Inspection method and device for power grid

Abstract

The application discloses a power grid inspection method and device, and relates to the technical field of electric power. The method comprises the following steps: detecting whether a user instruction is received; if the user instruction is received, determining a communication protocol of each inspection point in the power grid based on a distributed service architecture, and collecting power grid data through the communication protocol; processing missing values and abnormal values in the power grid data, and integrating the processed power grid data through supplementary data to obtain integrated power grid data; predicting the integrated power grid data based on a prediction model to obtain an analysis result; wherein the analysis result at least comprises a comprehensive operation state of the power grid; the prediction model is a data analysis model constructed based on a machine learning algorithm; determining an inspection strategy and executing the inspection strategy according to the analysis result; wherein the inspection strategy comprises an optimal inspection path; and the optimal inspection path is obtained based on an inspection time, energy consumption and equipment health state prediction through a path planning model. The application is used for realizing the inspection function of the power grid.

Description

Technical Field

[0001] This application relates to the field of power technology, and in particular to a method and apparatus for inspecting power grids. Background Technology

[0002] With the continuous growth of electricity demand and the increasing scale of the power grid, traditional power grid inspection methods face numerous challenges. Traditional inspection methods mainly rely on regular manual inspections, which are not only inefficient but also limited by human resources and testing equipment, making it difficult to achieve comprehensive and real-time monitoring of power grid equipment. For example, when determining inspection strategies for transmission lines in remote areas using manual methods, the time-consuming process makes it difficult to promptly detect problems such as conductor wear and insulator aging. If these hidden dangers are not addressed in time, they could potentially lead to line faults, power outages, and affect the stability of power supply.

[0003] In recent years, although some power grid inspection methods have attempted to incorporate information technology to improve inspection efficiency, these methods often require connecting different data processing systems or devices to different dimensions of power grid data. These systems and devices collect and process the power grid data, which is then sent to a dedicated analysis server for analysis. Inspection strategies are generated based on these analysis server systems, and finally, the power grid management system executes these strategies to achieve the power grid inspection function. However, in practical applications, existing power grid inspection methods, whether for data collection, analysis, or strategy generation, all rely on different devices or terminals. This results in extensive data interaction between different systems or devices throughout the inspection process. Especially when dealing with inspection points in the power grid containing a large amount of data, existing power grid inspection methods are clearly constrained by the communication interfaces between systems, thus affecting processing efficiency. Summary of the Invention

[0004] This application provides a method and apparatus for inspecting a power grid, the main purpose of which is to solve the problem that the inspection efficiency of the power grid is easily affected by the communication interface between systems during the current inspection process.

[0005] To address the aforementioned technical problems, this application provides the following technical solutions:

[0006] Firstly, this application provides a power grid inspection method, applied to a power grid inspection system, wherein the power grid inspection system directly communicates with each inspection point in the power grid, and the method includes:

[0007] Check if a user command has been received;

[0008] If received, the communication protocol with each inspection point in the power grid is determined based on the distributed service architecture, and power grid data is collected through the communication protocol. The power grid data is collected based on multiple inspection points in the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators.

[0009] The process involves processing missing and outlier values ​​in power grid data, and then integrating the processed power grid data with supplementary data to obtain integrated power grid data. The supplementary data is a dataset designed to make the processed power grid data more accurate. The supplementary data includes equipment status and environmental information. Equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status. Environmental information includes weather conditions and geographical location.

[0010] The integrated power grid data is predicted based on a predictive model to obtain analysis results; the analysis results include at least the comprehensive operating status of the power grid; the predictive model is a data analysis model built based on machine learning algorithms.

[0011] Based on the analysis results, an inspection strategy is determined and executed; wherein, the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on the prediction of inspection time, energy consumption and equipment health status; the path planning model is a model built based on a deep reinforcement learning algorithm.

[0012] Secondly, this application also provides a power grid inspection device, applied to a power grid inspection system, wherein the power grid inspection system directly communicates with each inspection point in the power grid, and the device includes:

[0013] The detection unit is used to detect whether a user command has been received.

[0014] The determining unit is used, if received, to determine the communication protocol with each inspection point in the power grid based on a distributed service architecture, and to collect power grid data through the communication protocol. The power grid data is collected based on multiple inspection points in the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators.

[0015] An integration unit is used to process missing and outlier values ​​in the power grid data, and integrate the processed power grid data with supplementary data to obtain integrated power grid data. The supplementary data is a dataset designed to make the processed power grid data more accurate. The supplementary data includes equipment status and environmental information. Equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status. Environmental information includes weather conditions and geographical location.

[0016] The prediction unit is used to predict the integrated power grid data based on the prediction model and obtain the analysis results; wherein the analysis results include at least the comprehensive operating status of the power grid; the prediction model is a data analysis model built based on machine learning algorithms;

[0017] An execution unit is used to determine and execute an inspection strategy based on the analysis results; wherein the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on the prediction of inspection time, energy consumption and equipment health status; the path planning model is a model built based on a deep reinforcement learning algorithm.

[0018] Thirdly, embodiments of this application provide a storage medium including a stored program, wherein, when the program is executed, it controls the device where the storage medium is located to perform the power grid inspection method described in any one of the first aspects.

[0019] Fourthly, embodiments of this application provide a power grid inspection device, the device including a storage medium; and one or more processors, the storage medium being coupled to the processors, the processors being configured to execute program instructions stored in the storage medium; the program instructions, when executed, perform the power grid inspection method as described in any one of the first aspects.

[0020] By employing the above-described technical solution, the technical solution provided in this application has at least the following advantages:

[0021] This application provides a power grid inspection method and apparatus, applied to a power grid inspection system. The system directly communicates with each inspection point within the power grid. This application can detect whether a user command has been received. If received, based on a distributed service architecture, it determines the communication protocol with each inspection point in the power grid and collects power grid data through the communication protocol. This power grid data is collected from multiple inspection points within the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. Then, it processes missing and outlier values ​​in the power grid data and integrates the processed data with supplementary data to obtain integrated power grid data. The supplementary data is used to make the processed power grid data more accurate. The dataset includes supplementary data such as equipment status and environmental information. Equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status. Environmental information includes weather conditions and geographical location. Then, the integrated power grid data is predicted using a predictive model to obtain analysis results. These results include at least the overall operating status of the power grid. The predictive model is a data analysis model built using machine learning algorithms. Finally, based on the analysis results, an inspection strategy is determined and executed. This strategy includes an optimal inspection path, which is obtained through a path planning model based on inspection time, energy consumption, and equipment health status predictions. The path planning model is a model built using deep reinforcement learning algorithms, thereby realizing the power grid inspection function. Compared with existing technologies, the inspection method of this application can be based on a power grid inspection system. Since this power grid inspection system can communicate directly with each inspection point in the power grid, and the communication protocol of each inspection point is determined based on a distributed architecture during the process of acquiring power grid data, it can be ensured that data is collected separately from the equipment at each inspection point based on a distributed architecture. This avoids the process of collecting data from different inspection points through different systems and then summarizing it, which is required in existing technologies. Since the data generated by each inspection point is significantly less than the summed data, the power grid inspection process is not constrained by interface bandwidth, thus avoiding the problem of processing efficiency being affected by interface constraints during the acquisition of power grid data.Meanwhile, because the entire process of power grid inspection—from handling missing and outlier values ​​in power grid data to predicting and analyzing results based on preset models, and even determining inspection strategies based on the analysis results—is conducted within the power grid inspection system, this avoids the need for different systems or devices to perform data acquisition, data analysis, and prediction based on analysis, as is common in existing technologies. It also avoids data interaction between these systems or devices, thus preventing interface constraints that may exist when interacting via interfaces. This further improves the efficiency of the entire power grid inspection process and solves the problem of interface constraints in communication interactions between multiple systems or devices involved in the inspection process. Furthermore, since power grid data is collected based on the communication protocol of each inspection point, it ensures that the power grid data during inspection encompasses various devices in the power grid, such as substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. This ensures data comprehensiveness and lays the foundation for the accuracy of subsequent power grid inspection strategy formulation. Furthermore, since missing and outlier values ​​in the power grid data are processed before prediction, and the data is integrated based on supplementary data after processing, it is ensured that the data being analyzed is complete data containing equipment status and environmental information before the analysis results are obtained. This further ensures the accuracy of the analysis results obtained after prediction, and also ensures the accuracy of the inspection strategy obtained based on the analysis results.

