Intelligent prediction method and system for line tripping under extreme and abrupt weather conditions

Abstract

The present invention relates to the field of transmission line tripping fault prediction, and disclosed are an intelligent prediction method and system for line tripping under extreme and abrupt weather conditions. The method comprises: acquiring first data that reflects at least one of an extreme weather condition and an abrupt weather condition, and second data that reflects line status; performing deep learning training on the basis of the first data and the second data to obtain a first prediction model; performing visualization and iterative optimization on the first prediction model to obtain a second prediction model; introducing a tower positioning box into the second prediction model to obtain a line tripping prediction model; and substituting prediction data, which is later than acquisition times of the first data and the second data, into the line tripping prediction model to obtain a tripping fault range and level. In the present invention, by introducing a tower positioning box, a fault location can be accurately determined on the basis of prediction models, thereby enabling prediction of 10 kV line tripping under an extreme weather condition, and the fault severity and risk level are automatically determined for each tower.

Description

A Smart Prediction Method and System for Line Tripping under Extreme and Transitional Weather Conditions Technical Field

[0001] This invention relates to the field of predicting power transmission line tripping faults, and in particular to an intelligent prediction method and system for power transmission line tripping under extreme and transitional weather conditions. Background Technology

[0002] Line tripping induced by extreme and abrupt weather changes occurs frequently. Examples include tripping caused by direct lightning strikes to insulators, external damage from heavy rain, and tree branches pressing on lines during heavy snow. Given the large number of 10kV distribution lines, their wide power supply area, and the complex terrain, current responses to line tripping are largely reactive. Actions are taken only after notification of a trip, or rely entirely on manual experience. While some preliminary predictions can be made based on extensive analysis of power grid and weather data, current methods only allow for section identification and isolation of faults, failing to accurately pinpoint the fault location or efficiently prevent or predict faults. This impacts the safe and stable operation of the power grid and the reliability of power supply. Therefore, there is an urgent need to research an intelligent prediction tool for line tripping under extreme and abrupt weather conditions. This tool should utilize historical weather and fault data as a foundation, combined with the influence of future weather conditions, to automatically and quickly predict the probability of 10kV line tripping. Existing patents and technologies combine power flow data and weather data from the power grid system to output fault prediction information, but they are lacking in fault location. This invention combines the latitude and longitude of the line towers to more accurately locate the fault location. Summary of the Invention

[0003] In view of the above-mentioned problems, the present invention is proposed.

[0004] Therefore, the problem that this invention aims to solve is that existing patents and technologies combine power flow data and weather data from the power grid system to output fault prediction information, which is lacking in fault location.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a method for intelligent prediction of line tripping under extreme and transitional weather conditions, comprising: acquiring first data that satisfies at least one criterion reflecting extreme or transitional weather, and second data that reflects the line condition; performing deep learning training based on the first and second data to obtain a first prediction model; visualizing and iteratively optimizing the first prediction model to obtain a second prediction model; introducing tower positioning frames into the second prediction model to obtain a line tripping prediction model; and substituting prediction data acquired later than the first and second data into the line tripping prediction model to obtain the tripping fault range and fault level.

[0006] As a preferred embodiment of the intelligent prediction method for line tripping under extreme and transitional weather conditions described in this invention, the following steps are taken: Before deep learning training, the first data and the second data must be kept consistent in both time scale and event scale, i.e., the time when the first data is acquired corresponds to the time when the second data is acquired and is consistent; the calibration event when the first data is acquired and the calibration event when the second data is acquired are the same event; the first time and the second data that are consistent in both time scale and event scale are used as a set of data for deep learning training.

[0007] As a preferred embodiment of the intelligent prediction method for line tripping under extreme and transitional weather conditions described in this invention, the training content of the first prediction model includes, based on the first data and the second data, training to obtain the occurrence status of the second data under different first data conditions according to the second data under the condition of obtaining the first data, that is: the probability of occurrence, cause of occurrence, scope of influence and degree of influence of the fault under different weather conditions.

[0008] As a preferred embodiment of the intelligent prediction method for line tripping under extreme and transitional weather conditions described in this invention, the visualization requirement of the second prediction model is that the prediction model, prediction process and prediction results are displayed in a visual graphic, and the visual graphic contains the prediction information of the prediction model.

