Typhoon emergency event data processing method and device and computer equipment

Abstract

The invention relates to a typhoon emergency event data processing method and device and computer equipment. The method comprises the following steps: acquiring original data of a typhoon emergency event; performing standardization processing on the original data to obtain standard data; inputting the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result comprises at least one of a power grid damaged area, a power failure range and recovery time; and generating a report document of the typhoon emergency event according to the prediction result. By adopting the method, the accuracy of the evaluation report can be improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly to a data processing method, apparatus, and computer device for typhoon emergency events. Background Art

[0002] With the acceleration of global climate change and urbanization, natural disasters, especially typhoons, have an increasingly serious impact on social economy. In the power industry, typhoons may cause large-scale power outages, seriously affecting people's lives and production.

[0003] In traditional technologies, for typhoon emergency events, disaster assessment is usually carried out manually and an assessment report is generated. However, disaster assessment is usually affected by many factors, and manual assessment is usually not accurate enough, affecting the accuracy of the impact assessment report.

[0004] Therefore, there is a problem of low accuracy of assessment reports in the current data processing technology for typhoon emergency events. Summary of the Invention

[0005] Based on this, in view of the above technical problems, it is necessary to provide a data processing method, apparatus, computer device, computer-readable storage medium, and computer program product for typhoon emergency events that can improve the accuracy of assessment reports.

[0006] In a first aspect, this application provides a data processing method for typhoon emergency events, including:

[0007] Obtaining the original data of the typhoon emergency event;

[0008] Performing standardization processing on the original data to obtain standard data;

[0009] Inputting the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a power grid damaged area, a power outage range, and a restoration time;

[0010] Generating a report document of the typhoon emergency event according to the prediction result.

[0011] In a second aspect, this application further provides a data processing apparatus for typhoon emergency events, including:

[0012] An obtaining module, configured to obtain the original data of the typhoon emergency event;

[0013] A processing module, configured to perform standardization processing on the original data to obtain standard data;

[0014] A prediction module, configured to input the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a power grid damaged area, a power outage range, and a restoration time;

[0015] A reporting module, configured to generate a report document of the typhoon emergency event according to the prediction result.

[0016] In a third aspect, the present application further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0017] Obtain the original data of the typhoon emergency event;

[0018] Perform standardization processing on the original data to obtain standard data;

[0019] Input the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a power grid damaged area, a power outage range, and a restoration time;

[0020] Generate a report document of the typhoon emergency event according to the prediction result.

[0021] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0022] Obtain the original data of the typhoon emergency event;

[0023] Perform standardization processing on the original data to obtain standard data;

[0024] Input the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a power grid damaged area, a power outage range, and a restoration time;

[0025] Generate a report document of the typhoon emergency event according to the prediction result.

[0026] In a fifth aspect, the present application further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0027] Obtain the original data of the typhoon emergency event;

[0028] Perform standardization processing on the original data to obtain standard data;

[0029] Input the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a power grid damaged area, a power outage range, and a restoration time;

[0030] Generate a report document of the typhoon emergency event according to the prediction result.

[0031] The above data processing method, device, computer device, computer-readable storage medium, and computer program product for typhoon emergency events obtain the original data of typhoon emergency events, perform standardization processing on the original data to obtain standard data, input the standard data into a multi-model fusion deep learning model to obtain a prediction result of typhoon emergency events, and generate a report document of typhoon emergency events according to the prediction result; it can predict the power grid damaged area, power outage range, restoration time, etc. that a typhoon may cause based on a multi-model fusion deep learning model and using various original data related to typhoon emergency events. Since the original data used can be sufficient and does not rely on manual evaluation, the accuracy of the evaluation report can be improved. Brief Description of the Drawings

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0033] Figure 1 It is a schematic flowchart of a data processing method for typhoon emergency events in an embodiment;

[0034] Figure 2 It is a schematic flowchart of a data processing method for typhoon emergency events in another embodiment;

[0035] Figure 3 It is a structural block diagram of a data processing device for typhoon emergency events in an embodiment;

[0036] Figure 4 It is an internal structure diagram of a computer device in an embodiment. Detailed Description of the Embodiments

[0037] In order to make the purpose, technical solutions, and advantages of the present application clearer, the following further details the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.

[0038] It should be noted that the terms "first", "second", etc. in the specification, claims and above-mentioned drawings of the present disclosure are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure described here can be implemented in an order other than those illustrated or described here. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are only examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0039] In an exemplary embodiment, as Figure 1 shown, a data processing method for typhoon emergency events is provided. In this embodiment, it is exemplified that the method is applied to a server. It can be understood that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0040] Step S102, obtaining the original data of the typhoon emergency event.

[0041] Among them, a typhoon emergency event refers to a failure such as power grid damage and power outage caused by a typhoon that requires emergency disposal measures. The original data includes but is not limited to meteorological data, satellite remote sensing data, disaster information, and grid operation data and fault reports of power enterprises.

[0042] In a specific implementation, for a typhoon emergency event, the server can obtain meteorological data, satellite remote sensing data, disaster information, and grid operation data and fault reports of power enterprises.

[0043] In practical applications, meteorological data, satellite remote sensing data, grid operation data and fault reports can be obtained from an authorized terminal or server through an Application Programming Interface (API). A web crawler program can also be developed to crawl pictures and text descriptions reflecting typhoon disasters from authorized web pages.

[0044] Step S104, performing standardization processing on the original data to obtain standard data.

[0045] Among them, the standard data can be the original data after standardization processing. The standardization processing includes but is not limited to data cleaning, data integration, data standardization, data verification, etc.

[0046] In specific implementation, the server can first perform data cleaning on the original data. For example, it can remove invalid data in the original data, fill in missing values in the original data, denoise the original data, etc. Then, it can integrate the cleaned data. For example, it can convert the cleaned data into a unified format, merge multi-source data into a complete data set, align timestamps, etc. It can also label the integrated data. For example, it can label key data, geographical information, etc. Finally, it can verify the integrity, rationality, accuracy, etc. of the data, and use the data that passes the verification as standard data.

[0047] Step S106: Input the standard data into a multi-model fusion deep learning model to obtain a prediction result for the typhoon emergency event. The prediction result includes at least one of the power grid damaged area, power outage range, and restoration time.

[0048] Among them, the multi-model fusion deep learning model can be a model obtained by fusing multiple models such as machine learning and deep learning. The power grid damaged area can be the predicted possible damaged area of the power grid. The power outage range can be the predicted possible power outage range. The restoration time can be the predicted possible restoration time of the power grid.

[0049] In specific implementation, the server can pre-train or set up a multi-model fusion deep learning model, input the standard data obtained through standardization processing into the multi-model fusion deep learning model, and predict the possible power grid damaged area, power outage range, and restoration time, etc. through the multi-model fusion deep learning model.

[0050] In practical applications, the multi-model fusion deep learning model can be obtained by fusing multiple models. In addition to predicting the power grid damaged area, power outage range, and restoration time, the multi-model fusion deep learning model can also identify evaluation results such as the timeliness of early warning and response, the timeliness and rationality of emergency team and equipment allocation, the timeliness and rationality of emergency repair task allocation, the risks and operation safety control during the emergency repair process, the timeliness of power restoration, and the problems existing in the power grid network structure under the influence of disasters. The server can use all the results obtained through the multi-model fusion deep learning model as the prediction results of the typhoon emergency event.

[0051] Step S108: Generate a report document for the typhoon emergency event according to the prediction result.

[0052] Among them, the report document can be, but is not limited to, an evaluation report of the typhoon emergency event.

[0053] In specific implementation, the server can use the artificial intelligence generated content (AIGC) technology to automatically generate an evaluation report of the typhoon emergency event by using the prediction result.

[0054] In practical applications, technologies such as natural language processing and pattern recognition can be used to extract outliers and potential problems from the prediction results. For example, in the timeliness evaluation of early warning and response, the system can identify problems such as delayed early warning release and untimely response initiation. It can also comprehensively analyze data from multiple dimensions, including meteorological data, power grid operation status, resource allocation situation, etc., to comprehensively identify problems. For example, in the timeliness evaluation of emergency team and equipment allocation, in addition to paying attention to the arrival time, the effectiveness of the dispatching strategy and resource matching degree can also be analyzed. In addition, a real-time monitoring and feedback mechanism can be introduced to ensure that problems can be captured and recorded in a timely manner. For example, in the timeliness evaluation of emergency repair task allocation, the task execution situation can be monitored in real time to identify unreasonable task allocations.

