on multi-source data and analyze the occurrence frequency of each defect in the defect handling strategy.

[0133] e4. Based on the resource cost and time cost required to handle each defect, analyze the difficulty of handling each defect.

[0134] e5. Based on the fuzzy logic control algorithm and the processing difficulty, generation frequency and threat level, the priority of each defect handling action in the defect handling strategy is calculated.

[0135] For example, a possible implementation method of using the fuzzy logic control algorithm to determine the priority of the defect handling action is:

[0136] Step 1: The difficulty of defect handling, frequency of occurrence, and threat level are used as input variables of the fuzzy logic control algorithm; for each input variable, a corresponding fuzzy membership function is defined to map the precise input value to a fuzzy value.

[0137] Step 2: Describe the relationship between the fuzzy value combinations of input variables and the defect handling priority based on a pre-set fuzzy rule base. This pre-set fuzzy rule base is based on expert experience and grid operation data, expressing rules in the form of "if... then..." The pre-set fuzzy rule base also includes various possible input combinations, ensuring that reasonable priority judgments can be made under different conditions.

[0138] Step 3: Utilize the fuzzy inference engine to perform inference calculations based on the fuzzy rule base and the fuzzy values ​​of the input variables. If multiple rules are activated, a pre-defined synthesis method is used to combine the results of these rules to output the fuzzy priority for each defect. The pre-defined synthesis method can be either the maximum synthesis method or the minimum synthesis method.

[0139] Step 4: Based on the fuzzy priority results obtained by fuzzy reasoning, defuzzification is performed to convert them into specific priority scores to obtain the priority of each defect handling action.

[0140] The method of converting the fuzzy priority result into a specific priority score is defuzzification, wherein the defuzzification methods include but are not limited to: the center of gravity method and the maximum membership method.

[0141] In this step, taking the centroid method as an example, by calculating the centroid of the area enclosed by the fuzzy set membership function curve and the horizontal axis, an accurate priority value is obtained. This value is the priority score of the defect handling action and can be used for subsequent path planning.

[0142] e6. Based on the priority of each defect handling action and the real-time resource status, determine the resource allocation priority corresponding to each defect. Based on the resource allocation priority of the defect and the execution path of defect elimination, determine the resource allocation priority of the corresponding defect handling action.

[0143] In this step, the resource allocation strategy corresponding to each defect handling action is determined based on the real-time resource status and the priority of each defect handling action. The purpose is to prioritize the allocation of processing resources for high-priority defects under the constraints of the real-time resource status.

[0144] S1052. Based on the resource allocation strategy corresponding to each defect handling action, update the defect handling strategy to obtain an optimized defect handling strategy.

[0145] In this step, based on the resource allocation strategy for each defect handling action, the execution order of the optimized defect handling actions is determined, and the defect handling strategy is updated to obtain the optimized defect handling strategy. The resource allocation strategy for each defect handling action dynamically adjusts the execution order of the defect handling actions under resource constraints.

[0146] It should be noted that in the process of determining the resource allocation strategy in this step, the resource constraints in the defect handling process are determined by the real-time resource status. Under the premise of resource constraints, the execution order of the defect handling process is optimized through dynamic planning, so that high-priority defects are handled first, while avoiding resource conflicts and unnecessary time delays.

[0147] In this step, the obtained optimization processing strategy includes the execution order of defect processing actions, the defects corresponding to each defect processing action, and the estimated completion time of each defect processing action.

[0148] It should be noted that the defect handling actions in the optimization handling strategy can be executed not only serially but also in parallel according to the specific resource allocation situation.

[0149] S1053. Based on the optimized defect handling strategy, environmental data in the multi-source data, and the root cause analysis results, a defect handling process is designed to obtain a defect handling solution.

[0150] In this step, the environmental data refers to the environmental data corresponding to the power grid equipment in the multi-source data.