[0022] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0023] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of this application are illustrated by way of example and not limitation, with the same or corresponding reference numerals denoteing the same or corresponding parts, wherein:

[0024] Figure 1 A flowchart of a power grid inspection method provided in an embodiment of this application is shown;

[0025] Figure 2 This paper shows a block diagram of a power grid inspection device according to an embodiment of the present application;

[0026] Figure 3 A block diagram of another power grid inspection device provided in an embodiment of this application is shown. Detailed Implementation

[0027] Exemplary embodiments of this application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of this application are shown in the drawings, it should be understood that this application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0028] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0029] This application provides a power grid inspection method, applied to a power grid inspection system. The power grid inspection system directly communicates with each inspection point in the power grid. Specifically, as follows... Figure 1 As shown, the method includes:

[0030] 101. Check if a user command has been received.

[0031] In this embodiment, the power grid inspection system can be understood as a system based on a distributed architecture, with multiple distributed terminals. Each distributed terminal can correspond to and communicate with at least one inspection point. Thus, the process of collecting power grid data is actually the process of each distributed terminal in the power grid inspection system collecting data from the equipment at the corresponding inspection point. Since each inspection point's equipment generates its own data, the amount of data acquired by each distributed terminal is relatively small. Compared with the prior art, which deploys one or more servers according to different terminal types for data docking and collection, the power grid inspection system of this application collects significantly less data from each distributed terminal than the amount of data collected by a server based on a dimension setting. This avoids the problem of a large amount of data collected by one server when there are many inspection points and equipment, which would require a large amount of interface bandwidth when handed over to other servers or systems. This avoids the situation of being constrained by the interface, and thus avoids the problem of the overall inspection efficiency being affected by the interface constraint from the data collection process.

[0032] In addition, for the sake of simplicity, in this embodiment, the power grid inspection system will be referred to as "this system" in subsequent descriptions. This will be explained here for the sake of simplicity and convenience in subsequent descriptions.

[0033] In this embodiment, before conducting a power grid inspection, a judgment is first made based on the method described in this step to determine whether a power grid inspection is necessary. Specifically, the system continuously listens for user command input, which can be received through a graphical interface, command-line tool, or API interface. When the user clicks the "Start Inspection" button on the interface or sends a corresponding API request, the system detects that a user command has been received.

[0034] 102. If received, the communication protocol with each inspection point in the power grid is determined based on the distributed service architecture, and power grid data is collected through the communication protocol.

[0035] The power grid data is collected from multiple inspection points within the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators.

[0036] Based on a distributed service architecture, this system pre-configures communication protocol templates for various inspection point devices (substation monitoring equipment, transmission line monitoring equipment, etc.). When a user command is detected, the system matches and determines the corresponding communication protocol from the templates according to the equipment type of each inspection point in the power grid.

[0037] For example, for substation monitoring equipment using the Modbus protocol, the system connects to the equipment via the Modbus TCP / IP protocol to read the power grid data stored in the equipment, including equipment operating parameters and environmental parameters. For transmission line monitoring equipment using the DL / T634.5104 protocol, the system assembles and parses data frames according to this protocol to achieve data acquisition.

[0038] 103. Process missing and outlier values ​​in the power grid data, and integrate the processed power grid data with supplementary data to obtain integrated power grid data.

[0039] The supplementary data is a dataset designed to make the processed power grid data more accurate. The supplementary data includes equipment status and environmental information. The equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status. The environmental information includes weather conditions and geographical location.

[0040] In this data processing step, the first step is to clean the collected power grid data, identifying and processing missing and outlier values. For example, if a monitoring device uploads data showing a negative current value, the system marks it as an outlier and corrects or removes it according to preset rules. Then, the system integrates the processed power grid data with supplementary data. This supplementary data includes equipment status (such as the operating status of power equipment, line load conditions, etc.) and environmental information (such as weather conditions, geographical location, etc.). This data comes from multiple sources, such as meteorological APIs and power grid topology databases.

[0041] For example, this system integrates the processed substation equipment operation data with the current weather conditions (obtained from the meteorological interface) and the substation's geographical location information (obtained from the power grid topology database) to form more comprehensive and accurate data.

[0042] 104. Based on the prediction model, the integrated power grid data is predicted, and the analysis results are obtained.

[0043] The analysis results include at least the overall operation status of the power grid; the prediction model is a data analysis model built based on machine learning algorithms.

[0044] In this embodiment, the system pre-trains a predictive model based on machine learning algorithms. This model uses historical power grid data as training samples to learn the complex relationship between the power grid's operating status and various factors.

[0045] When forecasting is required, the integrated power grid data is input into the forecasting model. This model analyzes and predicts based on the input power grid data, yielding the predicted results, i.e., the analysis results. These results include at least the overall operating status of the power grid. In practical applications, users can also set other results as part of the analysis, such as the health status of equipment at various inspection points. Furthermore, in this embodiment, the overall operating status of the power grid can be characterized by specific levels. For example, the analysis results predicted by the forecasting model can represent the current overall operating status of the power grid as "stable," "mild warning," or "severe anomaly," among other levels. Of course, besides the overall operating status of the power grid, the analysis results obtained by the forecasting model can also include other content; this is not limited here and can be selected and set based on the user's actual needs.

[0046] 105. Based on the analysis results, determine the inspection strategy and implement it.

[0047] The inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on inspection time, energy consumption, and equipment health status prediction; the path planning model is a model built based on a deep reinforcement learning algorithm.

[0048] After analyzing the results from the prediction model, the system will then call the path planning module to process the results and obtain the inspection strategy. This path planning module is equipped with a path planning model built on a deep reinforcement learning algorithm. This model generates the optimal inspection path by considering factors such as inspection time, energy consumption, and equipment health status.

[0049] For example, for a power grid area that includes multiple substations and transmission lines, the path planning model plans a path that can cover all key equipment and has the shortest inspection time and lowest energy consumption, based on the health status prediction results of the equipment, the location information of each equipment, and the endurance of the inspection robot.

[0050] Once the inspection strategy is determined, the system will dispatch inspection robots or drones to perform inspection tasks according to the optimal inspection path in the strategy.

[0051] Thus, in this embodiment, by defining the communication protocols with various inspection point devices, accurate collection of power grid operation data is ensured. Data cleaning and integration further improve data quality, providing a reliable data foundation for subsequent prediction and inspection strategy formulation. Simultaneously, the prediction model based on machine learning algorithms can comprehensively analyze the impact of multiple factors on the power grid's operating status, improving the accuracy and reliability of predictions. Furthermore, understanding the overall operating status of the power grid in advance helps in preparing for inspections and addressing potential problems. Moreover, in determining the inspection strategy, the path planning model constructed using deep reinforcement learning algorithms can dynamically generate the optimal inspection path based on actual conditions, effectively reducing inspection time and energy consumption, improving inspection efficiency, and ensuring the comprehensiveness and effectiveness of the inspection work, thereby guaranteeing the safe and stable operation of the power grid.