[0055] The server can automatically generate an evaluation report according to predefined templates and rules, combined with the deep prediction results. The content of the report can cover problem descriptions, cause analyses, and improvement suggestions for each emergency response link. The report can also include charts and visualization elements, such as heat maps, time axes, path planning diagrams, etc., to help readers more intuitively understand the problems and their solutions. In addition, experts in the power industry can be invited to participate in the review, and combined with expert opinions and best practice cases, the content of the report and improvement suggestions can be further optimized.

[0056] The above data processing method for typhoon emergency events obtains the original data of typhoon emergency events, performs standardization processing on the original data to obtain standard data, inputs the standard data into a multi-model fusion deep learning model to obtain the prediction results of typhoon emergency events, and generates a report document for typhoon emergency events according to the prediction results; it can predict the power grid damaged areas, power outage ranges, recovery times, etc. that may be caused by typhoons based on the multi-model fusion deep learning model and using various original data related to typhoon emergency events. Since the original data used can be sufficient and does not rely on manual evaluation, the accuracy of the evaluation report can be improved.

[0057] In an exemplary embodiment, the above step S102 may specifically include: obtaining meteorological data, remote sensing data, and power grid operation data through an application programming interface, and obtaining disaster data through a web crawler; storing the meteorological data, remote sensing data, power grid operation data, and disaster data in a database in a pre-set data format to obtain the original data of typhoon emergency events.

[0058] Among them, the meteorological data can be real-time meteorological observation data, such as wind speed, air pressure, rainfall, etc. The remote sensing data can be, but is not limited to, satellite remote sensing images. The power grid operation data can be various data related to the operation of the power grid, such as voltage, current, load, etc. The disaster data can be typhoon disaster pictures, text descriptions, etc. publicly available on the Internet through social media or news reports, etc.

[0059] In specific implementation, the server can obtain meteorological observation data, satellite remote sensing data, as well as power grid operation data and fault reports, etc. from authorized terminals or servers through API interfaces, and can also develop web crawler programs to scrape disaster pictures and text descriptions, etc. from authorized social media or news websites, and define a unified data format to store the obtained meteorological observation data, satellite remote sensing data, power grid operation data, fault reports, disaster pictures, text descriptions, etc. in the database in a unified data format as the original data of typhoon emergency events.

[0060] In this embodiment, by obtaining meteorological data, remote sensing data, and power grid operation data through application programming interfaces, and obtaining disaster data through web crawlers, and storing the meteorological data, remote sensing data, power grid operation data, and disaster data in the database in a preset data format, the original data of typhoon emergency events can be obtained, and multi-dimensional data related to typhoon emergency events can be obtained, improving the accuracy of disaster monitoring, early warning, and assessment.

[0061] In an exemplary embodiment, the above step S104 may specifically include: performing data cleaning on the original data to obtain first data; performing data integration on the first data to obtain second data; performing data annotation on the second data to obtain third data; and obtaining standard data when the third data passes the verification.

[0062] Among them, the first data can be the original data after cleaning. The second data can be the original data after cleaning and integration. The third data can be the original data after cleaning, integration, and annotation.

[0063] In specific implementation, the server can perform data cleaning on the original data, including removing invalid data in the original data, filling in missing values in the original data, denoising the original data, etc., to obtain first data; then perform integration on the first data, including converting the first data into a unified format, merging multi-source first data into a complete data set, aligning the time stamps of the first data, etc., to obtain second data; and can also perform annotation on the second data, including annotating key data, geographical information, etc. in the second data, to obtain third data; finally, verify the integrity, rationality, accuracy, etc. of the third data, and use the third data that passes the verification as the standard data.

[0064] In this embodiment, by performing data cleaning on the original data, the first data is obtained. By integrating the first data, the second data is obtained. By performing data annotation on the second data, the third data is obtained. When the third data passes the verification, the standard data can be obtained, which can realize the standardization of the original data and ensure the reliable implementation of subsequent data analysis and report generation.

[0065] In an exemplary embodiment, the above step S106 may specifically include: extracting features from the standard data to obtain the data features of the standard data, and selecting a target model from the multi-model fusion deep learning model; inputting the data features into the target model to obtain the target prediction result corresponding to the target model; and obtaining the prediction result of the typhoon emergency event according to the target prediction result.

[0066] Among them, the data features may be the features of the data related to the typhoon emergency event. For example, wind speed, rainfall, grid load, number of faults, etc. The target model may be the model selected from the multi-model fusion deep learning model. The target prediction result may be the result predicted by the target model.

[0067] In specific implementation, the multi-model fusion deep learning model may include multiple models. The server may extract features from the standard data to obtain the data features, and select the model related to the current focus content from the multi-model fusion deep learning model as the target model. Input the data features into the target model, and the target model identifies the data features to obtain the target prediction result. By fusing the target prediction results identified by multiple target models, the prediction result of the typhoon emergency event can be obtained.

[0068] For example, when focusing on the timeliness of early warning and response, the server may select the fusion model of the time series prediction model and the anomaly detection model from the multi-model fusion deep learning model as the target model, input the data features extracted from the meteorological data and the grid operation data into the target model to obtain the timeliness evaluation result of early warning and response. Similarly, the timeliness and rationality evaluation results of the allocation of emergency teams and equipment, the timeliness and rationality evaluation results of the distribution of repair tasks, the risk and operation safety control evaluation results during the repair process, the timeliness evaluation result of power restoration after repair, and the problems existing in the grid network structure under the influence of disasters can be obtained. By fusing the above contents, for example, through AIGC recognition, the prediction results such as the damaged area of the power grid, the power outage range, and the restoration time can be obtained, as well as the evaluation of the prediction results.

[0069] In this embodiment, by extracting features from the standard data, the data features of the standard data are obtained. Moreover, a target model is selected from the multi-model fusion deep learning model, the data features are input into the target model, the target prediction result corresponding to the target model is obtained, and according to the target prediction result, the prediction result of the typhoon emergency event is obtained, so that the typhoon emergency event can be evaluated multidimensionally, and the comprehensiveness and reliability of the evaluation report are improved.

[0070] In an exemplary embodiment, the step of inputting the data features into the target model to obtain the target prediction result corresponding to the target model may specifically include: inputting the first feature into the fusion model of the time series prediction model and the anomaly detection model to obtain the target prediction result reflecting the early warning response situation; or inputting the second feature into the fusion model of the optimization scheduling model and the simulation model to obtain the target prediction result reflecting the emergency handling situation; or inputting the third feature into the multi-objective optimization model to obtain the target prediction result reflecting the task allocation situation; or inputting the fourth feature into the fusion model of the Bayesian network and the fault tree to obtain the target prediction result reflecting the safety control situation; or inputting the fifth feature into the fusion model of the shortest path model and the dynamic scheduling model to obtain the target prediction result reflecting the emergency repair and power restoration situation; or inputting the sixth feature into the fusion model of the complex network, scenario simulation and stress test to obtain the target prediction result reflecting the power grid network situation.

[0071] Among them, the first feature may be the data features of meteorological data and power grid operation data. The second feature may be the data features of Geographic Information System (GIS) data and emergency resource inventory records. The third feature may be the data features of power grid fault reports and repair team information. The fourth feature may be the data features of Internet of Things sensor data and video surveillance data. The fifth feature may be the data features of power grid topology data and power grid operation data. The sixth feature may be the data features of power grid topology data and disaster data.

[0072] In specific implementation, a time series prediction model (Long Short Term Memory, LSTM) can be selected to process real-time meteorological observation data and other relevant information to predict the typhoon path and its impact on the power grid, combined with an anomaly detection model (Isolation Forest) to identify potential delay points in the early warning system, and obtain the target prediction result reflecting the timeliness of early warning and response.

[0073] Alternatively, adopt an optimized scheduling model (linear programming), combined with a simulation model (discrete event simulation), analyze the effects of different scheduling strategies, construct scheduling scenarios and simulate the execution situation through GIS data, emergency resource inventory records and historical scheduling records, evaluate its rationality and efficiency, and obtain the target prediction results reflecting the timeliness and rationality of the allocation of emergency teams and equipment.