[0009] As a preferred embodiment of the intelligent prediction method for line tripping under extreme and transitional weather conditions described in this invention, the step of introducing the tower positioning frame into the second prediction model includes introducing a corresponding tower positioning frame for the calibration event of the second data; the tower positioning frame includes the latitude and longitude, height, and tower shape of the tower in the calibration event of the second data.

[0010] As a preferred embodiment of the intelligent prediction method for line tripping under extreme and transitional weather conditions described in this invention, the step of substituting prediction data acquired later than the first and second data into the line tripping prediction model includes: using the first and second data acquired at the first time for model training, and substituting prediction data acquired at the second time, which is later than the first time, into the line tripping prediction model for prediction; the prediction data includes the first and second data at the second time; the amount of data in the first and second data at the second time is less than the amount of data in the first and second data at the first time.

[0011] As a preferred embodiment of the intelligent prediction method for line tripping under extreme and transitional weather conditions described in this invention, the step of obtaining the tripping fault range and fault level includes: obtaining the tripping fault range based on the influence range in the training content; obtaining the tripping fault level based on the probability of occurrence, cause of occurrence, and degree of influence of the fault in the training content; the fault level is divided into at least two levels.

[0012] Another objective of this invention is to provide an intelligent prediction system for line tripping under extreme and transitional weather conditions. This system can predict the probability, scope, and degree of impact of tripping faults based on future weather data and line data.

[0013] To address the aforementioned technical problems, this invention provides the following technical solution: a system for intelligent prediction of line tripping under extreme and transitional weather conditions, comprising: a data acquisition module, a model training module, and a prediction module; the data acquisition module acquires first data that satisfies at least one criterion reflecting extreme or transitional weather, and second data that reflects the line condition; the model training module performs deep learning training based on the first and second data to obtain a first prediction model, visualizes the first prediction model, and iteratively optimizes it to obtain a second prediction model; the prediction module incorporates the tower positioning frame into the second prediction model to obtain a line tripping prediction model, and substitutes prediction data acquired later than the first and second data into the line tripping prediction model to obtain the tripping fault range and fault level.

[0014] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the intelligent prediction method for line tripping under extreme and transitional weather conditions as described above.

[0015] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent prediction method for line tripping under extreme and transitional weather conditions as described above.

[0016] The beneficial effects of this invention are as follows: Based on historical weather and 10kV line tripping information, a prediction model is formed using intelligent technologies such as deep learning and visualization. This model can support the prediction of line tripping under complex weather conditions.

[0017] Furthermore, existing patents and technologies combine power flow data and weather data from the power grid system to output fault prediction information, which is lacking in fault location. This invention introduces a tower positioning frame, based on a prediction model and combined with visualization capabilities, to accurately locate the fault location, realize 10kV line tripping prediction under extreme weather conditions, and automatically provide the fault severity and risk level according to the tower classification, assisting operation and maintenance units to dynamically and quickly predict risks in order to reduce risks or take preventive measures. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0019] Figure 1 is a flowchart of a method for intelligent prediction of line tripping under extreme and transitional weather conditions in Example 1.

[0020] Figure 2 is a module structure diagram of a line tripping intelligent prediction system under extreme and transitional weather conditions in Example 3. Detailed Implementation

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0023] Example 1, referring to Figure 1, is the first embodiment of the present invention. This embodiment provides a method for intelligent prediction of line tripping under extreme and transitional weather conditions, as shown in Figure 1:

[0024] S1. Obtain first data that satisfies at least one of the criteria for extreme or transitional weather, and second data that reflects the condition of the line.

[0025] In this embodiment, the first data can be extreme weather data or transitional weather data. Extreme weather includes weather data such as thunderstorms, strong winds, heavy rain, heavy snow, high temperatures, and low temperatures. Transitional weather can be weather type data such as drastic changes in temperature, sudden changes in humidity, sudden changes in wind direction, and sudden changes in air pressure.

[0026] It should be noted that turning point weather refers to weather in which the weather type parameters change drastically; therefore, extreme weather includes turning point weather.

[0027] To further explain, the first data in this invention are parameter data such as temperature, precipitation, and wind level under extreme or transitional weather conditions. Since the difficulty of parameter measurement increases under extreme weather conditions, the amount of data is not required.