[0151] For example, a substation detects a main transformer heat dissipation fault (fault ID: D001). The corresponding optimized defect handling strategy includes: [{"Action":"Replace radiator","Equipment":"Transformer T1","Priority":8.9, "Resource requirement":["Maintenance team A","Heat dissipation module X"]},{"Action":"Reinforce insulator","Equipment":"Transmission tower L5","Priority":7.2,"Resource requirement":["Maintenance team B","Composite insulation material"]}].

[0152] Environmental data includes: {"temperature":38.5,"humidity":70,"wind speed":4.2,"weather":"sunny"}.

[0153] The adaptive adjustment method for the optimized defect handling strategy based on environmental data can be as follows: when the temperature in the environmental data is greater than 38 degrees Celsius and the defect handling action in the optimized strategy is an outdoor operation, the execution time window of the defect handling action is adjusted to 7:00 to 10:00, and the corresponding safety measure of "wearing a cooling vest" is added; when the wind speed in the environmental data is greater than level 6 and the defect handling action is an aerial operation, the high-altitude backup plan of "enabling drone assistance" is added.

[0154] In the adjusted optimization strategy, the main adjustment is made to "Action": "Replace radiator". The adjusted information is: {"Action": "Replace radiator","Equipment":"Transformer T1","Time":"07:00-09:00","Safety Measures":["Cooling vest","Portable air conditioner"],"High temperature backup plan":"Temporarily install auxiliary fan"}.

[0155] After adjusting the optimization processing strategy based on environmental data, use the root cause analysis results to add associated measures to the processing actions.

[0156] The root cause analysis results are: {"Root cause node":"Cooling system failure","Chain impact path":["Cooling failure → Transformer overheating → Capacitor aging → Voltage fluctuation"],"Chain risk value":0.92}.

[0157] The logic for associated processing is as follows: when the chain risk is greater than 0.9, redundancy measures ["Real-time temperature monitoring", "Backup cooling system"] are added to the "Action": "Replace radiator" action. If "Capacitor aging" is found in the root cause analysis impact path, the associated task "Capacitor bank insulation test" is added.

[0158] The processing result for the "Replace radiator" action is: {"Action":"Replace radiator","Redundancy measures":["Install temperature sensor (report every 5 minutes)","Deploy backup liquid cooling device"],"Related tasks":[{"Action":"Capacitor bank insulation test","Execution timing":"Immediately after radiator replacement"}]}.

[0159] Combine the above content to optimize the defect handling strategy, clarify the execution time and execution order of different defect handling actions, and generate corresponding defect handling solutions.

[0160] S106: Push the defect handling solution to the client terminal and instruct the client terminal to execute the defect handling solution.

[0161] In this step, the customer terminal generally refers to the control system or user interface of the power grid equipment, through which the operator can receive information and execute instructions; pushing the defect handling solution refers to sending the generated defect handling solution to the customer terminal via the network or other communication means.

[0162] It should be noted that, when the defect handling solution is pushed to the client terminal, a visual defect handling report may also be generated based on the defect handling solutions corresponding to various different grid devices in the current grid environment.

[0163] Optionally, the ZC's visual defect handling report can use a clustering algorithm to group defects corresponding to the defect handling steps in the defect handling plan, categorizing them into three priority categories: high priority, medium priority, and low priority. For the initial clustering results, a weight adjustment metric can be introduced to improve classification accuracy. Clustering quality can be verified using visual scatter plots and silhouette coefficients. The classification results can be compared with historical defect handling records to verify the rationality of the priorities. The classification model parameters can also be adjusted based on expert review feedback.

[0164] Calculate the proportion of each defect type across different priority levels and count the number of defects by type and priority. Analyze defect trends over time. For defect types with smaller sample sizes but higher risk, calculate their proportions and mark them as requiring special attention. For skewed data distributions, use boxplots to identify anomalies and provide appropriate explanations.

[0165] Optionally, after the client executes the defect handling solution, further model optimization may be performed based on the client's execution results. One possible implementation of the model optimization is as follows:

[0166] S1061. Obtain the execution result of the defect handling solution fed back by the client terminal.

[0167] In this step, a method for obtaining the execution result of the defect handling solution fed back by the customer terminal may be: collecting structured data fed back by the customer terminal through a mobile application.