[0074] Alternatively, apply a multi-objective optimization model (genetic algorithm), combined with a task scheduling algorithm, consider the severity of the fault, priority, workload and skill matching degree of the repair team, optimize the repair task allocation plan, and evaluate the rationality and execution effect of the task allocation through power grid fault reports, repair team information and historical repair records, and obtain the target prediction results reflecting the timeliness and rationality of the repair task allocation.

[0075] Alternatively, conduct risk assessment based on Bayesian networks and fault tree analysis, combined with Internet of Things sensor data (temperature, humidity, wind speed, harmful gas concentration) and video surveillance data, monitor on-site environmental parameters in real time, evaluate operation risks, and obtain the target prediction results reflecting the risks in the repair process and the operation safety control.

[0076] Alternatively, use the shortest path algorithm (Dijkstra algorithm) in graph theory, combined with a dynamic scheduling model (reinforcement learning), plan the optimal power restoration path to ensure safe and reliable operation, and evaluate the planning and execution effects of the power restoration process through power grid topology data, real-time power grid operation status data and historical power restoration records, and obtain the target prediction results reflecting the timeliness of power restoration.

[0077] Alternatively, apply complex network theory, combined with scenario simulation and stress testing, evaluate the impact of disasters on the power grid network structure, identify potential vulnerable links, analyze the performance of the power grid under extreme weather conditions through power grid topology data, historical disaster data and simulation tools, and obtain the target prediction results reflecting the problems existing in the power grid network structure under the influence of disasters.

[0078] In this embodiment, by inputting the first feature into the fusion model of the time series prediction model and the anomaly detection model, the target prediction result reflecting the early warning response situation is obtained. Or, by inputting the second feature into the fusion model of the optimized scheduling model and the simulation model, the target prediction result reflecting the emergency handling situation is obtained. Or, by inputting the third feature into the multi-objective optimization model, the target prediction result reflecting the task allocation situation is obtained. Or, by inputting the fourth feature into the fusion model of the Bayesian network and the fault tree, the target prediction result reflecting the safety control situation is obtained. Or, by inputting the fifth feature into the fusion model of the shortest path model and the dynamic scheduling model, the target prediction result reflecting the emergency repair and power restoration situation is obtained. Or, by inputting the sixth feature into the fusion model of the complex network, scenario simulation and stress test, the target prediction result reflecting the power grid framework situation is obtained, so as to comprehensively evaluate the typhoon emergency event and its response decision, and improve the comprehensiveness and accuracy of the evaluation.

[0079] In an exemplary embodiment, the above step S108 may specifically include: identifying the abnormal situation in the prediction result; obtaining the description of the potential problems of the typhoon emergency event according to the abnormal situation; and performing natural language processing on the description of the potential problems to obtain a report document.

[0080] Among them, the abnormal situation may be the abnormal result in the prediction result. The description of the potential problems may be the descriptive statement of the potential problems of the typhoon emergency event.

[0081] In specific implementation, the server may identify the abnormal situation in each prediction result through a pre-trained identification model, fuse the obtained multiple abnormal situations to obtain the description of the potential problems of the typhoon emergency event, and generate an evaluation report of the typhoon emergency event according to the description of the potential problems through natural language processing.

[0082] In this embodiment, by identifying the abnormal situation in the prediction result, obtaining the description of the potential problems of the typhoon emergency event according to the abnormal situation, and performing natural language processing on the description of the potential problems to obtain a report document, a report document can be automatically generated, improving the efficiency of report generation.

[0083] To facilitate those skilled in the art to deeply understand the embodiments of the present application, the following will be described with a specific example.

[0084] This application integrates technologies such as multi-source data collection, automated data processing, multi-model fusion deep learning, and Natural Language Generation (NLG) algorithms to achieve full-process automation in the power industry from data acquisition of emergency events to problem identification, assessment, and summary during the emergency response process. The system can not only integrate information from multiple sources such as micro-meteorological stations, satellite remote sensing, social media, news announcements, and the power industry to ensure the comprehensiveness and integrity of data, but also adopt an adaptive learning optimization model to deeply explore specific problems and bottlenecks during the emergency response process and provide objective and detailed problem analysis. Finally, with the help of the NLG algorithm, the system automatically generates a report focusing on evaluating the problems existing in the emergency response process, providing a scientific basis for improving future emergency response measures. This solution overcomes the problems of low efficiency, incomplete data, and strong subjectivity in traditional methods, significantly improving the efficiency, accuracy, and comprehensiveness of disaster response and providing scientific and timely disaster information support for decision-makers.

[0085] In one embodiment, a method for generating an assessment and summary report on typhoon emergency events based on AIGC technology is provided, which is applicable to the power industry. Through information technology means, this method realizes full-process automation from data collection, processing, analysis of typhoon emergency events to the generation of an assessment and summary report, significantly improving the efficiency, accuracy, and comprehensiveness of the power system in disaster response and providing scientific and timely decision-making support for the power industry. The specific content of this method is as follows:

[0086] The data collection module 2100 is responsible for obtaining real-time and historical data from multiple data sources to ensure the comprehensiveness and timeliness of the data. This is the basis of the entire technical solution, providing high-quality raw data for subsequent data processing, analysis, and report generation. Especially in the power industry, the comprehensiveness and accuracy of data are crucial for evaluating the damage of the power grid and formulating restoration plans.

[0087] The execution entity of this module is the server; the input is none; the output is raw data (weather station data, satellite remote sensing data, social media data, news announcements, power data, etc.); the data processing relationship is to transmit the collected raw data to the data processing module 2200.

[0088] The detailed steps are as follows:

[0089] Step S2111, determine the data source.

[0090] Weather station data: Real-time meteorological observation data, including wind speed, air pressure, rainfall, etc.

[0091] Satellite remote sensing data: High-resolution satellite images for monitoring the typhoon path and influence range.

[0092] Social media data: pictures and text descriptions of the disaster situation uploaded by users.

[0093] News announcements: disaster information and news reports published on the Internet.

[0094] Power data: power grid operation data, fault reports, maintenance records, etc.

[0095] Step S2112, data acquisition.

[0096] API interface:

[0097] Obtain real-time meteorological data through the API interface.

[0098] Obtain power grid operation data and fault reports through the API interface.

[0099] Web crawler technology:

[0100] Develop a web crawler program to scrape relevant information from public channels such as social media and the Internet.

[0101] For example, use the Python web crawler framework (Scrapy) to scrape the disaster situation uploaded by users from the Internet platform.

[0102] Manual entry:

[0103] For some data that cannot be obtained through the API or web crawler, it can be supplemented by manual entry.

[0104] For example, some power data, such as maintenance records, may need to be manually entered.

[0105] It should be noted that the above data are all information and data authorized by users or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0106] Step S2113, data storage.

[0107] Database selection:

[0108] Select a suitable database system, such as MySQL, PostgreSQL, MongoDB, etc., for storing different types of data.

[0109] Data format:

[0110] Define a unified data format and structure to ensure data consistency and readability.

[0111] For example, define a standard data table structure, including fields such as timestamp, data type, and data value.

[0112] Data entry into the database:

[0113] Store the acquired data in the database according to the defined format to ensure data security and integrity.

[0114] For example, use Structured Query Language (SQL) to insert meteorological data into a MySQL database.

[0115] Store grid operation data and fault reports in a MongoDB database.

[0116] Data backup:

[0117] Regularly back up the database to prevent data loss or corruption.

[0118] For example, use an automated script to back up the database once every morning at midnight.

[0119] Step S2114, data verification.

[0120] Data integrity check:

[0121] Ensure that all expected data fields have been successfully acquired and stored.

[0122] For example, check whether each piece of data contains a timestamp, data type, and data value.

[0123] Data quality check:

[0124] Check the rationality and accuracy of the data. For example, whether the wind speed is within a reasonable range, whether the location coordinates are correct, and whether the grid data is consistent, etc.

[0125] For example, use the data analysis library (Pandas) in Python to perform data quality checks and filter out abnormal data.