[0028] In this embodiment, the second data can be a combination of several types of information, including basic information of the 10kV line, line GIS information, line tripping information, and line defect information.

[0029] The basic information includes line number, line type, line length, material type, commissioning date, and maintenance unit; the line GIS information includes spatial location, grounding facilities, line route, topography, and environmental conditions; the line tripping information includes tripping time, tripping location, tripping cause, tripping type, recovery time, affected area, and handling measures; and the line defect information includes defect type, defect level, discovery time, defect description, and handling results.

[0030] It should be noted that in order to achieve the purpose of predicting tripping faults in this invention, the second data must include basic information and line tripping information. If the acquired information lacks either basic information or line tripping information, the acquired information cannot be used as the second data and should be discarded.

[0031] In an optional embodiment, the first and second data can be obtained using web crawling technology. For extreme weather and transitional weather data, they are usually obtained from meteorological department websites, meteorological data open platforms, or professional meteorological service websites. For 10kV line information, the data may be stored in the power company's internal database, operation and maintenance management system, or public data of relevant government regulatory platforms.

[0032] Before web scraping, choose a suitable programming language and framework, such as Python's Scrapy, BeautifulSoup, or requests libraries, and prepare corresponding storage and computing resources, such as cloud servers and databases. Use browser developer tools to analyze the target website structure and determine the location of the data to be scraped. Determine the URL pattern, parameters, and pagination logic of the target data. Write code to simulate login (if needed). Write code to parse the webpage content and extract the required data. Handle pagination and page turning logic to ensure data integrity. Add exception handling, delay, and proxy mechanisms to avoid being restricted by the website's anti-scraping policies. Store the scraped data in a database, such as MySQL or MongoDB. Design a reasonable data table structure to ensure data integrity and query efficiency. Clean the scraped data, removing invalid, erroneous, or duplicate data. Perform necessary preprocessing on the data, such as format conversion and data validation.

[0033] In an optional embodiment, the first and second data can also be collected offline. This involves confirming the data provider and data format, checking the integrity and format of the data files, and performing preliminary data cleaning, such as removing irrelevant formatting and correcting errors. A reasonable database structure, including table structure, field types, and indexes, should be designed based on the data content. The data files should be imported into the database using database management tools (such as MySQL Workbench, SQL Server Management Studio, etc.) or command-line tools (such as mysqlimport, sqlcmd, etc.). For CSV or Excel files, an import script may need to be written or a database import wizard may need to be used. After data import, data validation should be performed to ensure accuracy and integrity. The number of data records before and after import should be compared, and key fields should be checked for missing or outlier values. If the data comes from different sources, data integration may be necessary to eliminate duplication and unify data formats. After successful data import, a database backup should be performed to prevent data loss.

[0034] S2. The first prediction model is obtained by deep learning training based on the first and second data.

[0035] Before using the first and second data for deep learning training, it is necessary to ensure that the first and second data are consistent in terms of time scale and event scale. That is, the time when the first data is acquired corresponds to the time when the second data is acquired and is consistent; the calibration event when the first data is acquired is the same event when the second data is acquired.

[0036] It should be noted that the first data acquisition time refers to the time when parameter data is collected under extreme or transitional weather conditions, and the second data acquisition time refers to the time when the tripping fault occurs, such as xx month xx day xx hour. The first data acquisition time and the second data acquisition time correspond to each other and are consistent, but it does not mean that the first data acquisition time and the second data acquisition time are exactly the same. There may be a time difference, and the threshold of the time difference depends on the specific operating conditions, but the maximum cannot exceed 24 hours.

[0037] To further clarify, the calibration event during the first data acquisition, such as "heavy snow on xx month xx day xx hour", and the calibration event during the second data acquisition, such as "power outage due to heavy snow on xx month xx day xx hour", are the same event, meaning the calibrated weather events are the same.

[0038] The first and second data points, which are consistent in time scale and event scale, are used as a data pair for deep learning training.

[0039] The training content of the first prediction model includes, based on the first data and the second data, training to obtain the occurrence status of the second data under different conditions of the first data, i.e., the probability of tripping failure, cause of occurrence, scope of impact and degree of impact under different weather conditions.