[0168] Exemplarily, the collected execution results include: the code of the defect handling solution, abnormal events in executing the solution, actual environment parameters, resource consumption, and time.

[0169] For example, the collected execution results can be: {"Scheme Code":"Code 001","Execution Result":{"Actual Time":"07:15-09:45 (15 minutes over schedule)","Resource Consumption":{"Manpower":"3 people × 2.5 hours","Materials":["Radiator X-2","Insulation Adhesive - 1 roll"]},"Quality Acceptance":{"Transformer Temperature":"62°C (meets the standard)","Voltage Fluctuation":"±3% (exceeds the standard)"},"Newly Discovered Problems":["Capacitor has slight oil leakage","Heat Dissipation Bracket is Rusted"]}}.

[0170] S1062: Determine the defect handling solution, the execution result, and the multi-source data corresponding to the defect handling solution as historical processing records, and add the historical processing records to a preset incremental database.

[0171] In this step, the historical processing record is obtained based on the association of multiple data. The historical processing record includes but is not limited to: the code or identifier of the record, the characteristics of the defect handling solution, the execution effect of the defect handling solution, and the resource consumption value generated by the execution result.

[0172] S1063: When the number of accumulated historical processing records in the preset incremental database reaches a preset threshold, update the root cause analysis model and the defect elimination model based on the accumulated historical processing records.

[0173] It should be noted that the pre-trained root cause analysis model and the pre-trained defect elimination model mentioned in this embodiment are both obtained by model training based on historical fault processing data of power grid equipment.

[0174] In this step, the method of triggering the model update may also be to set an update time period, or to set a threshold for the decline of the model performance.

[0175] For example, the model update condition can be set to any of the following:

[0176] The cumulative number of historical processing records reaches 1,000; the model's operating cycle meets 1 month; the model's performance award ratio is greater than 10%.

[0177] The present invention provides a method for handling a fault in a power grid device. When the power grid device initiates a fault handling request, the method obtains multi-source data of the power grid device within a preset time period; performs data processing on the multi-source data to obtain fault tree data; inputs the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result for the power grid device fault; inputs the root cause analysis result into a pre-trained defect elimination model to obtain a defect handling strategy for the power grid device fault; combines the multi-source data and the defect handling strategy to allocate resources, obtains a final defect handling solution, and prompts the client terminal to execute the defect handling solution. Compared with the prior art, the present invention uses a root cause analysis model to analyze the deep-seated causes of power grid device faults, providing more accurate root cause location of the fault; and uses a defect handling model to generate a defect handling strategy for fault handling, combines multi-source data to reasonably allocate resources, obtains a defect handling solution, improves the efficiency of defect elimination, and thus achieves the technical effect of improving fault handling efficiency.

[0178] Figure 2 Schematic diagram of the process of troubleshooting the power grid equipment provided in this application Figure 2 ,like Figure 2 As shown, the method includes:

[0179] S201 : Input the logical path in the fault tree data into the fault tree analysis module in the root cause analysis model to obtain the root cause probability distribution result and the path characteristics of the fault event.

[0180] In this step, the root cause probability distribution refers to the probability of each root cause node causing the top event to occur. The path characteristics of the failure time refer to the extracted structural features of the roadbed path. The structural features of the logical path include: path hierarchy, logic gate type, and event weight.

[0181] For example, the input fault tree logic path is: A [transformer overload fault] -> B [cooling system failure]; A -> C [load surge]; B -> D [cooling fan failure]; B -> E [cooling oil leakage].

[0182] The root cause probability distribution results are: cooling fan failure P = 0.6, cooling oil leakage P = 0.3, and load surge P = 0.1. The path characteristics of the failure time are: the hierarchical feature corresponds to cooling system failure as a secondary intermediate event, and the logic gate refers to the cooling system failure and load surge as an OR logic gate.

[0183] In this step, the training process of the root cause analysis model is as follows:

[0184] S2011. Collect historical operating status, historical defect records, and equipment topology data of power grid equipment to obtain multi-source historical data.

