Substation inspection strategy generation method

Abstract

The invention relates to the technical field of substation monitoring, in particular to a substation inspection strategy generation method. Comprising the following steps: acquiring monitoring video information in a transformer substation in real time, and identifying abnormal transformer substation equipment and abnormal behaviors; if identifying that the substation equipment is on fire, acquiring equipment information and environment information of primary and secondary equipment near the on-fire equipment; based on a graph neural network, capturing a dependency relationship and an interaction state between the ignition equipment and each piece of primary and secondary equipment, analyzing an abnormal behavior of the equipment, and determining a corresponding substation inspection strategy in combination with environmental information; and if oil contamination at the top of the substation equipment is identified, determining an oil leakage starting point, calling historical video data of the oil contamination equipment in a preset time period to perform oil leakage identification, obtaining an oil leakage monitoring result, and determining a corresponding risk monitoring strategy. Through real-time monitoring and analysis, the abnormal condition of the substation equipment is positioned in time, an effective routing inspection and risk monitoring strategy is formulated, and the routing inspection efficiency and safety are effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of substation monitoring, and more specifically, to a method for generating a substation inspection strategy. Background Art

[0002] With the development of social economy and the advancement of science and technology, the stability and safety of power systems have become particularly important. As a key link in the power system, the operating status of substations directly affects the safety and stability of the power grid. In recent years, with the advancement of digital transformation, the traditional method of relying on manual inspections has gradually failed to meet the requirements of modern power grids for efficiency and safety. Therefore, automated and intelligent substation monitoring systems have become a development trend. Among them, 360-degree panoramic video monitoring technology is favored because it can provide all-round, full-process, and multi-perspective real-time monitoring capabilities, and has become an important technical means to promote the construction of digital substations.

[0003] However, although the application of 360-degree panoramic video monitoring technology in substations has achieved certain results, it still faces some challenges in actual application. Mainly manifested in the following aspects: First, in terms of intelligent analysis and recognition, the existing system still has certain limitations. Secondly, due to the limitation of the monitoring angle of view, some areas may be blocked, resulting in the inability to fully capture the detailed information of the equipment. Furthermore, for foreign object detection, the accuracy of existing technology needs to be improved. Finally, in the face of emergencies such as fire and intrusion, there is room for further improvement in the response speed and recognition accuracy of the monitoring system.

[0004] In response to the above problems, the invention patent with publication number CN114069844A discloses a method for substation secondary equipment status image recognition based on deep learning. The method includes obtaining relevant pictures, encoding and classifying the identification and status of relevant components in the substation secondary equipment, and locating safety hazards and fault points by accurately identifying and coupling the signs and status of the air switches, pressure plates and signal indicator lights of the secondary equipment. It can be seen that this solution uses a deep learning model to automatically identify equipment failures in substations, but it mainly focuses on static identification of equipment and does not involve how to deal with real-time monitoring problems in dynamic scenarios.

[0005] The invention patent with publication number CN117641275A discloses a substation environment monitoring method, device, equipment and storage medium. The device integrates multiple sensors and can transmit data to the control center via a wireless network. However, this technology mainly focuses on the monitoring of environmental parameters and does not involve the field of video surveillance, especially panoramic video surveillance technology and its related intelligent analysis functions. It can be seen that although the above two patents have improved the automation level of substation monitoring to a certain extent, they still have shortcomings in solving the above-mentioned technical difficulties. Therefore, it is particularly necessary to develop a new generation of panoramic video surveillance system that can significantly improve intelligent analysis, recognition accuracy and emergency response. Summary of the invention

[0006] In view of this, the present application provides a method for generating a substation inspection strategy, which uses real-time video monitoring and intelligent analysis technology to timely detect and locate abnormal conditions of substation equipment, and formulate effective inspection and risk monitoring strategies based on this to improve inspection efficiency and safety.

[0007] A method for generating a substation inspection strategy, comprising:

[0008] Acquire monitoring video information in the substation in real time, identify abnormal substation equipment and abnormal behaviors, including fire or oil stains on the top; if a fire is detected in the substation equipment, obtain the equipment information and environmental information of each primary and secondary equipment located near the burning equipment; capture the dependency and interaction status between the burning equipment and each primary and secondary equipment based on the graph neural network, and analyze the abnormal behaviors of each primary and secondary equipment; determine the corresponding substation inspection strategy based on the abnormal behaviors and environmental information, including key inspection targets and inspection tasks; if oil stains are detected on the top of the substation equipment, determine the starting point of the oil leakage, and call the historical video data of the oil pollution equipment within a preset time period for oil leakage identification, and obtain the oil leakage monitoring results; according to the oil leakage monitoring results, determine the corresponding risk monitoring strategy, including oil leakage risk monitoring and risk alarm for oil pollution equipment.