[0126] Exception handling:

[0127] For the discovered abnormal data, mark and record it, and correct or re-acquire it if necessary.

[0128] For example, record the abnormal data in a log file and notify relevant personnel for processing via email.

[0129] Step S2115, data update mechanism.

[0130] Scheduled tasks:

[0131] Set up scheduled tasks to regularly obtain the latest data from the data source to ensure data timeliness.

[0132] For example, use a Cron expression to set to obtain meteorological data from an API interface once every hour.

[0133] Real-time update:

[0134] For important data sources, a real-time update mechanism can be set up to ensure the timeliness of the data.

[0135] For example, use the WebSocket technology of the network communication protocol to achieve real-time data push.

[0136] Data synchronization:

[0137] Ensure data synchronization between various data sources to avoid data conflicts and inconsistencies.

[0138] For example, use a message queue (such as RabbitMQ) to achieve data synchronization and ensure data consistency.

[0139] Example process:

[0140] Step S2121, meteorological data collection:

[0141] Obtain real-time meteorological data including wind speed, air pressure, rainfall, etc. from an authorized meteorological website through an API interface.

[0142] Store the obtained data in a MySQL database to ensure uniform data format.

[0143] Step S2122, satellite remote sensing data collection:

[0144] Obtain high-resolution satellite images from an authorized satellite remote sensing image website through an API interface.

[0145] Use image processing technology to extract relevant information on typhoon paths and affected areas.

[0146] Store the processed data in a MongoDB database.

[0147] Step S2123, social media data collection:

[0148] Develop a web crawler program to capture pictures and text descriptions of the disaster situation uploaded by users from public social platforms with user authorization.

[0149] Use natural language processing technology to extract valuable information such as the disaster location and damage situation.

[0150] Store the extracted information in a MongoDB database.

[0151] Step S2124, news announcement collection:

[0152] Obtain disaster information and news reports from an authorized news website through an API interface or crawler technology.

[0153] Store the obtained information in a MySQL database to ensure data reliability.

[0154] Step S2125, Power data collection:

[0155] With authorization, obtain power grid operation data and fault reports through the API interface.

[0156] Store the obtained data in a MongoDB database to ensure a unified data format.

[0157] Regularly back up the database to prevent data loss.

[0158] Through the above steps, the data collection module 2100 can efficiently and comprehensively obtain relevant data on typhoon emergency events, especially power grid operation data and fault reports unique to the power industry, providing a solid foundation for subsequent data processing and analysis.

[0159] The data processing module 2200 is responsible for cleaning, integrating, and annotating the collected raw data to form a standardized data format. This process ensures the quality and consistency of the data, providing a reliable basis for subsequent data analysis and report generation. Especially in the power industry, the accuracy and consistency of data are crucial for evaluating the damage of the power grid and formulating recovery plans.

[0160] The execution entity of this module is the server; the input is raw data (weather station data, satellite remote sensing data, social media data, news announcements, power data, etc.); the output is standardized data; the data processing relationship is to transmit the processed standardized data to the data analysis module 2300.

[0161] The detailed steps are as follows:

[0162] Step S2211, Data cleaning.

[0163] Remove invalid data:

[0164] Filter out obviously incorrect or irrelevant data, such as null values, duplicate data, data with incorrect formats, etc.

[0165] For example, use the dropna() function in the Pandas library to remove null values and the drop_duplicates() function to remove duplicate data.

[0166] Handle missing values:

[0167] For missing data, methods such as interpolation, mean filling, and median filling can be used for processing.

[0168] For example, use the fillna() function in the Pandas library for mean filling.

[0169] Noise removal:

[0170] Identify and remove noisy data through statistical methods or machine learning methods.

[0171] For example, use the Z-score method to identify and remove outliers.

[0172] Step S2212, data integration.

[0173] Unify data format:

[0174] Convert data from different data sources into a unified format to ensure data consistency and readability.

[0175] For example, unify all timestamps into the Universal Time Coordinated (UTC) time format and convert all numerical data into floating-point numbers.

[0176] Data merging:

[0177] Merge multi-source data into a complete dataset to ensure data comprehensiveness.

[0178] For example, use the merge() function in the Pandas library to merge meteorological data and power grid operation data.

[0179] Data alignment:

[0180] Ensure that the timestamps of different data sources are aligned for time series analysis.

[0181] For example, use the resample() function in the Pandas library for time alignment.

[0182] Step S2213, data annotation.

[0183] Key data annotation:

[0184] Annotate key data, such as disaster-stricken areas, damaged facilities, power grid fault points, etc.

[0185] For example, use labels (such as "damaged", "normal") to annotate power grid operation data.

[0186] Geographic information annotation:

[0187] Annotate the data with geographic information, such as longitude and latitude, administrative regions, etc.

[0188] For example, use the geospatial library GeoPandas to perform geospatial annotation on satellite remote sensing data.

[0189] Step S2214, data verification.

[0190] Data integrity check:

[0191] Ensure that all expected data fields have been successfully obtained and stored.

[0192] For example, check whether each piece of data contains a timestamp, data type, and data value.

[0193] Data quality check:

[0194] Check the rationality and accuracy of the data. For example, whether the wind speed is within a reasonable range, whether the location coordinates are correct, and whether the power grid data is consistent, etc.

[0195] For example, use the describe() function in the Pandas library for statistical analysis to check the data distribution.

[0196] Exception handling:

[0197] For the detected abnormal data, mark and record it, and correct or re-obtain it if necessary.

[0198] For example, record the abnormal data in a log file and notify relevant personnel for processing via email.

[0199] Example process:

[0200] Step S2221, data cleaning:

[0201] Remove null values and duplicate data from the meteorological data:

[0202] import pandas as pd

[0203] # Read meteorological data

[0204] weather_data = pd.read_csv('weather_data.csv')

[0205] # Remove null values

[0206] weather_data.dropna(inplace=True)

[0207] # Remove duplicate data

[0208] weather_data.drop_duplicates(inplace=True)

[0209] Fill the missing values in the power grid operation data with the mean value:

[0210] # Read the power grid operation data

[0211] grid_data = pd.read_csv('grid_data.csv')

[0212] # Mean value filling

[0213] grid_data.fillna(grid_data.mean(), inplace=True)

[0214] Step S2222, data integration:

[0215] Merge the meteorological data and the power grid operation data:

[0216] # Merge data

[0217] merged_data = pd.merge(weather_data, grid_data, on='timestamp')

[0218] Align the data in terms of time:

[0219] # Time alignment

[0220] merged_data.set_index('timestamp', inplace=True)

[0221] resampled_data = merged_data.resample('H').mean()

[0222] Step S2223, data annotation:

[0223] Annotate the power grid operation data:

[0224] # Annotate damaged and normal states

[0225] resampled_data['status'] = resampled_data['fault_count'].apply(lambda x: 'Damaged' if x > 0 else 'Normal')

[0226] Annotate the geographical information of the satellite remote sensing data:

[0227] import geopandas as gpd

[0228] # Read satellite remote sensing data

[0229] satellite_data = gpd.read_file('satellite_data.shp')

[0230] # Add geographic information annotation

[0231] satellite_data['latitude'] = satellite_data.geometry.centroid.y

[0232] satellite_data['longitude'] = satellite_data.geometry.centroid.x

[0233] Step S2224, data verification:

[0234] Check the integrity and quality of the data:

[0235] # Check data integrity

[0236] print(resampled_data.isnull().sum())

[0237] # Check data quality

[0238] print(resampled_data.describe())

[0239] Record abnormal data:

[0240] # Record abnormal data

[0241] anomalies = resampled_data[(resampled_data['wind_speed'] < 0) |(resampled_data['wind_speed'] > 100)]

[0242] anomalies.to_csv('anomalies.csv')

[0243] Through the above steps, the data processing module 2200 can efficiently and accurately clean, integrate and annotate the collected raw data to form a standardized data format, providing a solid foundation for subsequent data analysis and report generation.

[0244] The multi-model fusion deep learning data analysis module 2300. Multi-model fusion deep learning analysis is one of the core links, aiming to deeply analyze the standardized data through advanced machine learning and deep learning algorithms to identify specific problems and bottlenecks in the emergency response process. The following is a detailed description, especially for the timeliness of early warning and response, the timeliness and rationality of the allocation of emergency teams and equipment, the timeliness and rationality of emergency repair task allocation, the risks and operation safety control during the emergency repair process, the timeliness of power restoration after emergency repair, and the problems existing in the power grid network structure under the influence of disasters in the emergency response process of the power industry.