[0185] S2012: Perform data structuring processing on the acquired multi-source historical data to obtain fault tree structured data, graph structured data, and feature data.

[0186] In this step, the fault tree structured data includes top events, intermediate events, root causes, and logic gates between events. The logic gates accurately describe the logical relationships between events. Graph structured data includes nodes, edges, and the corresponding node feature matrix and adjacency matrix. Feature data refers to statistical features derived from analyzing multi-source historical data.

[0187] For example, the statistical features may be: current variance, number of faults, and average temperature of the power grid equipment.

[0188] S2013. Perform data preprocessing on the fault tree structured data, the graph structured data, and the feature data to obtain a training data set.

[0189] In this step, data preprocessing methods include: missing value processing, outlier extraction and numerical standardization.

[0190] S2014: Perform model training on the root cause analysis model based on the training data set, and terminate the model training when the root cause analysis model meets a preset termination condition to obtain the root cause analysis model.

[0191] In this step, the preset termination condition may be: the loss value of the model is lower than the preset loss value, the accuracy of the model is higher than the preset accuracy, or the number of iterations of the model is greater than the preset number of iterations.

[0192] For example, the formula of the root cause analysis model is shown in Formula 1:

[0193]

[0194] Where A represents the final root cause analysis result; φ represents the overall nonlinear activation function; H i represents the potential root cause nodes and their logical paths after the fault tree is layered; F(X) represents the result of feature extraction from multi-source data; N represents the number of layers in the fault tree; G(E, T) represents the correlation feature matrix of the fused fault tree and graph neural network; E represents the causal relationship matrix generated by the fault tree; T represents the device correlation feature matrix extracted by the graph neural network; λ1, λ2, λ3, and λ1 represent weighting parameters. P(X, T) represents the interaction function, which calculates the coupling effect between multi-source features and device correlation paths.

[0195] S202: Input the graph data in the fault tree data into the graph neural network module in the root cause analysis model to obtain the correlation characteristics of the power grid equipment and the propagation path of the fault event.

[0196] In this step, the association characteristics of power grid devices refer to the vectorized representation of the hard dependencies between devices, such as load correlation and fault propagation intensity; the fault time propagation path refers to the chain conduction path of the quantified fault between devices.

[0197] For example, the input graph data is: transformer -->|power output| busbar; cooling fan -->|heat dissipation| transformer; oil pump -->|oil supply| cooling system.

[0198] Correlation features extracted from the graph data include: a load correlation of 0.85 between the transformer and the cooling fan; and a co-occurrence frequency of 0.7 between the cooling oil pump and the cooling fan. The resulting propagation path is: cooling fan failure -> transformer overheating -> bus voltage fluctuation.

[0199] S203: Input the root cause probability distribution result, the path characteristics of the fault event, the correlation characteristics of the power grid equipment, and the propagation path of the fault event into the fusion module in the root cause analysis model to obtain the root cause analysis result.

[0200] In this step, the root cause analysis result refers to the root cause and chain impact map obtained by integrating the structured causality of the fault tree with the correlation characteristics of the graph neural network.

[0201] For example, combining the examples in S201 and S202, the root cause probability distribution results, the path characteristics of the fault event, the correlation characteristics of the power grid equipment, and the propagation path of the fault event are weightedly fused, and the root cause analysis results obtained are: A [fan failure]--P=0.72-->B [cooling system failure]; B--intensity 0.8-->C [transformer overheating]; C-->D [bus voltage fluctuation].

[0202] Figure 3 Schematic diagram of the process of troubleshooting the power grid equipment provided in this application Figure 3 ,like Figure 3 As shown, the method includes:

[0203] S301 : Acquire the real-time resource status corresponding to the power grid equipment, construct a state space based on the root cause analysis result and the real-time resource status, and obtain a state vector.

[0204] In this step, the implementation resource state refers to the operating state of the power grid equipment at the current point in time, including parameters such as voltage, current, and temperature. The root cause analysis results refer to the results obtained through the root cause analysis model. The state space refers to the set of all possible states, and the state vector refers to the specific representation of the current state.