[0040] The deep learning training methods can include training models such as multilayer perceptrons (MLP), convolutional neural networks (CNN), recurrent neural networks (RNN) and their variants, temporal convolutional networks (TCN), autoencoders, generative adversarial networks (GAN), and attention mechanisms.

[0041] In an optional embodiment, the first model may be trained by a convolutional neural network (CNN), specifically including data normalization, expressed by the formula:

[0042] Among them, X norm Let X represent the standardized data, μ represent the original data pair, and σ represent the mean of the data.

[0043] Building a CNN model:

[0044] Input layer: Receives standardized data X norm .

[0045] Convolutional layer: Y conv =ReLU(conv(X) norm W conv )+bconv )

[0046] Where conv represents the convolution operation, W conv Represented as convolution kernel weights, b conv ReLU represents the bias term, and ReLU represents the activation function.

[0047] Pooling layer: Y Pool =Pool(Y conv )

[0048] Here, Pool represents the pooling operation.

[0049] Fully connected layer: Y fc =ReLU(W fc Y Pool +b fc )

[0050] Among them, Y fc Represented as the output of a fully connected layer, W fc Let b represent the weights of the fully connected layer. fc This represents the bias term for the fully connected layer.

[0051] Output layer:

[0052] in, Let W represent the prediction result, ε represent the activation function of the output layer, and W represent the prediction result. out The weights of the output layer, b out This is represented as the bias term of the output layer.

[0053] The loss function L is constructed as follows:

[0054] Where N represents the sample size, y i Represented as a real label, This is expressed as the predicted probability.

[0055] Repeat the following steps until convergence: Forward propagation calculates the predicted value Calculate the loss L and update the weights W of each layer.

[0056] In an alternative embodiment, the first model may also be trained by a recurrent neural network (RNN), with the data normalization steps being the same as those for training a convolutional neural network (CNN).

[0057] Building an RNN model:

[0058] Input layer: Receives standardized data.

[0059] RNN layer: For each time step t, compute the hidden state h. tRepresented as h t =tanh(W ih X t +W hh h t-1 +b h )

[0060] Among them, X t W is represented as the input at time step t. ih W represents the weights input to the hidden layer. hh The weight b represents the weight from one hidden layer to another. h denoted as the bias term of the hidden layer, and tanh represents the activation function.

[0061] Output layer: Calculates the predicted output for each time step t. Represented as,

[0062] Among them, W ho The weights, b, represent the weights from the hidden layer to the output layer. o This represents the bias term of the output layer of the RNN model.

[0063] The loss function is constructed in the same way as the training process for a convolutional neural network (CNN).

[0064] Updating the weights using, for example, the Adam optimizer, is represented as follows:

[0065] Where α represents the learning rate. It is represented as the gradient of the loss function with respect to the weights.

[0066] Repeat the following steps until convergence: Initialize the weights, and for each time step t, compute h during forward propagation. t and Calculate the loss L, and calculate the gradient through backpropagation. Update the weight W.

[0067] S3. Visualize the first prediction model and iteratively optimize it to obtain the second prediction model.

[0068] The visualization requirements for the second prediction model are that the prediction model, prediction process, and prediction results are displayed in a visual graphic, and the visual graphic contains the prediction information of the prediction model.

[0069] S4. Introduce the tower positioning frame into the second prediction model to obtain the line tripping prediction model.

[0070] A corresponding tower positioning frame is introduced for the calibration event of the second data.

[0071] The information in the tower positioning frame includes the tower's latitude, longitude, height, and shape in the calibration event corresponding to the second data.

[0072] S5. Substitute the predicted data that is later than the acquisition time of the first and second data into the line tripping prediction model to obtain the tripping fault range and fault level.

[0073] The first and second data obtained at the first time point are used for model training, and the prediction data obtained at the second time point, which is later than the first time point, are substituted into the line trip prediction model for prediction.

[0074] The forecast data includes the first and second data at the second time point.

[0075] The amount of data in the first and second data points at the second time point is less than the amount of data in the first and second data points at the first time point.

[0076] It should be noted that the first time in this embodiment is the time when the data is acquired, which is the historical time, while the second time is the time of the current point to be predicted or a future time. Since the amount of information to be acquired for the future is relatively small, the amount of data in the second time is generally less than that in the first time.