[0245] The execution subject of this module is the server; the input is standardized data (meteorological data, satellite remote sensing data, social media data, news announcements, power data, etc.); the output is the analysis result (power grid damaged area, power outage range, restoration time, etc.); the data processing relationship is to transmit the analysis result to the report generation module 2400.

[0246] The detailed steps are as follows:

[0247] Step S2311, data preprocessing.

[0248] Normalize the data:

[0249] Normalize data with different features to the same scale to eliminate the dimension difference and improve the efficiency and accuracy of model training.

[0250] For example, use the Min-Max normalization or Z-score standardization method.

[0251] from sklearn.preprocessing import MinMaxScaler

[0252] scaler = MinMaxScaler()

[0253] normalized_data = scaler.fit_transform(standardized_data)

[0254] Feature extraction:

[0255] Extract key features from the standardized data, such as wind speed, rainfall, power grid load, number of faults, etc.

[0256] For example, use the Pandas library to select specific columns as features.

[0257] features = standardized_data[['wind_speed', 'rainfall', 'grid_load','fault_count']]

[0258] Step S2312, model selection.

[0259] Multi-model fusion deep learning analysis:

[0260] (1) Timeliness of early warning and response

[0261] Task requirements: Evaluate the response speed and effectiveness of the existing early warning system before a disaster occurs, identify the delay points in the early warning issuance and emergency response activation processes, and ensure that early warning information can be conveyed in a timely manner and trigger appropriate emergency measures.

[0262] Data preparation: It is necessary to prepare real-time meteorological observation data (such as wind speed, air pressure, rainfall) and satellite remote sensing images, combined with historical disaster data (such as affected areas, types of damaged facilities) and power grid operation status data (such as voltage, current, load), as well as auxiliary information in social media and news reports, to evaluate the actual performance and response time of the early warning system.

[0263] (2) Timeliness and rationality of the allocation of emergency teams and equipment

[0264] Task requirements: Evaluate the rationality and implementation efficiency of the emergency team and equipment allocation plan, analyze the effects of different dispatching strategies, ensure that resource allocation conforms to the actual situation and can reach the designated location quickly, and improve the overall efficiency of emergency response.

[0265] Data preparation: It is necessary to prepare GIS data (such as coordinates of each location, road network topology structure), emergency resource inventory records (such as types of materials, quantities, storage locations), historical dispatching records (such as dispatching logs in previous events), and real-time traffic data (such as current road conditions, estimated travel time) to support the evaluation of dispatching strategies and implementation situations.

[0266] (3) Timeliness and rationality of the distribution of repair tasks

[0267] Task requirements: Evaluate the rationality and implementation effects of the repair task distribution plan, ensure that task distribution takes into account the severity of the fault, priority, workload of the repair team, and skill matching, and can be dynamically adjusted according to the actual situation to ensure the efficient progress of the repair work.

[0268] Data Preparation: It is necessary to prepare power grid fault reports (such as the time, location, and type of faults), repair team information (such as personnel composition, skill levels, and current status), historical repair records (such as comparison of completion time and actual workload), and GIS data (such as the current location of the repair team and the coordinates of the fault location) to support the evaluation of task assignment and execution.

[0269] (4) Risks in the repair process and operation safety control

[0270] Task Requirements: Evaluate the effectiveness of risk management and operation safety control measures during the repair process to ensure that on-site environmental parameter monitoring, instant feedback, and early warning mechanisms can effectively guide operators to take appropriate protective measures, record and analyze safety incidents, and propose improvement suggestions.

[0271] Data Preparation: It is necessary to prepare Internet of Things (IoT) sensor data (such as temperature, humidity, wind speed, and concentration of harmful gases), video surveillance data (such as real-time video streams and screenshots of abnormal events), historical safety incident records (such as accident cause analysis reports), and weather forecast data (such as weather forecasts for the next few days) to support the evaluation and improvement of safety control measures.

[0272] (5) Timeliness of power restoration

[0273] Task Requirements: Evaluate the planning and execution effects of the power restoration process to ensure that the power restoration path is reasonably selected, the operation is safe and reliable, analyze the power restoration effect, and evaluate the overall restoration speed to ensure that the power supply can quickly return to normal.

[0274] Data Preparation: It is necessary to prepare power grid topology data (such as the topological relationship between power grid nodes and lines), real-time power grid operation status data (such as node voltage, line current, and substation load), historical power restoration records (such as comparison of completion time and actual workload), and GIS data (such as the locations of substations and distribution centers) to support the evaluation of the power restoration process and effects.

[0275] (6) Problems existing in the power grid network structure under the influence of disasters

[0276] Task Requirements: Evaluate the impact of disasters on the power grid network structure, identify potential vulnerable links, analyze the performance of the power grid under extreme weather conditions, and propose suggestions for optimizing the power grid structure to enhance its resilience and adaptability.

[0277] Data Preparation: It is necessary to prepare power grid topology data (such as the topological relationship between power grid nodes and lines), historical disaster data (such as affected areas and lists of damaged facilities), simulation tools (such as complex network theory models), and expert knowledge bases (such as best practice cases) to support the evaluation of the stability and disaster resistance of the power grid structure.

[0278] Select a suitable model:

[0279] (1) Timeliness of early warning and response: Select LSTM to process real-time meteorological observation data and other relevant information to predict typhoon paths and their impacts on the power grid. Combine with an anomaly detection model (Isolation Forest) to identify potential delay points in the early warning system. Use historical disaster data and power grid operation status data for model training and verification to ensure the accuracy of the evaluation results.

[0280] (2) Timeliness and rationality of emergency team and equipment allocation: Adopt an optimized scheduling model (linear programming), combined with a simulation model (discrete event simulation), to analyze the effects of different scheduling strategies. Through GIS data, emergency resource inventory records, and historical scheduling records, construct scheduling scenarios and simulate the execution situation to evaluate its rationality and efficiency. Use real-time traffic data to dynamically adjust the scheduling plan to ensure that resources can reach the designated location quickly.

[0281] (3) Timeliness and rationality of repair task allocation: Apply a multi-objective optimization model (genetic algorithm), combined with a task scheduling algorithm, considering the severity of the fault, priority, workload of the repair team, and skill matching, to optimize the repair task allocation plan. Through power grid fault reports, repair team information, and historical repair records, evaluate the rationality and execution effect of the task allocation. Use GIS data for path optimization to ensure the efficient progress of the repair work and be able to dynamically adjust according to the actual situation.

[0282] (4) Risks and operation safety control during the repair process: Conduct risk assessment based on Bayesian networks and fault tree analysis, combined with Internet of Things sensor data (temperature, humidity, wind speed, harmful gas concentration) and video surveillance data, to monitor on-site environmental parameters in real time and evaluate operation risks. Through an instant feedback and early warning mechanism, guide operators to take appropriate protective measures. Use historical safety event records for post-event analysis and put forward improvement suggestions to ensure the effectiveness of safety control measures.

[0283] (5) Timeliness of power restoration after repair: Use the shortest path algorithm in graph theory (Dijkstra algorithm), combined with a dynamic scheduling model (reinforcement learning), to plan the optimal power restoration path to ensure safe and reliable operation. Through power grid topology data, real-time power grid operation status data, and historical power restoration records, evaluate the planning and execution effect of the power restoration process. Use GIS data to support path selection, analyze the power restoration effect, and evaluate the overall restoration speed to ensure that the power supply can quickly return to normal.

[0284] Problems existing in the power grid network structure under the influence of disasters: Apply complex network theory, combine scenario simulation and stress testing to evaluate the impact of disasters on the power grid network structure, and identify potential vulnerable links. Analyze the performance of the power grid under extreme weather conditions through power grid topology data, historical disaster data, and simulation tools. Use an expert knowledge base to provide optimization suggestions to enhance the resilience and adaptability of the power grid structure, and ensure its stability and disaster resistance.