[0009] In one possible implementation, the monitoring video information in the substation is obtained by the inspection equipment configured in the substation, including a panoramic camera device, a visible light camera device, a dual-spectrum camera device, and a main transformer interval camera device.

[0010] In one possible implementation, abnormal substation equipment and abnormal behavior are identified based on a preset fault identification model, and the fault identification model is a deep learning neural network model based on an AI identification algorithm;

[0011] Generating the fault identification model includes:

[0012] Collect video data with smoke, flames, foreign objects, oil stains and personnel violations, and annotate the video data, including the location and time of smoke and flames, the type of foreign objects, oil stains, and personnel violations;

[0013] Build deep learning neural network models;

[0014] The deep learning neural network model is trained using the annotated video data, and the model parameters are iteratively optimized using a back propagation algorithm to generate the fault recognition model.

[0015] In one possible implementation, identifying the type of substation equipment and abnormal behavior that occurs in the monitoring video information includes:

[0016] Extracting feature information from monitoring video information through multi-layer convolution and pooling layers of the fault recognition model;

[0017] The feature information is classified through the fully connected layer of the fault recognition model to identify the type of substation equipment and abnormal behavior that occurs in the monitoring video information.

[0018] In one possible implementation, the device information of each primary and secondary device corresponding to the fire device includes basic device information, fire detection information, location information and operation status information; the environmental information includes temperature information and smoke particle size information.

[0019] In one possible implementation, determining a corresponding substation inspection strategy according to the device information and the environment information includes:

[0020] Based on the equipment information and environmental information, key inspection targets are determined, including primary and secondary equipment with abnormal operating status and / or smoke particle size exceeding a preset threshold;

[0021] The environmental information corresponding to the key inspection targets is compared with the preset fire information to determine the fire type and danger level, and uploaded to the background system to determine the corresponding inspection tasks, including appearance, temperature and voiceprint monitoring of key inspection targets.

[0022] One possible implementation method for identifying oil contamination on the top of substation equipment includes:

[0023] Collecting image data of the top of the substation equipment from the monitoring video information, and preprocessing the image data, including image denoising and image enhancement;

[0024] Based on the pre-processed image data, the suspected oil pollution area is determined, and the color, texture and shape characteristics of the suspected oil pollution area are extracted to determine whether it is top oil pollution.

[0025] In one possible implementation, determining the starting point of oil leakage of substation equipment includes:

[0026] According to the morphological information and position information of the top oil stain, the oil leakage starting point corresponding to the oil stain is determined; wherein the morphological information includes the oil stain area and diffusion trend.

[0027] In one possible implementation, historical video information of the oil pollution equipment within a preset time period is called to identify oil leakage and obtain the oil leakage monitoring result, including:

[0028] Through continuous frame difference technology, the historical video information of the oil pollution equipment within a preset time period is identified to obtain the dynamic change information of the oil pollution formation;

[0029] The oil leakage monitoring result is determined according to the dynamic change information, including the oil leakage position, the oil leakage time information and the oil leakage speed.

[0030] In one possible implementation, a corresponding risk monitoring strategy is determined based on the oil leakage monitoring results, including:

[0031] Assess the oil leakage risk level of oil-contaminated equipment based on the oil leakage monitoring results;

[0032] According to the oil leakage risk level, determine the risk monitoring cycle to regularly monitor the oil leakage status of oil pollution equipment;

[0033] When the monitored oil pollution range exceeds the second preset threshold, the risk warning mechanism is triggered to indicate that there is an abnormality in the oil pollution equipment.