[0077] The scope of the tripping fault is estimated based on the impact range obtained from the prediction model.

[0078] The tripping fault level is determined by comprehensively considering the probability of occurrence, causes, and impact of the fault obtained from the prediction model.

[0079] To further explain, the tripping fault level can be calculated by weighting the probability of occurrence, cause of occurrence and degree of impact of the fault. The specific fault level is obtained by comparing it with the preset threshold. The fault level is divided into at least two levels.

[0080] In an optional embodiment, the fault level can be divided into two levels, such as severe and non-severe. If the fault level is severe, a repair plan is immediately formulated. If the fault level is non-severe, the specific repair time can be determined according to the actual situation.

[0081] In an optional embodiment, the fault level can be further divided into three levels, such as low, medium and high. If the fault level is low, idle personnel are dispatched to inspect and repair it. If the fault level is medium, the specific repair time can be determined according to the actual situation, and a professional team can be organized to carry out the repair. If the fault level is high, a plan is immediately formulated for repair.

[0082] Example 2 is the second embodiment of the present invention, which differs from the first embodiment in that: a method for intelligent prediction of line tripping under extreme and transitional weather conditions further includes visualizing the first prediction model to obtain a second prediction model.

[0083] It should be noted that visualization methods can include knowledge graph visualization, geographic information system (GIS) visualization, interactive data visualization, and other methods.

[0084] In an optional embodiment, the second model can be obtained from knowledge graph visualization, specifically including defining entities and relationships:

[0085] Physical entities: such as power lines, substations, weather events, power outages, defect reports, etc.

[0086] Relationships: such as "located in", "caused", "affected", etc.

[0087] Design the ontology structure, define entity types and relationship types, use entity recognition technology to extract entities from the data, link them with entities in the ontology, and finally extract the relationships between entities from the data.

[0088] Knowledge graph storage:

[0089] Choose a storage medium: You can use graph databases such as Neo4j and OrientDB to store the knowledge graph and import the constructed knowledge graph into the graph database.

[0090] Visual design:

[0091] Choose a visualization tool: You can use tools such as Neo4j Browser, Cytoscape, and Gephi for visualization. Customize the color, shape, size, and other attributes of nodes and edges as needed to better express information.

[0092] In an optional embodiment, the second model can also be obtained from Geographic Information System (GIS) visualization (Note: if this method is used, it can generally be done after the location box is introduced in step S4), specifically including ensuring that all data contains geospatial information (such as latitude and longitude) and preparing the data format, such as CSV, Shapefile or GeoJSON.

[0093] Choose a GIS software, such as ArcGIS, QGIS, or Google Earth Pro.

[0094] Import data into GIS software.

[0095] For QGIS, open QGIS, select "Layer" > "Add Layer" > "Add Vector Layer" to import a Shapefile or GeoJSON file.

[0096] For non-spatial data, you can import a CSV file by using "Add Delimited Text Layer" and specify the geospatial field.

[0097] Different symbols and colors can be assigned to different data types. For example, lines can be set to different colors, tripping events can be marked in red, and defect reports can be marked in yellow.

[0098] Create different layers, each representing a data type (such as line, trip, defect, weather event).

[0099] Create a new layer by going to "Layer" > "Add Layer" > "Create Layer" > "New Shapefile Layer".

[0100] Use spatial connections or attribute connections to link data from different layers.

[0101] In QGIS, you can use the "Vector" > "Data Management Tools" > "Join Attributes by Location" function.

[0102] Spatial analysis can be performed using GIS software analysis tools, such as buffer analysis and overlay analysis.

[0103] Examine the spatial relationships between different layers, such as whether tripping events are concentrated near specific lines or weather events.

[0104] Utilize the interactive features of GIS software, such as clicking to view detailed information and filtering data for specific time periods.

[0105] Export the visualization results as images or PDF files. If you are using a web GIS platform, you can generate a shareable link or embed it into a webpage.