[0285] Model fusion strategy:

[0286] To improve the analysis accuracy and decision-making support ability for each specific task, a model fusion strategy can be adopted. First, by integrating multiple types of machine learning and deep learning models and combining their respective advantages, a comprehensive evaluation framework is formed. For example, in the timeliness evaluation of early warning and response, combining a time series prediction model with an anomaly detection model can more comprehensively capture the actual performance of the early warning system. Second, introduce an adaptive learning framework to enable the system to dynamically adjust data analysis strategies and model parameters according to the latest environmental information, ensuring that the model gradually improves its performance during long-term operation. Finally, through methods such as cross-validation, hyperparameter tuning, and error analysis, ensure the stability and reliability of each model, so as to provide scientific and timely decision-making support for the emergency response in the power industry.

[0287] Step S2313, model training:

[0288] Input real-time data. The system receives historical and real-time data from multiple channels, including meteorological observation data (such as wind speed, air pressure, rainfall), satellite remote sensing images, power grid operation status data (such as voltage, current, load), GIS data (such as coordinates of each location, road network topology), emergency resource inventory records (such as types of materials, quantities, storage locations), power grid fault reports (such as time, location, type of fault occurrence), IoT sensor data (such as temperature, humidity, wind speed, concentration of harmful gases), video surveillance data, historical dispatching and repair records, weather forecasts, and auxiliary information in social media and news reports. After preprocessing and feature extraction, these data provide comprehensive and accurate input for the model.

[0289] Output the prediction results. Based on the input real-time and historical data, the system uses selected machine learning and deep learning models to output the evaluation results. These results are used to evaluate the rationality of the existing emergency response process, identify potential delay points and bottlenecks, and propose improvement suggestions. The specific evaluation contents include: the response speed and effectiveness of the early warning system, the rationality and execution efficiency of the emergency team and equipment allocation plan, the rationality of the emergency repair task allocation and its dynamic adjustment ability, the effectiveness of the risk management and operation safety control measures during the emergency repair process, the rationality of the power restoration path planning and the overall restoration speed, as well as the performance and vulnerable links of the power grid network structure under extreme weather conditions.

[0290] For example, for the full-process review and optimization of typhoon emergency response, assume that a certain area has just experienced a severe typhoon attack, and the power company needs to review the entire emergency response process to evaluate its rationality and find room for optimization. The system provides scientific decision-making support for each emergency response link by integrating multi-source data.

[0291] (1) Timeliness of early warning and response.

[0292] Input real-time data: Historical meteorological observation data (such as typhoon path, wind speed change); Power grid operation status data (such as voltage and current fluctuation conditions); Disaster impact feedback in social media and news reports.

[0293] Output evaluation results: The system evaluates the time node and accuracy of early warning issuance, and identifies the delay points in the early warning information transmission process. For example, it is found that the early warning issuance time in some areas is too late, resulting in residents failing to take preventive measures in time.

[0294] (2) Timeliness and rationality of emergency team and equipment allocation.

[0295] Input real-time data: Historical dispatching records (such as the actual arrival time of emergency teams and equipment), traffic condition data (such as road conditions, transportation time); Emergency resource inventory records.

[0296] Output evaluation results: The system analyzes the effects of different dispatching strategies and identifies the problems existing in resource allocation. For example, it is found that some key equipment fails to reach the designated location in time, affecting the initial response speed.

[0297] (3) Timeliness and rationality of emergency repair task allocation.

[0298] Input real-time data: Historical emergency repair records (such as the comparison of completion time and actual workload), power grid fault reports; The working status and skill matching degree of emergency repair teams.

[0299] Output evaluation results: The system evaluates the rationality of the emergency repair task allocation and identifies the unreasonable aspects in the task allocation. For example, it is found that some high-priority tasks are not processed in a timely manner, affecting the overall emergency repair progress.

[0300] (4) Risks during the emergency repair process and operation safety control.

[0301] Input real-time data: Historical safety event records (such as accident cause analysis reports), Internet of Things sensor data (such as on-site environmental parameter monitoring); video surveillance data.

[0302] Output evaluation results: The system evaluates the risk management and operation safety control measures during the emergency repair process and identifies existing potential safety hazards. For example, it is found that there are insufficient safety protection measures in some operating environments, increasing the accident risk.

[0303] (5) Timeliness of emergency power restoration.

[0304] Input real-time data: Historical power restoration records (such as comparison of completion time and actual workload), power grid topology data; locations of substations and distribution centers.

[0305] Output evaluation results: The system evaluates the rationality of the power restoration path planning and the overall restoration speed, and identifies the bottlenecks during the power restoration process. For example, it is found that the power restoration sequence of some key facilities is unreasonable, prolonging the overall power restoration time.

[0306] (6) Problems existing in the power grid network structure under the influence of disasters.

[0307] Input real-time data: Historical disaster data (such as affected areas, lists of damaged facilities), power grid topology data; simulation tools.

[0308] Output evaluation results: The system evaluates the impact of disasters on the power grid network structure and identifies potential vulnerable links. For example, it is found that some lines are prone to damage under extreme weather conditions, affecting the stability of the power grid.

[0309] Step S2314, result verification.

[0310] To ensure the accuracy and reliability of the prediction results for each specific task, the following verification methods are adopted:

[0311] Cross-validation: Use the K-fold cross-validation method to evaluate the performance of the model and ensure its generalization ability and stability. For example, in the evaluation of the timeliness of early warning and response, the data set is divided into a training set and a test set, and the prediction accuracy of the model is verified through multiple iterations.

[0312] Hyperparameter Tuning: Apply methods such as grid search, random search, or Bayesian optimization to find the optimal hyperparameter configuration and improve model performance. For example, in the evaluation of the timeliness and rationality of emergency team and equipment allocation, adjust the parameters of the linear programming model to optimize the scheduling strategy.

[0313] Error Analysis: Analyze in detail the error between the prediction results and the actual observations, identify the factors that may cause deviations, and improve the model accordingly. For example, in the evaluation of the timeliness and rationality of emergency repair task allocation, compare the actual completion time and the model prediction time to evaluate the rationality of task allocation.

[0314] Visualization: Intuitively present the analysis results in the form of charts for easy user understanding and interpretation. For example, in the evaluation of the risk and operation safety control during emergency repair, generate a heat map to show the distribution of risk areas and draw an event evolution diagram on the time axis, etc.

[0315] Actual Application Testing: Deploy the model in a real environment, collect feedback, and continuously improve it to ensure that the model can provide reliable decision-making support in actual operations. For example, in the evaluation of the timeliness of emergency repair and power restoration, test the effect of the power restoration path planning through on-site drills and evaluate the overall restoration speed.

[0316] Expert Review: Invite experts in the power industry to participate in the review, and further optimize the model and evaluation results in combination with expert opinions and best practice cases. For example, in the evaluation of the problems existing in the power grid network structure under the influence of disasters, use the optimization suggestions provided by the expert knowledge base to enhance the resilience and adaptability of the power grid structure.

[0317] Example Process:

[0318] Data Preprocessing:

[0319] Normalize Data:

[0320] from sklearn.preprocessing import MinMaxScaler

[0321] scaler = MinMaxScaler()

[0322] normalized_data = scaler.fit_transform(standardized_data)

[0323] Feature Extraction:

[0324] features = standardized_data[['wind_speed', 'rainfall', 'grid_load', 'fault_count']]

[0325] Through the above steps, the data analysis module 2300 can efficiently and accurately conduct in-depth analysis on the standardized data, identify the scope and degree of the impact of typhoons on the power system, and provide a scientific basis for the generation of the evaluation summary report. This not only helps the power company make rapid and accurate decisions in disaster response but also provides important reference information for relevant agencies.

[0326] Based on the results of the deep learning analysis, the report generation module 2400 automatically generates a professional report focusing on evaluating the problems existing in the emergency response process using the NLG algorithm. This report not only records the deficiencies in the emergency response but also puts forward specific improvement suggestions, providing a scientific basis for the optimization of future emergency measures and enhancing the effectiveness of decision-making support.

[0327] Step S2411, problem identification. After completing the deep learning analysis, the system will automatically identify the key problems in the emergency response process. This step aims to discover the deficiencies in the emergency response through a data-driven approach and provide a clear direction for subsequent improvements.