[0205] For example, the grid equipment currently causing the fault is a transformer. The resource status of the transformer shows that the temperature exceeds the threshold and the current exceeds the threshold. The root cause analysis result indicates that the cause of the fault is excessive load. The information contained in the state vector may be: temperature, current and load.

[0206] S302: Input the state vector into the defect elimination model for inference to obtain multiple defect handling actions and the distribution probability of each defect handling action.

[0207] In this step, the defect elimination model is a reinforcement learning-based model used to infer the optimal strategy for handling defects. The defect handling model calculates multiple possible defect handling actions based on the input state vector and assigns a probability to each defect handling action. This probability represents the applicability of the action under the current state.

[0208] In this step, the formula for training the defect elimination model is shown in Formula 2:

[0209]

[0210] in, represents the defect handling strategy output by the model, Refers to the information of the current defect, Indicates status The defect handling action under the plan; T represents the maximum processing time range; D represents the operating status data of the power grid equipment; E represents the environmental data; represents the reward function, which is used to measure the efficiency of defect handling and the optimal utilization of resources; represents the cost function for defect handling; represents the grid operation stability function; γ represents the discount factor; λ represents the cost control weight; μ represents the stability priority weight.

[0211] In this step, by involving a comprehensive reward function in the defect elimination model, considering the defect handling efficiency, resource optimization utilization and the stability of the power grid operation, the learning direction of the strategy is dynamically adjusted; the reinforcement learning algorithm is combined with the turntable transfer rules in the model to dynamically generate the defect handling strategy, and the adaptive learning rate is used to optimize the learning efficiency and convergence speed of the strategy.

[0212] S303: Based on the real-time resource status, the root cause analysis results, the defect handling actions and their corresponding distribution probabilities, a policy optimization process is performed to obtain a defect handling policy.

[0213] In this step, strategy optimization refers to selecting the optimal defect handling strategy by combining real-time status, root cause analysis, and model inference results. The defect handling strategy refers to the finalized action plan used to effectively resolve current equipment defects.

[0214] For example, consider a transformer whose real-time resource status indicates excessive temperature and current. Root cause analysis indicates this is due to excessive load. The state vector includes: temperature = 85°C, current = 150A, and load = 120%. The defect mitigation model, based on state vector reasoning, determines the following defect resolution actions: 1. Reduce load (probability 0.6), 2. Increase cooling (probability 0.3), and 3. Check circuit (probability 0.1). Combining the real-time status, root cause analysis, and model reasoning results, the optimal defect resolution strategy is selected: first, reduce load, as this has the highest probability and immediately reduces equipment stress. If the temperature still does not decrease, then increase cooling measures.

[0215] Figure 4 A schematic diagram of the structure of the fault handling device for power grid equipment provided in this application is shown as follows: Figure 4 As shown, the fault handling device for power grid equipment provided in this embodiment includes:

[0216] The acquisition module 401 is configured to collect multi-source data corresponding to the power grid equipment within a preset time period in response to a power grid equipment fault processing request.

[0217] The first processing module 402 is configured to process data based on multi-source data to obtain fault tree data.

[0218] The second processing module 403 is used to input the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result.

[0219] The third processing module 404 is configured to input the root cause analysis result into a pre-trained defect elimination model to obtain a defect handling strategy.

[0220] The fourth processing module 405 is configured to perform resource allocation analysis based on the defect handling strategy and multi-source data to obtain a defect handling solution.

[0221] The fifth processing module 406 is configured to push the defect handling solution to the client terminal and instruct the client terminal to execute the defect handling solution.

[0222] Among them, the pre-trained root cause analysis model and the pre-trained defect elimination model are both obtained by model training based on historical fault processing data of power grid equipment.

[0223] In a possible implementation, the first processing module 402 is further configured to:

[0224] Feature extraction is performed based on multi-source data to obtain a feature enhanced data table.

[0225] The logical path conversion is performed based on the feature enhancement data table and the device identification of the power grid equipment to obtain the logical path of the fault tree.