[0052] This embodiment provides a power grid inspection method. Compared with the prior art, the inspection method of this application can be based on a power grid inspection system. Since the power grid inspection system can directly communicate with each inspection point in the power grid, and the communication protocol of each inspection point is determined based on a distributed architecture during the process of acquiring power grid data, it can ensure that data is collected separately from the equipment at each inspection point based on a distributed architecture. This avoids the process of collecting data from different inspection points through different systems and then summarizing it, which is required in the prior art. Since the data generated by each inspection point is significantly less than the summed data, the power grid inspection process is not limited by interface bandwidth, thus avoiding the problem of processing efficiency being affected by interface constraints during the acquisition of power grid data. Meanwhile, because the entire process of power grid inspection—from handling missing and outlier values ​​in power grid data to predicting and analyzing results based on preset models, and even determining inspection strategies based on the analysis results—is conducted within the power grid inspection system, this avoids the need for different systems or devices to perform data acquisition, data analysis, and prediction based on analysis, as is common in existing technologies. It also avoids data interaction between these systems or devices, thus preventing interface constraints that may exist when interacting via interfaces. This further improves the efficiency of the entire power grid inspection process and solves the problem of interface constraints in communication interactions between multiple systems or devices involved in the inspection process. Furthermore, since power grid data is collected based on the communication protocol of each inspection point, it ensures that the power grid data during inspection encompasses various devices in the power grid, such as substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. This ensures data comprehensiveness and lays the foundation for the accuracy of subsequent power grid inspection strategy formulation. Furthermore, since missing and outlier values ​​in the power grid data are processed before prediction, and the data is integrated based on supplementary data after processing, it is ensured that the data being analyzed is complete data containing equipment status and environmental information before the analysis results are obtained. This further ensures the accuracy of the analysis results obtained after prediction, and also ensures the accuracy of the inspection strategy obtained based on the analysis results.

[0053] Furthermore, in some embodiments, after "step 105, determine the inspection strategy and execute it", the method may also perform the following steps respectively:

[0054] Step 1: When it is determined that there is an anomaly in the power grid based on the analysis results, a power grid anomaly alarm is sent to prompt the user to handle the anomaly.

[0055] During the execution of the inspection strategy, this system monitors and analyzes the results in real time. If certain indicators in the analysis results are found to exceed the preset normal range, such as abnormally high equipment temperature or sudden and significant fluctuations in grid load, it is determined that there is an anomaly in the power grid.

[0056] At this point, the system immediately triggers its alarm mechanism, sending power grid anomaly alerts to relevant personnel, such as maintenance staff and on-duty engineers, via SMS, email, or system push notifications, so that they can promptly understand the anomaly and take appropriate action. In this way, by monitoring and analyzing results in real time and sending anomaly alerts, the system ensures that maintenance personnel can quickly become aware of power grid anomalies and take timely measures to address them, effectively reducing fault handling time and improving the reliability of power grid operation.

[0057] Step 2: Save the power grid data, the integrated power grid data, and the analysis results.

[0058] This system stores power grid data, integrated power grid data, and analysis results on different storage media. During the caching phase, the high-speed read / write performance of Redis database is used to temporarily cache data for rapid response to data query and processing requests. For data requiring long-term storage, PostgreSQL database is used for persistent storage to ensure data integrity and security. Furthermore, to meet the needs of efficient analysis of large-scale data, ClickHouse database is used for columnar storage, facilitating complex data analysis and report generation.

[0059] This approach, combining multiple databases for data storage, fully leverages the strengths of each. Redis provides fast data access, PostgreSQL ensures long-term secure data storage, and ClickHouse column-oriented storage meets the needs of efficient analysis of large-scale data, providing a solid foundation for comprehensive data application and decision support.

[0060] Step 3: Localize the abnormal data in the real-time collected power grid data through edge computing nodes, and when it is determined that there is a risk of equipment failure after processing, modify the inspection strategy through an adaptive resource scheduling algorithm in order to reallocate the inspection priority of the inspection points.

[0061] The adaptive resource scheduling algorithm is based on a mixed integer linear programming algorithm, with the goal of minimizing task delay time and resource idle rate, and determines the inspection priority of inspection points based on constraints; the constraints include at least one of the following: the endurance of the inspection robot, the working range of the inspection robot, and the emergency level of the faulty equipment.

[0062] In this system, edge computing nodes are deployed close to inspection points, enabling them to receive and process power grid data collected by inspection equipment in real time. When an edge computing node detects anomalies in the data, such as abnormal fluctuations in equipment operating parameters, it performs localized analysis to determine if there is a risk of equipment failure. If a failure risk is identified, the edge computing node invokes an adaptive resource scheduling algorithm. This algorithm is based on mixed-integer linear programming, with the optimization objective of minimizing task delay time and resource idle rate. Under constraints, including the inspection robot's endurance, working range, and the urgency level of the faulty equipment, the inspection priority of the inspection point can be recalculated and determined. For example, for faulty equipment with a high urgency level, its inspection priority is increased, and the inspection robot is scheduled to inspect it first.

[0063] By combining edge computing nodes with adaptive resource scheduling algorithms, dynamic optimization of inspection resources is achieved. During the inspection process, inspection priorities can be adjusted in a timely manner based on equipment failure risks, ensuring that limited inspection resources are used most effectively, improving the targeting and efficiency of inspection work, and further guaranteeing the safe and stable operation of the power grid.

[0064] Furthermore, in some embodiments, after "step 101, detecting whether a user instruction has been received" in the aforementioned embodiment, the method further includes:

[0065] When the current mode is determined to be manual mode, if a stop inspection command is received, a confirmation message is output to prompt the user whether to confirm the execution of the stop inspection command.

[0066] When feedback information is received and it is determined based on the feedback information that the stop inspection instruction needs to be executed, the inspection strategy determination behavior to be executed is stopped or the operation based on the inspection strategy is stopped.

[0067] In this embodiment, the system has both automatic and manual operating modes. During operation, the system monitors the current mode status in real time. When the current mode is detected to be manual, it begins listening for commands to stop the inspection.

[0068] During this process, users can trigger the suspension of the inspection operation through the manual console in the system interface or by sending specific API commands. For example, in the system interface, the user clicks the "Stop Inspection" button and enters confirmation information in the pop-up confirmation dialog box.

[0069] Upon receiving a stop inspection command, the system first outputs a confirmation message to ensure the user intends to perform the operation. This confirmation can be presented to the user via pop-up notifications, SMS verification codes, or other methods. When the user confirms the stop inspection command again, the system takes appropriate stop actions based on the current execution status of the inspection task. If the inspection strategy has not yet started, the system directly cancels the task scheduling; if the inspection equipment has already started performing inspection operations, the system sends a stop command to the inspection equipment, causing it to stop the current inspection task and safely return to its initial position or standby state.

[0070] The method described in this embodiment allows users to suspend inspection operations at any time in manual mode according to actual needs, providing flexible control over system operation. This is particularly useful in scenarios such as handling emergencies, equipment maintenance, or testing, improving the system's adaptability and controllability. Furthermore, this embodiment utilizes a command confirmation mechanism to prevent inspection interruptions or equipment malfunctions caused by misoperation. It ensures that the suspension of inspection operations is performed with the user's explicit intent, guaranteeing the safe and stable operation of the system and protecting the inspection equipment and power grid from unnecessary interference and damage.

[0071] Furthermore, in some embodiments, "step 101, detecting whether a user instruction has been received" in the aforementioned embodiments can be specifically executed as follows:

[0072] First, output the power grid type information to prompt the user to select the current power grid type. The power grid type information includes substations, transmission lines, distribution transformers, etc.

[0073] Then, upon receiving feedback information about the power grid type, the current power grid type is determined based on the feedback information, and a corresponding inspection plan is determined based on the current power grid type, and the corresponding inspection operation is executed.