[0328] (1) Automated problem detection:

[0329] Method: The system uses NLP and pattern recognition technologies to extract outliers and potential problems from the model prediction results. For example, in the evaluation of the timeliness of early warning and response, the system can identify problems such as delayed early warning issuance and untimely response initiation.

[0330] Application example: By comparing the actual early warning time and the predicted time, the system finds that the early warning issuance time in some areas is too late, affecting the residents' preparedness for prevention.

[0331] (2) Multi-dimensional analysis:

[0332] Method: The system conducts comprehensive analysis on data from multiple dimensions, including meteorological data, power grid operation status, resource allocation situation, etc., to comprehensively identify problems. For example, in the evaluation of the timeliness of emergency team and equipment allocation, the system not only focuses on the arrival time but also analyzes the effectiveness of the dispatching strategy and resource matching degree.

[0333] Application example: The system finds that some key equipment fails to reach the designated location in time, affecting the initial response speed.

[0334] (3) Real-time feedback mechanism:

[0335] Method: Introduce a real-time monitoring and feedback mechanism to ensure that problems can be captured and recorded in a timely manner. For example, in the evaluation of the timeliness of repair task allocation, the system monitors the task execution situation in real time and identifies unreasonable task allocations.

[0336] Application example: The system found that high-priority tasks were not processed in a timely manner, affecting the overall emergency repair progress.

[0337] Step S2412, Evaluation summary generation. Based on the results of deep learning analysis, the system automatically generates a professional report focusing on evaluating the problems existing in the emergency response process using the NLG algorithm. This report not only records the deficiencies in the emergency response but also puts forward specific improvement suggestions, providing a scientific basis for optimizing future emergency measures and enhancing the effectiveness of decision-making support.

[0338] (1) Automated report generation:

[0339] Method: The system automatically generates an evaluation report according to predefined templates and rules, combined with the results of deep learning analysis. The report content covers problem descriptions, cause analyses, and improvement suggestions for each emergency response link.

[0340] Application example: For the timeliness issue of early warning and response, the report generated by the system points out the reasons for the late release of the early warning and puts forward specific suggestions for optimizing the early warning process.

[0341] (2) Structured report content:

[0342] Problem description: Specifically record the problems existing in each emergency response link, such as delayed early warning release, unreasonable resource allocation, improper task assignment, etc.

[0343] Cause analysis: Deeply analyze the root causes of the problems, such as poor information transmission, lack of flexibility in the scheduling strategy, insufficient resource reserves, etc.

[0344] Improvement suggestions: Put forward specific improvement suggestions for each problem, such as optimizing the early warning process, adjusting the scheduling strategy, increasing resource reserves, etc.

[0345] Application example: For the timeliness issue of the allocation of emergency teams and equipment, the report recommends optimizing the scheduling strategy to ensure that key equipment can be deployed more quickly.

[0346] (3) Visualization display:

[0347] Method: The report contains rich charts and visualization elements, such as heat maps, timelines, path planning diagrams, etc., to help readers more intuitively understand the problems and their solutions.

[0348] Application example: The report contains a heat map showing the risk distribution in different regions and a timeline showing the key time nodes of the emergency response.

[0349] (4) Expert review and feedback:

[0350] Method: Invite experts in the power industry to participate in the review, and combine expert opinions and best practice cases to further optimize the report content and improvement suggestions.

[0351] Application example: After expert review, the system adjusts some improvement suggestions according to expert opinions to ensure their feasibility and effectiveness.

[0352] Step S2413, report storage.

[0353] Store the generated report file in the database to ensure the security and accessibility of the report.

[0354] For example, use SQL statements to store the report file path in the database.

[0355] Step S2414, report review.

[0356] Conduct a manual review of the generated report to ensure the accuracy and integrity of the report content.

[0357] For example, send the report to the designated reviewer for review.

[0358] Through the above steps, the report generation module 2400 can efficiently and accurately generate a detailed evaluation summary report, providing scientific and timely decision-making support for power companies and relevant institutions. This not only helps power companies make quick and accurate decisions in disaster response but also provides important reference information for relevant institutions.

[0359] The report display and distribution module 2500 is responsible for displaying the generated evaluation summary report to users and providing download and sharing functions. This process ensures the usability and accessibility of the report, enabling power companies and relevant institutions to quickly obtain and utilize the information in the report, thus making scientific and timely decisions.

[0360] The execution entity is the client; the input is the evaluation summary report (in formats such as PDF, Word, etc.); the output is the report display page visible to users; the data processing relationship is to read the report data from the database and display it on the client interface.

[0361] The detailed steps are as follows:

[0362] Step S2511, report display.

[0363] Report list display:

[0364] Display the report list on the client interface, and users can choose to view specific reports.

[0365] For example, use web front-end technologies (such as React, Vue.js) to build the report list page.

[0366] Report details display:

[0367] After the user clicks on an item in the report list, the detailed content of the report is displayed, including text descriptions and charts.

[0368] For example, use web front-end technology to display the report details page.

[0369] Step S2512, report download.

[0370] Download button:

[0371] A download button is provided on the report details page, allowing users to download the report file.

[0372] For example, use HTML and JavaScript to implement the download function.

[0373] Step S2513, report sharing.

[0374] A sharing function is provided, enabling users to share the report via email, social media, etc.

[0375] For example, use JavaScript to implement the sharing function.

[0376] Through the above steps, the report display and distribution module 2500 can efficiently and conveniently display the evaluation summary report to users, and provide download and sharing functions, ensuring the usability and accessibility of the report. This not only helps power companies make quick and accurate decisions in disaster response, but also provides important reference information for relevant agencies.

[0377] The device and method provided in this embodiment have efficient data processing and analysis capabilities. Through the automated data cleaning and integration module, it can efficiently process the raw data from multiple data sources, ensuring the quality and consistency of the data, thereby reducing the need for manual intervention and improving the efficiency of data processing. Moreover, by using advanced machine learning and deep learning models (such as LSTM, random forest, etc.), it can accurately predict the impact range and degree of typhoons on the power system. These models are fully trained and can provide high-precision analysis results in a short time, providing a scientific basis for decision-making.

[0378] The device and method provided in this embodiment also have comprehensive report generation and display functions. Through natural language generation technology, it can automatically generate a detailed evaluation summary report, converting complex analysis results into easy-to-understand natural language text, which not only improves the readability of the report, but also ensures the professionalism and accuracy of the report. Moreover, it supports multiple report formats (such as PDF, Word, etc.) to meet the needs of different users. Users can conveniently view, download and share the report on the client interface, ensuring the rapid dissemination and application of information.

[0379] The device and method provided in this embodiment can also provide a user-friendly interaction experience. An intuitive report list and detailed display page are provided on the client interface, allowing users to easily browse and find the information they need. The report details page not only contains text descriptions but also embeds charts and images, making the information more intuitive and vivid. Moreover, users can download the report file through simple click operations and can also share the report in various ways such as by email and social media, facilitating communication and collaboration with other team members or external institutions. In addition, the built-in user feedback function allows users to submit their usage experiences and improvement suggestions, helping developers continuously optimize the system and enhance the user experience.

[0380] The above method for generating an evaluation summary report under typhoon emergency events based on AIGC technology improves the efficiency and accuracy of typhoon emergency event evaluation. First, the automated data cleaning and integration module can efficiently process the raw data from multiple data sources, remove invalid data, handle missing values, and remove noise, ensuring the quality and consistency of the data. Second, by using advanced machine learning and deep learning models (such as LSTM, random forest, etc.), it can quickly and accurately predict the impact range and degree of typhoons on the power system, providing a scientific basis for decision-making. In addition, the application of natural language generation technology enables complex data analysis results to be automatically generated into easy-to-understand evaluation reports, supporting output in multiple formats (such as PDF, Word), ensuring the professionalism and readability of the reports. Finally, the user-friendly report display and distribution module provides an intuitive interface, supporting convenient report viewing, downloading, and sharing functions, enhancing the dissemination and application efficiency of information. In summary, this method not only greatly improves the efficiency of data processing and analysis but also ensures the accuracy and usability of the evaluation report, providing a powerful decision-making support tool for power companies and related institutions.