[0226] Based on the feature enhancement data table, the device identification of the power grid equipment, and the topological structure data contained in the multi-source data, graph data is constructed to obtain the graph data of the fault tree; among them, the topological structure data refers to the connection relationship between the power grid equipment and its other associated power grid equipment.

[0227] Generate fault tree data based on the logical paths and diagram data.

[0228] In a possible implementation, the second processing module 403 is further configured to:

[0229] The logical path in the fault tree data is input into the fault tree analysis module in the root cause analysis model to obtain the root cause probability distribution results and the path characteristics of the fault event.

[0230] The graph data in the fault tree data is input into the graph neural network module in the root cause analysis model to obtain the correlation characteristics of the power grid equipment and the propagation path of the fault event.

[0231] The root cause probability distribution results, the path characteristics of the fault event, the correlation characteristics of the power grid equipment and the propagation path of the fault event are input into the fusion module in the root cause analysis model to obtain the root cause analysis results.

[0232] In a possible implementation, the third processing module 404 is further configured to:

[0233] Obtain the real-time resource status corresponding to the power grid equipment, construct the state space based on the root cause analysis results and the real-time resource status, and obtain the state vector.

[0234] The state vector is input into the defect elimination model for reasoning to obtain multiple defect handling actions and the distribution probability of each defect handling action.

[0235] Based on real-time resource status, root cause analysis results, defect handling actions and their corresponding distribution probabilities, policy optimization is performed to obtain a defect handling strategy.

[0236] In a possible implementation, the fourth processing module 405 is further configured to:

[0237] Based on multi-source data, real-time resource status, and demand information of power grid equipment, the resource allocation strategy corresponding to each defect handling action in the defect handling strategy is determined.

[0238] Based on the resource allocation strategy corresponding to each defect handling action, the defect handling strategy is updated to obtain an optimized defect handling strategy.

[0239] Based on the optimized defect handling strategy, environmental data from multi-source data, and root cause analysis results, the defect handling process is designed to obtain a defect handling solution.

[0240] Among them, environmental data refers to the environmental data corresponding to power grid equipment in multi-source data.

[0241] In a possible implementation, the apparatus further includes a sixth processing module, which is configured to:

[0242] The execution results of the defect handling solution fed back by the customer terminal will be obtained.

[0243] The defect handling plan, execution results, and multi-source data corresponding to the defect handling plan are determined as historical processing records, and the historical processing records are added to a preset incremental database.

[0244] When the number of accumulated historical processing records in the preset incremental database reaches a preset threshold, the root cause analysis model and the defect elimination model are updated based on the accumulated historical processing records.

[0245] The device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail in this embodiment.

[0246] Figure 5 This is a schematic diagram of the structure of the electronic device provided in this application. Figure 5 As shown, the electronic device provided in this embodiment includes: at least one processor 501 and a memory 502. Optionally, the device further includes a communication component 503. The processor 501, the memory 502 and the communication component 503 are connected via a bus 504.

[0247] In a specific implementation process, at least one processor 501 executes the computer-executable instructions stored in the memory 502 , so that the at least one processor 501 executes the above-mentioned fault handling method or methods for power grid equipment.

[0248] The specific implementation process of the processor 501 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0249] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0250] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage.

[0251] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0252] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0253] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0254] The readable storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0255] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Alternatively, the readable storage medium may be an integral part of the processor. The processor and the readable storage medium may reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium may reside in a device as discrete components.

[0256] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, whether electrical, mechanical, or otherwise, through some interface.

[0257] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0258] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0259] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0260] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0261] Finally, it should be noted that those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The present invention is not limited to the precise structure described above and illustrated in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof. The scope of the present invention is limited solely by the appended claims.

Claims

1. A method for handling a fault of a power grid device, characterized in that: include: In response to a power grid device fault processing request, collecting multi-source data corresponding to the power grid device within a preset time period; Performing data processing based on the multi-source data to obtain fault tree data; Inputting the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result; Inputting the root cause analysis results into a pre-trained defect elimination model to obtain a defect handling strategy; Perform resource allocation analysis based on the defect handling strategy and the multi-source data to obtain a defect handling solution; Pushing the defect handling solution to the client terminal and instructing the client terminal to execute the defect handling solution; The pre-trained root cause analysis model and the pre-trained defect elimination model are both obtained by model training based on historical fault processing data of the power grid equipment.