[0074] In this example, the system automatically outputs power grid type information upon startup. Specifically, it displays common power grid type options, such as substations, transmission lines, and distribution transformers, through a graphical interface. Users can then select the appropriate option based on the type of power grid they wish to inspect. For example, when inspecting a large substation, a user can select the "Substation" type from the drop-down menu and click "Confirm."

[0075] After receiving feedback information about the power grid type selected by the user, the system proceeds to the corresponding inspection plan determination module. This module matches the inspection plan corresponding to the selected power grid type based on a pre-configured inspection strategy library. This inspection plan can be understood as outlining the behaviors and operations required during the inspection process for different power grid types. For example, for substations, the inspection plan might include detailed inspection steps, inspection frequency, and required inspection equipment types for key equipment such as transformers and circuit breakers. Based on the inspection plan, the system schedules appropriate inspection resources, such as inspection robots and drones, to execute the inspection operations according to the plan.

[0076] In this embodiment, corresponding inspection plans can be formulated according to different power grid types, ensuring that the inspection work is more in line with actual needs and improving the pertinence and effectiveness of the inspection. For example, for substation-type power grids, the focus is on the operating status and internal structure of the equipment; for transmission line-type power grids, the focus is on checking the external environment and insulation of the lines. At the same time, by pre-determining inspection plans that match the power grid type, the system can rationally allocate inspection resources, avoiding resource waste and over-inspection, thus improving inspection efficiency, reducing inspection costs, and ensuring the comprehensiveness and quality of power grid inspection work.

[0077] Furthermore, in some embodiments, the aforementioned embodiment's "step 105, determining and executing the inspection strategy based on the analysis results" can also be executed in one or more of the following ways, including:

[0078] Step A: Determine the maintenance plan based on the operating status of the power equipment.

[0079] This system predicts equipment maintenance needs based on operational status indicators of the power equipment in the analysis results, such as equipment operating time, failure frequency, and performance degradation trend. For equipment with poor operating conditions and high failure risk, the system automatically generates a maintenance plan, including maintenance time, maintenance content, and required spare parts.

[0080] For example, when the transformer oil temperature is consistently high and the insulation performance deteriorates, this system predicts that the transformer may fail in the near future. Therefore, it schedules maintenance for the next maintenance window, including replacing the insulating oil and inspecting the windings.

[0081] By developing maintenance plans, timely maintenance can be carried out on equipment at risk of failure, thereby extending equipment lifespan and improving equipment reliability and operating efficiency.

[0082] Step B: Optimize power distribution based on line load conditions.

[0083] Based on the line load conditions, this system adjusts the power distribution strategy. Specifically, by monitoring parameters such as current and voltage of each line, it identifies lines that are overloaded or underloaded. For overloaded lines, the system takes measures such as adjusting transformer taps and transferring some load to other lines to achieve a reasonable power distribution.

[0084] By adjusting power distribution and equipment operating parameters, it is possible to effectively balance the grid load, improve power quality, reduce grid losses, and ensure that the grid operates in a highly efficient and stable state.

[0085] Step C: Adjust the operating parameters of the power equipment according to the ambient temperature and weather conditions.

[0086] Based on the ambient temperature and weather conditions, this system will automatically adjust the operating parameters of the power equipment. For example, in hot weather, it will appropriately reduce the load rate of the transformer and increase the operating power of the heat dissipation equipment; in cold weather, it will adjust the heating devices of the equipment to prevent the equipment from experiencing performance degradation or malfunction due to low temperatures.

[0087] Step D: Generate improvement measures based on power quality.

[0088] In this embodiment, the system can also analyze power quality data from grid data, such as voltage deviation, frequency fluctuation, and harmonic content, and generate corresponding improvement measures. For areas with voltage deviation problems, the system recommends installing reactive power compensation devices or regulating transformer taps; for areas with excessive harmonic content, it recommends installing filters and other equipment.

[0089] Step E: Determine the security protection strategy based on the security monitoring status.

[0090] Based on security monitoring data, such as equipment alarm information and the security status of the surrounding environment, this system determines security protection strategies. For example, installing this intelligent security system around a substation will automatically activate audible and visual alarms and notify security personnel when unauthorized intrusion is detected.

[0091] This allows for the enhancement of the security protection level of the power grid and equipment based on security protection strategies, effectively preventing external threats and safety accidents, and ensuring the safety of power grid operation.

[0092] Step F: Generate a periodic inspection strategy based on the grid load and electricity demand, wherein the electricity demand is predicted based on the user-side demand response model.

[0093] This system can generate periodic inspection strategies based on the predicted results of power grid load and electricity demand. Electricity demand prediction can be achieved through a user-side demand response model, which takes into account various influencing factors such as user electricity consumption habits, production plans, and weather conditions.

[0094] For example, for the power grid of an industrial park, before the summer peak electricity consumption period, this system predicts that the electricity load will increase significantly. Therefore, it arranges to increase the frequency of inspections in advance, focusing on checking the heat dissipation of equipment, the current carrying capacity of power supply lines, etc., to ensure the safe and stable operation of the power grid during high load periods.

[0095] Based on this, by generating a periodic inspection strategy, the inspection work can be planned and anticipated. At the same time, the generation of this periodic inspection strategy takes into account the power demand and grid load, which can detect potential problems in advance, prevent the occurrence of faults, and improve the overall operational stability and power supply reliability of the power grid.

[0096] Furthermore, in some embodiments, the "step 102, determining the communication protocol with each inspection point in the power grid based on a distributed service architecture" in the aforementioned embodiments may include the following during execution:

[0097] First, determine the equipment type at each inspection point, and then determine the communication protocol for each equipment type based on the equipment type. The communication protocol includes at least the front-end communication protocol, data transmission and reception format, disconnection handling method, and status monitoring parameters.

[0098] Then, a communication channel is established with each of the inspection points based on the communication protocol, and data interaction is performed with each of the inspection points based on the communication channel.

[0099] This system has a pre-configured mapping table between device types and communication protocols. When communication needs to be established with an inspection point in the power grid, the system first identifies the device type of that inspection point. Specifically, the identification method can be obtained through the device's identification information, model parameters, or by conducting an initial handshake communication with the device. Then, according to the mapping table, the system matches the communication protocol corresponding to that device type. The communication protocol includes at least the front-end communication protocol (such as TCP / IP, UDP, etc.), the data transmission and reception format (such as XML, JSON, etc.), the disconnection handling method (such as the number of reconnections, reconnection interval, etc.), and the status monitoring parameters (such as the heartbeat packet sending frequency, connection timeout time, etc.).

[0100] After determining the communication protocol, the system then establishes a communication channel with the inspection point equipment according to the parameters and methods specified in the protocol. For example, for equipment using the TCP / IP protocol, the system creates a connection with the equipment through network socket programming.

[0101] After establishing a connection, the system interacts with the equipment at the inspection points. The system sends data request commands to the equipment according to the data transmission and reception format specified in the communication protocol, and the equipment returns the corresponding power grid data based on the commands. Simultaneously, the system monitors the status of the communication channel in real time. If a connection loss occurs, it performs a reconnection operation according to the connection handling method specified in the protocol, ensuring the continuity and stability of data interaction.

[0102] Based on the above process, the method in this embodiment can adapt to equipment from different manufacturers and of different models at the inspection points, achieving seamless communication with various types of equipment and improving the system's versatility and scalability. Simultaneously, through clear communication protocol specifications and an effective disconnection handling mechanism, the stability and reliability of data interaction are ensured, avoiding data loss or inspection work interruptions due to communication problems, and ensuring the integrity and timeliness of inspection data.