[0106] Example 3, referring to Figure 2, is the third embodiment of the present invention, which differs from the previous two embodiments in that: a system for intelligent prediction of line tripping under extreme and transitional weather conditions includes a data acquisition module 100, a model training module 200, and a prediction module 300; the data acquisition module 100 acquires first data that satisfies at least one criterion reflecting extreme or transitional weather, and second data that reflects the line condition; the model training module 200 performs deep learning training based on the first and second data to obtain a first prediction model, visualizes the first prediction model, and iteratively optimizes it to obtain a second prediction model; the prediction module 300 introduces the tower positioning frame into the second prediction model to obtain a line tripping prediction model, and substitutes prediction data acquired later than the first and second data acquisition times into the line tripping prediction model to obtain the tripping fault range and fault level.

[0107] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0108] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0109] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0110] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0111] It should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent prediction of line tripping under extreme and transitional weather conditions, characterized in that: include, Obtain first data that satisfies at least one criterion for extreme or transitional weather, and second data that reflects the condition of the line; A first prediction model is obtained by deep learning training based on the first data and the second data; The first prediction model is visualized and iteratively optimized to obtain the second prediction model; By incorporating the tower positioning frame into the second prediction model, a line tripping prediction model is obtained. The predicted data acquired later than the first and second data are substituted into the line tripping prediction model to obtain the tripping fault range and fault level.

2. The intelligent prediction method for line tripping under extreme and transitional weather conditions as described in claim 1, characterized in that: Before using the first and second data for deep learning training, it is necessary to ensure that the first and second data are consistent in terms of both time scale and event scale. The time of acquiring the first data corresponds to and is consistent with the time of acquiring the second data. The calibration event during the first data acquisition and the calibration event during the second data acquisition are the same event; The first and second data points, which are consistent in time scale and event scale, are used as a data pair for deep learning training.

3. The intelligent prediction method for line tripping under extreme and transitional weather conditions as described in claim 2, characterized in that: The training content of the first prediction model includes, based on the first data and the second data, training to obtain the occurrence status of the second data under different conditions of the first data, according to the second data, under the condition of obtaining the first data: The probability of failure, causes, scope of impact, and degree of impact under different weather conditions.

4. The intelligent prediction method for line tripping under extreme and transitional weather conditions as described in claim 3, characterized in that: The visualization requirements for the second prediction model are that the prediction model, prediction process, and prediction results are displayed in a visual graphic, and the visual graphic contains the prediction information of the prediction model.

5. The intelligent prediction method for line tripping under extreme and transitional weather conditions as described in claim 4, characterized in that: The step of introducing the tower positioning frame into the second prediction model includes introducing the corresponding tower positioning frame for the calibration event of the second data. The tower positioning frame includes the tower's latitude, longitude, height, and shape in the calibration event of the second data.

6. The intelligent prediction method for line tripping under extreme and transitional weather conditions as described in claim 5, characterized in that: The step of substituting the predicted data acquired later than the first data and the second data into the line tripping prediction model includes using the first data and the second data acquired at the first time for model training, and substituting the predicted data acquired at the second time, which is later than the first time, into the line tripping prediction model for prediction. The predicted data includes first and second data at the second time point; The amount of data in the first and second data at the second time is less than the amount of data in the first and second data at the first time.

7. The intelligent prediction method for line tripping under extreme and transitional weather conditions as described in claim 6, characterized in that: The process of obtaining the tripping fault range and fault level includes obtaining the tripping fault range based on the influence range in the training content. The tripping fault level is determined by comprehensively considering the probability of occurrence, causes, and impact of the faults in the training content. The fault levels are divided into at least two levels.

8. A system employing the intelligent prediction method for line tripping under extreme and transitional weather conditions as described in any one of claims 1 to 7, characterized in that: It includes a data acquisition module (100), a model training module (200), and a prediction module (300); The data acquisition module (100) acquires first data that satisfies at least one criterion for reflecting extreme weather or turning point weather, and second data that reflects the line condition; The model training module (200) performs deep learning training based on the first data and the second data to obtain a first prediction model, visualizes the first prediction model and iteratively optimizes it to obtain a second prediction model; The prediction module (300) introduces the tower positioning frame into the second prediction model to obtain the line tripping prediction model. The prediction data that is later than the acquisition time of the first data and the second data is substituted into the line tripping prediction model to obtain the tripping fault range and fault level.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent prediction method for line tripping under extreme and transitional weather conditions as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent prediction method for line tripping under extreme and transitional weather conditions as described in any one of claims 1 to 7.

Images