2. The method according to claim 1, characterized in that The performing data processing based on the multi-source data to obtain fault tree data includes: Perform feature extraction based on the multi-source data to obtain a feature enhancement data table; Performing logical path conversion based on the feature enhancement data table and the device identification of the power grid device to obtain a logical path of the fault tree; Based on the feature enhancement data table, the device identification of the power grid device, and the topology data contained in the multi-source data, graph data is constructed to obtain graph data of the fault tree; wherein the topology data refers to the connection relationship between the power grid device and other associated power grid devices; The fault tree data is generated based on the logic path and the graph data.

3. The method according to claim 2, characterized in that Inputting the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result includes: Inputting the logical path in the fault tree data into the fault tree analysis module in the root cause analysis model to obtain the root cause probability distribution result and the path characteristics of the fault event; Inputting the graph data in the fault tree data into the graph neural network module in the root cause analysis model to obtain the correlation characteristics of power grid equipment and the propagation path of the fault event; The root cause probability distribution result, the path characteristics of the fault event, the associated characteristics of the power grid equipment, and the propagation path of the fault event are input into a fusion module in the root cause analysis model to obtain the root cause analysis result.

4. The method according to any one of claims 1 to 3, characterized in that Inputting the root cause analysis results into a pre-trained defect elimination model to obtain a defect handling strategy includes: Acquire the real-time resource status corresponding to the power grid device, construct a state space based on the root cause analysis result and the real-time resource status, and obtain a state vector; Inputting the state vector into the defect elimination model for reasoning to obtain multiple defect handling actions and a distribution probability of each defect handling action; Based on the real-time resource status, the root cause analysis result, the defect handling action and its corresponding distribution probability, a policy optimization process is performed to obtain the defect handling policy.

5. The method according to claim 4, characterized in that The performing of resource allocation analysis based on the defect handling strategy and the multi-source data to obtain a defect handling solution includes: Determining a resource allocation strategy corresponding to each defect handling action in the defect handling strategy based on the multi-source data, the real-time resource status, and the demand information of the power grid equipment; Based on the resource allocation strategy corresponding to each defect handling action, the defect handling strategy is updated to obtain an optimized defect handling strategy; Based on the optimized defect handling strategy, the environmental data in the multi-source data, and the root cause analysis result, a defect handling process design is performed to obtain the defect handling solution; The environmental data refers to the environmental data corresponding to the power grid equipment in the multi-source data.

6. The method according to claim 5, characterized in that The method further comprises: Obtaining the execution result of the defect handling solution fed back by the client terminal; Determining the defect handling solution, the execution result, and the multi-source data corresponding to the defect handling solution as historical processing records, and adding the historical processing records to a preset incremental database; When the number of accumulated historical processing records in the preset incremental database reaches a preset threshold, the root cause analysis model and the defect elimination model are updated based on the accumulated historical processing records.

7. A fault handling device for power grid equipment, characterized in that: include: An acquisition module, configured to collect multi-source data corresponding to the power grid equipment within a preset time period in response to a power grid equipment fault processing request; A first processing module is configured to process the multi-source data to obtain fault tree data; A second processing module is used to input the fault tree data into a pre-trained root cause analysis model to obtain a root cause analysis result; a third processing module, configured to input the root cause analysis result into a pre-trained defect elimination model to obtain a defect treatment strategy; a fourth processing module, configured to perform resource allocation analysis based on the defect handling strategy and the multi-source data to obtain a defect handling solution; a fifth processing module, configured to push the defect handling solution to a client terminal and instruct the client terminal to execute the defect handling solution; The pre-trained root cause analysis model and the pre-trained defect elimination model are both obtained by model training based on historical fault processing data of the power grid equipment.

8. An electronic device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

10. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 6 when the computer program is executed by a processor.

Images