[0103] Furthermore, in some embodiments, the "step 2, saving the power grid data, the integrated power grid data, and the analysis results" in the aforementioned embodiments is specifically executed according to the following steps, which may be:

[0104] On one hand, Redis is used as a cache database to cache power grid data, integrated power grid data, and analysis results. This system utilizes Redis's high-performance caching capabilities to temporarily store real-time collected power grid data, integrated power grid data, and analysis results in Redis. For example, when inspection equipment uploads a new batch of power grid operation data, the system first stores it in the Redis cache to quickly respond to subsequent data query and processing requests. Because the Redis cache database provides extremely fast data read and write speeds, this meets the system's need for rapid access to real-time data, improving the system's response speed and processing efficiency.

[0105] On the other hand, PostgreSQL is used as the persistent database to store power grid data, integrated power grid data, and analysis results. For data requiring long-term storage, this system employs PostgreSQL. Before writing data to PostgreSQL, necessary data processing and formatting are performed to ensure data consistency and integrity. For example, equipment operating parameters and environmental information are stored in corresponding PostgreSQL tables according to a pre-defined table structure. Using PostgreSQL as the persistent database ensures long-term secure data storage, supports complex data queries and transaction processing, and provides a solid foundation for applications such as statistical analysis and historical data tracing.

[0106] On the other hand, the system uses ClickHouse as a columnar storage database to store power grid data, integrated power grid data, and analysis results. To meet the needs of efficient analysis of large-scale data, this system utilizes ClickHouse for columnar storage. During storage, data is organized and stored column-wise according to its characteristics and analytical requirements. For example, data of the same type from different inspection points (such as equipment temperature) can be stored in the same column, facilitating aggregation analysis and trend calculations. Storing data using ClickHouse's columnar storage database ensures efficient data compression and fast query performance when processing large-scale data, making it particularly suitable for complex operations such as data aggregation and trend analysis. This also provides a powerful tool for deeply mining data value and supporting decision-making.

[0107] Furthermore, in some embodiments, the Redis database, the PostgreSQL database, and the ClickHouse database in the foregoing embodiments are all configured with a dual-active architecture, wherein the dual-active architecture includes an active node and a standby node; wherein the active node and the standby node back each other up when processing data, and when the active node is in an invalid state, the standby node replaces the active node to respond to data storage operations, and when the active node resumes operation; therefore, an invalid state includes a fault state, a maintenance state, or an upgrade state.

[0108] In this embodiment, the Redis, PostgreSQL, and ClickHouse databases of this system are all deployed using a dual-active architecture. Each database contains one active node and one standby node, deployed on different physical servers and connected via a high-speed network to achieve real-time data synchronization. When the active node is running normally, all read and write requests are sent to it, and the active node synchronizes the data operations to the standby node in real time. For example, when the Redis active node receives a data write request, it completes the data storage locally and simultaneously sends the updated data information to the standby node via the network, allowing the standby node to synchronously update its own data.

[0109] In addition, this system monitors the operational status of active nodes in real time, determining their validity through heartbeat checks and connection tests. When an active node fails, undergoes maintenance, or is upgraded, the system automatically switches client requests to a standby node. For example, if a PostgreSQL active node becomes unavailable due to hardware failure, the standby node detects the failure and automatically becomes the active node, resuming read and write requests to ensure the continuity of data storage services.

[0110] Thus, in this embodiment, the deployment of an active-active architecture eliminates the risk of single point of failure. When the active node becomes invalid, the backup node can quickly take over the service, ensuring high availability of data storage services and uninterrupted operation of the power grid inspection system. Simultaneously, data is synchronized in real time between the active and backup nodes, guaranteeing data consistency and integrity. Even in the event of a failure in the active node, the backup node possesses the latest data copy, avoiding the risk of data loss and improving data security and reliability.

[0111] Furthermore, in some embodiments, the aforementioned embodiment "step 104, predicting the integrated power grid data based on the prediction model to obtain the analysis results" may include the following when executed:

[0112] First, using historical sample data and combining machine learning algorithms, a model for data analysis of the power grid is constructed to obtain a predictive model; wherein, each piece of data in the historical sample data contains at least monitoring data from substation monitoring equipment, monitoring data from transmission line monitoring equipment, monitoring data from distribution transformer monitoring equipment, monitoring data from power quality monitoring equipment, monitoring data from fault indicators, equipment status, and environmental information;

[0113] Then, based on the prediction model, a prediction operation is performed on the integrated power grid data to obtain the analysis results, wherein the analysis results also include at least one of the following: abnormal data point analysis results, power grid operation trends, and power grid security risks;

[0114] Furthermore, based on the federated learning framework, and combining historical inspection data with feedback results obtained after each execution of the inspection strategy, the prediction model is incrementally updated to optimize the accuracy of the analysis results.

[0115] In building the predictive model, this system first collects a large amount of historical sample data. This data includes monitoring data from substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators, as well as equipment status and environmental information. Then, machine learning algorithms, such as random forests, support vector machines, and neural networks, are used to train the historical sample data to build a predictive model for power grid data analysis. During training, the model's parameters and hyperparameters are adjusted to optimize its performance and improve prediction accuracy and recall.

[0116] After constructing and training the prediction model, this system inputs the integrated power grid data into the trained model. Based on the characteristics of the input data, the prediction model calculates and outputs analysis results, including the overall operating status of the power grid, analysis results of abnormal data points, power grid operating trends, and power grid safety risks. For example, the prediction model may output information such as the load trend of a transmission line in the next 24 hours and the probability of equipment failure in a substation, providing a basis for the formulation of inspection strategies.

[0117] Furthermore, in this embodiment, an incremental update process can be implemented based on the model. Specifically, based on a federated learning framework, the predictive model's parameters are incrementally updated by combining historical inspection data and feedback results obtained after each execution of the inspection strategy. During the federated learning process, each participant (such as different inspection areas or equipment groups) trains and updates the model locally, then encrypts and uploads the updated model parameters to the central server. The central server aggregates the collected model parameters, generates new global model parameters, and distributes them to all participants. In this way, the accuracy of the predictive model is continuously optimized, enabling it to better adapt to changes in the power grid's operating state.

[0118] In this way, the predictive model built using machine learning algorithms can deeply mine patterns in historical data, accurately predict the operating status and potential problems of the power grid, and provide a scientific basis for formulating reasonable inspection strategies in advance, thereby improving the reliability and safety of power grid operation. Furthermore, the introduction of a federated learning framework enables incremental updates to the model, fully utilizing the data resources of all participants while protecting data privacy, and continuously optimizing the model's parameters and performance. This allows the predictive model to adapt to the dynamic changes in power grid operation, maintain high predictive accuracy, and enhance the system's intelligence and decision support capabilities.

[0119] Furthermore, in some embodiments, the aforementioned embodiment "Step 1, when it is determined that there is an anomaly in the power grid based on the analysis results, send a power grid anomaly alarm" includes:

[0120] The analysis results will be output in the form of a report file, which will include at least one of the following: current power grid status assessment, anomaly warning, and safety optimization recommendations.

[0121] In this embodiment, the system analyzes the power grid's operating status in real time. When an anomaly is detected, such as equipment failure or parameter exceeding limits, a report generation mechanism is immediately triggered. The report generation process can extract relevant data from the analysis results, such as the name, location, anomaly type, and occurrence time of the abnormal equipment.

[0122] After extracting the relevant data, it is necessary to organize this data into a structured report file according to a preset report template. This report file should include at least a current power grid status assessment, anomaly warnings, and safety optimization recommendations. For example, the report might describe in detail an abnormal rise in the winding temperature of a transformer, assess its potential impact on power grid operation, issue an early warning signal, and propose corresponding safety optimization recommendations, such as increasing heat dissipation measures or reducing load.