[0381] In one embodiment, as Figure 2 shown, a data processing method for typhoon emergency events is provided. Taking the application of this method to a server as an example, it includes the following steps:

[0382] Step S301, obtain meteorological data, remote sensing data, and power grid operation data through an application programming interface, and obtain disaster data through a web crawler; store the meteorological data, remote sensing data, power grid operation data, and disaster data in a database in a pre-set data format to obtain the raw data of typhoon emergency events;

[0383] Step S302, perform data cleaning on the raw data to obtain first data; perform data integration on the first data to obtain second data; perform data annotation on the second data to obtain third data; when the third data passes the verification, obtain standard data;

[0384] Step S303: Extract features from the standard data to obtain the data features of the standard data, and select a target model from the multi-model fusion deep learning model; input the data features into the target model to obtain the target prediction result corresponding to the target model; and obtain the prediction result of the typhoon emergency event according to the target prediction result.

[0385] Step S304: Identify abnormal situations in the prediction result; obtain a description of potential problems of the typhoon emergency event according to the abnormal situations; and perform natural language processing on the description of potential problems to obtain a report document.

[0386] In specific implementation, the server can obtain meteorological data, remote sensing data, power grid operation data, etc. from authorized terminals or servers through API interfaces, extract pictures and text descriptions related to typhoon disasters from authorized web pages through web crawler technology, store the above information in a database in a unified data format to obtain the original data of the typhoon emergency event, and then perform standardized processing such as cleaning, integrating, annotating, and validating on the original data to obtain standard data, extract data features from the standard data, input the data features into the selected target model to obtain the corresponding target prediction result, obtain the prediction result of the typhoon emergency event according to multiple target prediction results, and perform natural language processing based on the prediction result to obtain document such as an evaluation summary report of the typhoon emergency event.

[0387] The above data processing method for typhoon emergency events can, based on a multi-model fusion deep learning model, use various original data related to typhoon emergency events to predict the power grid damaged areas, power outage ranges, restoration times, etc. that may be caused by typhoons. Since the original data used can be sufficient and does not rely on manual evaluation, the accuracy of the evaluation report can be improved.

[0388] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same moment, but can be executed at different moments, and the execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0389] Based on the same inventive concept, an embodiment of the present application further provides a data processing device for typhoon emergency events for implementing the data processing method of the typhoon emergency events involved above. The implementation solutions provided by this device to solve problems are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the data processing device for typhoon emergency events provided below can refer to the limitations on the data processing method of typhoon emergency events in the above text, and will not be elaborated here.

[0390] In an exemplary embodiment, as Figure 3 shown, a data processing device for typhoon emergency events is provided, including: an acquisition module 402, a processing module 404, a prediction module 406, and a reporting module 408, where:

[0391] The acquisition module 402 is configured to acquire the original data of the typhoon emergency event;

[0392] The processing module 404 is configured to perform normalization processing on the original data to obtain standard data;

[0393] The prediction module 406 is configured to input the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a power grid damaged area, a power outage range, and a recovery time;

[0394] The reporting module 408 is configured to generate a report document of the typhoon emergency event according to the prediction result.

[0395] In an exemplary embodiment, the above acquisition module 402 is further configured to acquire meteorological data, remote sensing data, and power grid operation data through an application programming interface, and acquire disaster data through a web crawler; store the meteorological data, the remote sensing data, the power grid operation data, and the disaster data in a database in a preset data format to obtain the original data of the typhoon emergency event.

[0396] In an exemplary embodiment, the above processing module 404 is further configured to perform data cleaning on the original data to obtain first data; perform data integration on the first data to obtain second data; perform data annotation on the second data to obtain third data; and obtain the standard data when the third data passes the verification.

[0397] In an exemplary embodiment, the above-mentioned prediction module 406 is further configured to extract features from the standard data to obtain the data features of the standard data, and select a target model from the multi-model fusion deep learning model; input the data features into the target model to obtain a target prediction result corresponding to the target model; and obtain the prediction result of the typhoon emergency event according to the target prediction result.

[0398] In an exemplary embodiment, the above-mentioned prediction module 406 is further configured to input the first feature into a fusion model of a time series prediction model and an anomaly detection model to obtain a target prediction result reflecting the early warning response situation; or input the second feature into a fusion model of an optimization scheduling model and a simulation model to obtain a target prediction result reflecting the emergency handling situation; or input the third feature into a multi-objective optimization model to obtain a target prediction result reflecting the task allocation situation; or input the fourth feature into a fusion model of a Bayesian network and a fault tree to obtain a target prediction result reflecting the safety control situation; or input the fifth feature into a fusion model of a shortest path model and a dynamic scheduling model to obtain a target prediction result reflecting the emergency repair and power restoration situation; or input the sixth feature into a fusion model of a complex network, scenario simulation, and stress test to obtain a target prediction result reflecting the power grid network situation.

[0399] In an exemplary embodiment, the above-mentioned reporting module 408 is further configured to identify abnormal situations in the prediction result; obtain a description of potential problems of the typhoon emergency event according to the abnormal situations; and perform natural language processing on the description of potential problems to obtain the report document.

[0400] Each module in the above-mentioned data processing device for typhoon emergency events can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0401] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 4As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data processing data for typhoon emergency events. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a data processing method for typhoon emergency events.

[0402] Those skilled in the art can understand that Figure 4 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0403] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0404] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0405] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0406] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0407] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0408] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in the present application.

[0409] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A data processing method for typhoon emergency events, characterized in that: The method comprises: Obtain the original data of typhoon emergency events; Performing standardization processing on the original data to obtain standard data; Inputting the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a damaged area of ​​the power grid, a power outage scope, and a restoration time; A report document of the typhoon emergency event is generated according to the prediction result.

2. The method according to claim 1, characterized in that The method of obtaining the original data of the typhoon emergency event includes: Obtain meteorological data, remote sensing data, and power grid operation data through application programming interfaces, and obtain disaster data through web crawlers; The meteorological data, the remote sensing data, the power grid operation data and the disaster data are stored in a database in a preset data format to obtain the original data of the typhoon emergency event.

3. The method according to claim 1, characterized in that The step of performing standardization processing on the original data to obtain standard data includes: Performing data cleaning on the original data to obtain first data; Performing data integration on the first data to obtain second data; Performing data labeling on the second data to obtain third data; When the third data is verified to be successful, the standard data is obtained.

4. The method according to claim 1, characterized in that The step of inputting the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event includes: Performing feature extraction on the standard data to obtain data features of the standard data, and selecting a target model from the multi-model fusion deep learning model; Inputting the data features into the target model to obtain a target prediction result corresponding to the target model; The prediction result of the typhoon emergency event is obtained according to the target prediction result.

5. The method according to claim 4, characterized in that The step of inputting the data features into the target model to obtain a target prediction result corresponding to the target model includes: Input the first feature into the fusion model of the time series prediction model and the anomaly detection model to obtain the target prediction result reflecting the warning response situation; or The second feature is input into the fusion model of the optimization scheduling model and the simulation model to obtain the target prediction result reflecting the emergency handling situation; or, Input the third feature into the multi-objective optimization model to obtain the target prediction result reflecting the task allocation situation; or, Input the fourth feature into the fusion model of Bayesian network and fault tree to obtain the target prediction result reflecting the safety control situation; or, Input the fifth feature into the fusion model of the shortest path model and the dynamic scheduling model to obtain the target prediction result reflecting the emergency repair and power restoration situation; or, The sixth feature is input into the fusion model of complex network, scenario simulation and stress test to obtain the target prediction result reflecting the grid grid situation.

6. The method according to claim 1, characterized in that Generating a report document of the typhoon emergency event according to the prediction result includes: identifying anomalies in the prediction results; According to the abnormal situation, a description of potential problems of the typhoon emergency event is obtained; Natural language processing is performed on the potential problem description to obtain the report document.

7. A data processing device for typhoon emergency events, characterized in that: The device comprises: An acquisition module is used to obtain the original data of typhoon emergency events; A processing module, used for performing standardization processing on the original data to obtain standard data; A prediction module, used for inputting the standard data into a multi-model fusion deep learning model to obtain a prediction result of the typhoon emergency event; the prediction result includes at least one of a damaged area of ​​the power grid, a power outage scope, and a restoration time; A reporting module is used to generate a report document of the typhoon emergency event according to the prediction result.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images