[0123] After generating the report, the report file can be output and pushed in various ways. This system supports saving report files to local storage devices in common formats such as PDF, Excel, and Word, making it convenient for users to view and archive. Simultaneously, the system can automatically push report files to relevant maintenance and management personnel via email, instant messaging tools, and other channels, ensuring they receive timely information on power grid anomalies and handling suggestions, enabling rapid response and action.

[0124] As described above, the power grid inspection method of this embodiment can present abnormal power grid conditions in a structured report format, enabling users to quickly and intuitively understand the problem, facilitating problem location and analysis, improving the efficiency of fault handling, and the safety optimization suggestions included in the report can provide professional guidance for operation and maintenance personnel, helping them to formulate reasonable countermeasures, reducing decision-making risks, and improving the safety and stability of power grid operation.

[0125] To achieve the above objectives, according to another aspect of this application, an embodiment of this application also provides a power grid inspection device, the device including a storage medium; and one or more processors, the storage medium being coupled to the processors, the processors being configured to execute program instructions stored in the storage medium; the program instructions, when executed, perform the power grid inspection method described above.

[0126] Furthermore, as a response to the above Figure 1 In addition to the implementation of the methods described in the above embodiments, another embodiment of this application also provides a power grid inspection device. This power grid inspection device embodiment corresponds to the foregoing method embodiments. For ease of reading, this power grid inspection device embodiment will not repeat the details of the foregoing method embodiments one by one, but it should be clear that the system in this embodiment can correspondingly implement all the contents of the foregoing method embodiments. Specifically, as follows... Figure 3 As shown, this is applied to a power grid inspection system, which directly communicates with each inspection point in the power grid. The power grid inspection device includes:

[0127] The detection unit 21 can be used to detect whether a user command has been received;

[0128] The determining unit 22 can be used to determine the communication protocol with each inspection point in the power grid based on the distributed service architecture if the detection unit 21 detects that a user instruction has been received, and to collect power grid data through the communication protocol. The power grid data is collected based on multiple inspection points in the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators.

[0129] The integration unit 23 can be used to process missing and outlier values ​​in the power grid data obtained by the determination unit 22, and integrate the processed power grid data with supplementary data to obtain integrated power grid data; wherein, the supplementary data is a dataset to make the processed power grid data more accurate; the supplementary data includes equipment status and environmental information; the equipment status includes the operating status of power equipment, line load, ambient temperature, power quality and safety monitoring status; the environmental information includes weather conditions and geographical location;

[0130] Prediction unit 24 can be used to predict the integrated power grid data of integration unit 23 based on the prediction model to obtain analysis results; wherein, the analysis results include at least the comprehensive operating status of the power grid; the prediction model is a data analysis model built based on machine learning algorithms;

[0131] The execution unit 25 can be used to determine and execute an inspection strategy based on the analysis results obtained by the prediction unit 24; wherein, the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on the prediction of inspection time, energy consumption and equipment health status; the path planning model is a model constructed based on a deep reinforcement learning algorithm.

[0132] Furthermore, such as Figure 3 As shown, the device further includes:

[0133] The alarm unit 26 can be used to send a power grid anomaly alarm when it is determined that there is an anomaly in the power grid based on the analysis results obtained by the prediction unit 24, so as to prompt the user to handle the anomaly.

[0134] The storage unit 27 can be used to store the power grid data obtained by the determination unit 22, the power grid data integrated by the integration unit 23, and the analysis results obtained by the prediction unit 24.

[0135] The adjustment unit 28 can be used to perform localized processing on abnormal data in the power grid data collected in real time by the determination unit 22 through edge computing nodes. When it is determined after processing that there is a risk of equipment failure, the inspection strategy is modified by an adaptive resource scheduling algorithm to reallocate the inspection priority of the inspection points. The adaptive resource scheduling algorithm can be used to minimize the task delay time and resource idle rate based on a mixed integer linear programming algorithm, and determine the inspection priority of the inspection points based on constraints. The constraints include at least one of the following: the endurance of the inspection robot, the working range of the inspection robot, and the emergency level of the faulty equipment.

[0136] Furthermore, such as Figure 3 As shown, the device further includes:

[0137] The abort unit 29 can be used to output confirmation information when the current mode is determined to be manual mode, so as to prompt the user whether to confirm the execution of the abort inspection command; and when feedback information is received and it is determined based on the feedback information that the abort inspection command needs to be executed, to abort the inspection strategy determination behavior to be executed by the execution unit 25 or to abort the operation based on the inspection strategy.

[0138] Furthermore, such as Figure 3 As shown, the detection unit 21 can specifically be used to output power grid type information to prompt the user to select the current power grid type. The power grid type information includes substations, transmission lines, distribution transformers, etc.; and, after receiving feedback information on the power grid type information, it determines the current power grid type based on the feedback information, determines the corresponding inspection plan based on the current power grid type, and executes the corresponding inspection operation.

[0139] Furthermore, such as Figure 3 As shown, the execution unit 25 can also be used to determine a maintenance plan based on the operating status of the power equipment;

[0140] The execution unit 25 can also be specifically used to optimize power distribution according to line load conditions;

[0141] The execution unit 25 can also be used to adjust the operating parameters of the power equipment according to the ambient temperature and weather conditions;

[0142] The execution unit 25 can also be specifically used to generate improvement measures based on power quality;

[0143] The execution unit 25 can also be specifically used to determine a security protection strategy based on the security monitoring status;

[0144] The execution unit 25 can also be used to generate a periodic inspection strategy based on the power grid load and electricity demand, wherein the electricity demand is obtained based on the prediction of the user-side demand response model.

[0145] Furthermore, such as Figure 3 As shown, the determining unit 22 can be specifically used to determine the equipment type of each inspection point, and determine the communication protocol with each equipment type based on the equipment type. The communication protocol includes at least a front-end communication protocol, data transmission and reception format, disconnection handling method, and status monitoring parameters; and establish a communication channel with each inspection point based on the communication protocol, and perform data interaction with each inspection point based on the communication channel.

[0146] Furthermore, such as Figure 3 As shown, the storage unit 27 can be specifically used to cache power grid data, integrated power grid data, and analysis results using a Redis database as a cache database; and to store power grid data, integrated power grid data, and analysis results using a PostgreSQL database as a persistent database; and to store power grid data, integrated power grid data, and analysis results using a ClickHouse database as a columnar storage database.

[0147] Furthermore, such as Figure 3 As shown, the Redis database, the PostgreSQL database, and the ClickHouse database are all configured using a dual-active architecture, which includes one active node and one standby node. The active node and the standby node back each other up when processing data, and when the active node is in an invalid state, the standby node takes over the data storage operation and responds to the data storage operation in place of the active node. When the active node resumes operation, the standby node will respond to the data storage operation in place of the active node. Therefore, the invalid state includes a fault state, a maintenance state, or an upgrade state.

[0148] Furthermore, such as Figure 3As shown, the prediction unit 24 can specifically be used to construct a model for data analysis of the power grid using historical sample data and machine learning algorithms, thereby obtaining a prediction model. Each piece of data in the historical sample data includes at least monitoring data from substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, fault indicator monitoring data, equipment status, and environmental information. Furthermore, based on the prediction model, a prediction operation is performed on the integrated power grid data to obtain analysis results. These analysis results include at least one of the following: abnormal data point analysis results, power grid operation trends, and power grid safety risks. Finally, based on a federated learning framework, combined with historical inspection data and feedback results obtained after each inspection strategy execution, the prediction model is incrementally updated to optimize the accuracy of the analysis results.

[0149] Furthermore, such as Figure 3 As shown, the alarm unit 26 can be used to output the analysis results in the form of a report file, which includes at least one of the following: current power grid status assessment, anomaly warning, and safety optimization suggestions.

[0150] Furthermore, in this embodiment, the inspection strategy can also be adjusted based on the equipment health status, which can specifically be an equipment health status indicator. The determination process of this instruction can be based on a multimodal data fusion algorithm. Specifically, the multimodal data fusion algorithm generates the equipment health status indicator through comprehensive processing of operating parameters, image data, and environmental parameters. The specific steps are as follows:

[0151] (1) Extraction of time-series features of operating parameters: For operating parameters with time-series characteristics (such as voltage, current, power, etc.), time series analysis methods (such as ARIMA model) are used to extract their trend, seasonality and periodicity features to generate equipment operating state vectors. For example, assuming the operating parameter is x(t), the equipment operating state vector S=[s1,s2,…,sn] is calculated by ARIMA model, where si represents different feature values, reflecting the operating trend and stability of the equipment over a period of time.

[0152] (2) Image Data Defect Detection: Convolutional Neural Networks (CNNs) are used to process the acquired equipment images, automatically learn the features in the images, and output the probability of defects on the equipment surface. For example, for images of transformer heat sinks, the CNN model can detect whether there are abnormalities such as heat sink deformation or damage, and give the corresponding abnormality probability P.

[0153] (3) Correlation between environmental parameters and historical failure cases: A knowledge graph is used to establish a correlation between environmental parameters (such as temperature, humidity, wind speed, etc.) and historical equipment failure cases. For example, excessively high temperatures and humidity may lead to a decrease in equipment insulation performance, increasing the risk of failure. By analyzing the correlation between environmental parameters and failure cases in the knowledge graph, the environmental risk coefficient R is calculated.

[0154] (4) Calculation of equipment health status index: The operating status vector S, the abnormal probability P and the environmental risk coefficient R are weighted and fused to calculate the equipment health status index H.

[0155] The formula is: H = w1S + w2P + w3R;

[0156] w1, w2, and w3 are weighting coefficients that can be set according to the importance of different data. This indicator comprehensively considers the equipment's operating parameters, image information, and environmental factors to more accurately assess the equipment's health status and provide a scientific basis for formulating inspection strategies.

[0157] With the optimized multimodal data fusion algorithm, the inspection work can more accurately locate equipment problems, rationally allocate inspection resources, improve inspection efficiency, and ensure the safe and stable operation of the power grid.

[0158] This application provides a power grid inspection method and apparatus, applied to a power grid inspection system. The system directly communicates with each inspection point within the power grid. This application can detect whether a user command has been received. If received, based on a distributed service architecture, it determines the communication protocol with each inspection point in the power grid and collects power grid data through the communication protocol. This power grid data is collected from multiple inspection points within the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. Then, it processes missing and outlier values ​​in the power grid data and integrates the processed data with supplementary data to obtain integrated power grid data. The supplementary data is used to make the processed power grid data more accurate. The system requires an accurate dataset; the supplementary data includes equipment status and environmental information; equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status; environmental information includes weather conditions and geographical location; subsequently, the integrated power grid data is predicted based on a prediction model to obtain analysis results; the analysis results include at least the comprehensive operating status of the power grid; the prediction model is a data analysis model built based on machine learning algorithms; finally, based on the analysis results, an inspection strategy is determined and executed; the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on inspection time, energy consumption, and equipment health status prediction; the path planning model is a model built based on deep reinforcement learning algorithms, thereby realizing the power grid inspection function. Compared with existing technologies, the inspection method of this application can be based on a power grid inspection system. Since this power grid inspection system can communicate directly with each inspection point in the power grid, and the communication protocol of each inspection point is determined based on a distributed architecture during the process of acquiring power grid data, it can be ensured that data is collected separately from the equipment at each inspection point based on a distributed architecture. This avoids the process of collecting data from different inspection points through different systems and then summarizing it, which is required in existing technologies. Since the data generated by each inspection point is significantly less than the summed data, the power grid inspection process is not constrained by interface bandwidth, thus avoiding the problem of processing efficiency being affected by interface constraints during the acquisition of power grid data.Meanwhile, because the entire process of power grid inspection—from handling missing and outlier values ​​in power grid data to predicting and analyzing results based on preset models, and even determining inspection strategies based on the analysis results—is conducted within the power grid inspection system, this avoids the need for different systems or devices to perform data acquisition, data analysis, and prediction based on analysis, as is common in existing technologies. It also avoids data interaction between these systems or devices, thus preventing interface constraints that may exist when interacting via interfaces. This further improves the efficiency of the entire power grid inspection process and solves the problem of interface constraints in communication interactions between multiple systems or devices involved in the inspection process. Furthermore, since power grid data is collected based on the communication protocol of each inspection point, it ensures that the power grid data during inspection encompasses various devices in the power grid, such as substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. This ensures data comprehensiveness and lays the foundation for the accuracy of subsequent power grid inspection strategy formulation. Furthermore, since missing and outlier values ​​in the power grid data are processed before prediction, and the data is integrated based on supplementary data after processing, it is ensured that the data being analyzed is complete data containing equipment status and environmental information before the analysis results are obtained. This further ensures the accuracy of the analysis results obtained after prediction, and also ensures the accuracy of the inspection strategy obtained based on the analysis results.

[0159] This application provides a storage medium that includes a stored program, wherein the program, when running, controls the device where the storage medium is located to execute the power grid inspection method described above.

[0160] Storage media may include non-permanent memory in the form of computer-readable media, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0161] This application also provides a power grid inspection device, the device including a storage medium and one or more processors, the storage medium being coupled to the processors, the processors being configured to execute program instructions stored in the storage medium; the program instructions, when executed, perform the power grid inspection method described above.

[0162] This application provides a device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps: detecting whether a user instruction has been received; if received, determining a communication protocol with each inspection point in the power grid based on a distributed service architecture, and collecting power grid data through the communication protocol. The power grid data is collected from multiple inspection points within the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators; processing missing and outlier values ​​in the power grid data, and integrating the processed power grid data with supplementary data to obtain integrated power grid data; wherein, the supplementary data is used to make the processed power grid data more powerful and efficient. A more accurate dataset of grid data is provided; the supplementary data includes equipment status and environmental information; equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status; the environmental information includes weather conditions and geographical location; predictions are made on the integrated grid data based on a prediction model to obtain analysis results; wherein, the analysis results include at least the comprehensive operating status of the grid; the prediction model is a data analysis model built based on machine learning algorithms; based on the analysis results, an inspection strategy is determined and executed; wherein, the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on inspection time, energy consumption, and equipment health status prediction; the path planning model is a model built based on deep reinforcement learning algorithms.

[0163] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing program code with the following initialization steps: detecting whether a user instruction has been received; if received, determining a communication protocol with each inspection point in the power grid based on a distributed service architecture, and collecting power grid data through the communication protocol, wherein the power grid data is collected based on multiple inspection points within the power grid, the inspection points including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators; processing missing values ​​and outliers in the power grid data, and integrating the processed power grid data with supplementary data to obtain integrated power grid data; wherein the supplementary data is for making the processed power grid data more... An accurate dataset is required; the supplementary data includes equipment status and environmental information; equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status; environmental information includes weather conditions and geographical location; predictions are made on the integrated power grid data based on a predictive model to obtain analysis results; wherein the analysis results include at least the comprehensive operating status of the power grid; the predictive model is a data analysis model built based on machine learning algorithms; based on the analysis results, an inspection strategy is determined and executed; wherein the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on inspection time, energy consumption, and equipment health status predictions; the path planning model is a model built based on deep reinforcement learning algorithms.

[0164] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create an implementation for the process. Figure 1 One or more processes and / or boxes Figure 1The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0165] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0166] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0167] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0168] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0169] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0170] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0171] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for inspecting a power grid, characterized in that, The method, applied to a power grid inspection system that directly communicates with each inspection point in the power grid, includes: Check if a user command has been received; If received, the communication protocol with each inspection point in the power grid is determined based on the distributed service architecture, and power grid data is collected through the communication protocol. The power grid data is collected based on multiple inspection points in the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. The process involves processing missing and outlier values ​​in power grid data, and then integrating the processed power grid data with supplementary data to obtain integrated power grid data. The supplementary data is a dataset designed to make the processed power grid data more accurate. The supplementary data includes equipment status and environmental information. Equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status. Environmental information includes weather conditions and geographical location. The integrated power grid data is predicted based on a predictive model to obtain analysis results; the analysis results include at least the comprehensive operating status of the power grid; the predictive model is a data analysis model built based on machine learning algorithms. Based on the analysis results, an inspection strategy is determined and executed; wherein, the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on inspection time, energy consumption, and equipment health status prediction; the path planning model is a model built based on a deep reinforcement learning algorithm; The detection of whether a user instruction has been received includes: Output power grid type information to prompt the user to select the current power grid type, wherein the power grid type information includes substation, transmission line, and distribution transformer; Upon receiving feedback information regarding the power grid type, the current power grid type is determined based on the feedback information, and a corresponding inspection plan is determined based on the current power grid type, and the corresponding inspection operation is executed.

2. The method according to claim 1, characterized in that, After determining and implementing the inspection strategy, the method further includes: When an anomaly is determined to exist in the power grid based on the analysis results, a power grid anomaly alarm is sent to prompt the user to handle the anomaly. Save the power grid data, the integrated power grid data, and the analysis results; Abnormal data in real-time collected power grid data is processed locally by edge computing nodes. When a risk of equipment failure is determined after processing, the inspection strategy is modified by an adaptive resource scheduling algorithm to reallocate the inspection priority of inspection points. The adaptive resource scheduling algorithm is based on a mixed integer linear programming algorithm to minimize task delay time and resource idle rate, and determines the inspection priority of inspection points based on constraints. The constraints include at least one of the following: the endurance of the inspection robot, the working range of the inspection robot, and the emergency level of the faulty equipment.

3. The method according to claim 1, characterized in that, After detecting whether a user instruction has been received, the method further includes: When the current mode is determined to be manual mode, if a stop inspection command is received, a confirmation message is output to prompt the user whether to confirm the execution of the stop inspection command. When feedback information is received and it is determined based on the feedback information that the stop inspection instruction needs to be executed, the inspection strategy determination behavior to be executed is stopped or the operation based on the inspection strategy is stopped.

4. The method according to claim 1, characterized in that, The step of determining and implementing inspection strategies based on analysis results also includes: The maintenance plan is determined based on the operating status of the power equipment. And / or, Optimize power distribution based on line load conditions; And / or, Adjust the operating parameters of power equipment according to ambient temperature and weather conditions; And / or, Based on power quality generation and improvement measures; And / or, Determine security protection strategies based on security monitoring status; And / or, Based on the power grid load and electricity demand, a periodic inspection strategy is generated, wherein the electricity demand is predicted based on the user-side demand response model.

5. The method according to claim 1, characterized in that, The distributed service architecture determines the communication protocol with each inspection point in the power grid, including: The equipment type at each inspection point is determined, and a communication protocol for each equipment type is determined based on the equipment type. The communication protocol includes at least a front-end communication protocol, data transmission and reception format, disconnection handling method, and status monitoring parameters. A communication channel is established with each of the inspection points based on the communication protocol, and data interaction is performed with each of the inspection points based on the communication channel.

6. The method according to claim 2, characterized in that, The process of saving the power grid data, the integrated power grid data, and the analysis results includes: The power grid data, the integrated power grid data, and the analysis results are cached using a Redis database. The power grid data, the integrated power grid data, and the analysis results are stored using a PostgreSQL database as a persistent database. The ClickHouse database is used as a columnar storage database to store power grid data, integrated power grid data, and analysis results.

7. The method according to claim 6, characterized in that, The Redis database, the PostgreSQL database, and the ClickHouse database are all configured with a dual-active architecture, which includes one active node and one standby node. The active node and the standby node back each other up when processing data, and when the active node is in an invalid state, the standby node takes over the data storage operation and responds to the data storage operation in place of the active node. When the active node resumes operation, the standby node will respond to the data storage operation in place of the active node. Therefore, the invalid state includes a fault state, a maintenance state, or an upgrade state.

8. The method according to claim 1, characterized in that, The analysis results obtained by predicting the integrated power grid data based on the prediction model include: By utilizing historical sample data and combining it with machine learning algorithms, a model for data analysis of the power grid is constructed to obtain a predictive model; wherein, each data point in the historical sample data contains at least monitoring data from substation monitoring equipment, monitoring data from transmission line monitoring equipment, monitoring data from distribution transformer monitoring equipment, monitoring data from power quality monitoring equipment, monitoring data from fault indicators, equipment status, and environmental information; Based on the prediction model, the integrated power grid data is used to perform prediction operations to obtain analysis results, wherein the analysis results also include at least one of the following: abnormal data point analysis results, power grid operation trends, and power grid security risks; Based on the federated learning framework, the prediction model is incrementally updated by combining historical inspection data and feedback results obtained after each execution of the inspection strategy, so as to optimize the accuracy of the analysis results.

9. The method according to claim 2, characterized in that, When it is determined that there is an anomaly in the power grid based on the analysis results, sending a power grid anomaly alarm includes: The analysis results will be output in the form of a report file, which will include at least one of the following: current power grid status assessment, anomaly warning, and safety optimization recommendations.

10. A power grid inspection device, characterized in that, An application in a power grid inspection system, wherein the power grid inspection system directly communicates with each inspection point in the power grid, the device includes: The detection unit is used to detect whether a user command has been received. The determining unit is used, if received, to determine the communication protocol with each inspection point in the power grid based on a distributed service architecture, and to collect power grid data through the communication protocol. The power grid data is collected based on multiple inspection points in the power grid, including substation monitoring equipment, transmission line monitoring equipment, distribution transformer monitoring equipment, power quality monitoring equipment, and fault indicators. An integration unit is used to process missing and outlier values ​​in the power grid data, and integrate the processed power grid data with supplementary data to obtain integrated power grid data. The supplementary data is a dataset designed to make the processed power grid data more accurate. The supplementary data includes equipment status and environmental information. Equipment status includes the operating status of power equipment, line load, ambient temperature, power quality, and safety monitoring status. Environmental information includes weather conditions and geographical location. The prediction unit is used to predict the integrated power grid data based on the prediction model and obtain the analysis results; wherein the analysis results include at least the comprehensive operating status of the power grid; the prediction model is a data analysis model built based on machine learning algorithms; An execution unit is used to determine and execute an inspection strategy based on the analysis results; wherein, the inspection strategy includes an optimal inspection path; the optimal inspection path is obtained by a path planning model based on inspection time, energy consumption, and equipment health status prediction; the path planning model is a model built based on a deep reinforcement learning algorithm; The detection unit is specifically used to output power grid type information to prompt the user to select the current power grid type. The power grid type information includes substation, transmission line, and distribution transformer. After receiving feedback information on the power grid type information, the unit determines the current power grid type based on the feedback information, determines the corresponding inspection plan based on the current power grid type, and executes the corresponding inspection operation.

11. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the storage medium to perform the power grid inspection method according to any one of claims 1 to 9.

12. A power grid inspection device, characterized in that, The device includes a storage medium; and one or more processors, the storage medium being coupled to the processors, the processors being configured to execute program instructions stored in the storage medium; the program instructions, when executed, perform the power grid inspection method according to any one of claims 1 to 9.

Images