[0034] Compared with the prior art, the technical solution of this application has the following beneficial effects:

[0035] By acquiring the monitoring video information in the substation in real time, the operating status of the substation equipment can be continuously monitored, abnormal situations can be discovered in time, and the monitoring efficiency and response speed can be improved. By using the preset fault identification model, abnormal situations in the monitoring video information can be automatically identified, reducing the workload of manual inspections and improving the accuracy of identification. When a fire is detected in the substation equipment, the system will obtain the primary and secondary equipment information and environmental information near the burning equipment, so as to determine a refined inspection strategy, including key inspection targets and inspection tasks, to improve the pertinence and effectiveness of the inspection. For equipment with oil pollution on the top, the starting point of the oil leakage can be determined, and the oil leakage can be identified by calling the historical video information within the preset time period, which is helpful for early detection of oil leakage problems and preventing potential safety risks. According to the oil leakage monitoring results, the risk monitoring strategy can be determined, and the oil leakage risk monitoring and risk alarm can be carried out on the oil pollution equipment, which is helpful for taking preventive measures in time to avoid accidents. Through the analysis of equipment information and environmental information, data-driven decision support can be provided to help operation and maintenance personnel formulate more scientific and reasonable inspection strategies and risk monitoring plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0037] Figure 1 A schematic diagram of the structure of a substation monitoring system provided in Example 1 of the present application.

[0038] Figure 2 A flowchart of a substation monitoring method provided in Example 2 of the present application.

[0039] Figure 3 A flow chart of a method for training a fault identification model provided in Example 2 of the present application.

[0040] Figure 4 This is a flow chart of a method for generating a substation fire equipment inspection strategy provided in Example 3 of the present application.

[0041] Figure 5 This is a flow chart of a method for generating an inspection strategy for oil contamination on top of a transformer provided in Example 4 of the present application.

[0042] Figure 6 A flowchart of a method for generating an inspection strategy for workers' illegal crossing of fences provided in Example 5 of the present application. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0044] Embodiment 1

[0045] See also Figure 1 , is a schematic diagram of the structure of a substation monitoring system provided in Example 1 of the present application. Figure 1 As shown in , the substation monitoring system includes but is not limited to various inspection devices configured in the substation. Specifically, it may include a panoramic camera device 1, a visible light camera device 2, a dual-spectrum camera device 3, and a main transformer interval camera device 4. Each camera device is connected to the central control system through a wired or wireless network, and the central control system analyzes the collected data.

[0046] The panoramic camera device 1 is installed at a key position inside the substation to monitor the substation equipment in real time with 360 degrees without blind spots, and is suitable for observing the overall environment. For example, the 360-degree panoramic camera device is used to focus on monitoring reactive equipment and large oil-filled equipment (main transformer, high-voltage reactor) that are prone to fire or cause serious damage to the substation after fire. In some embodiments, the panoramic camera device 1 can also be equipped with a high-resolution camera device to ensure clear capture of images at night.

[0047] The visible light camera devices 2 are installed near various important equipment in the substation, and are specifically used for detailed monitoring of the equipment under daily conditions to provide assistance for the subsequent analysis of the working status of the equipment.

[0048] The dual-spectrum camera device 3 may be a camera device including two imaging modes, visible light and infrared light, and is specifically used to effectively identify hidden dangers such as abnormal heating of substation equipment under different lighting conditions. It is particularly suitable for detecting overheating of electrical equipment and early warning of potential failures.

[0049] The main transformer interval camera device 4 is specifically used to focus on monitoring the main transformer and its surrounding areas to ensure the safe operation of the core equipment. In some embodiments, the main transformer interval camera device 4 can be equipped with intelligent analysis software to automatically identify abnormal situations and report them in a timely manner.

[0050] In addition, when inspecting difficult-to-reach or dangerous areas inside the substation, the drone can be controlled to perform aerial inspections by remote control or pre-programmed flight paths. In some embodiments, the drone can be equipped with a high-definition camera and an infrared imager to perform detailed inspections at different heights, and combined with the inspection results of the drone, in-depth inspections of specific areas can be performed to ensure the comprehensiveness and meticulousness of the inspection work.

[0051] The substation monitoring system provided in Example 1 of the present application comprehensively utilizes a variety of technical means such as panoramic cameras, visible light cameras, dual-spectrum cameras, main transformer interval cameras and drones, effectively improving the automation level of substation equipment inspections and enhancing the ability to adapt to complex environments. At the same time, it provides system support for realizing a substation inspection strategy generation method for which protection is requested in this application.

[0052] Below, the technical solution provided by this application is described in detail in conjunction with specific method embodiments.

[0053] Embodiment 2

[0054] See also Figure 2 , is a flow chart of a method for generating a substation inspection strategy provided in Example 2 of the present application. Figure 2 As shown in , the specific implementation steps of the above method include:

[0055] Step 101: Acquire monitoring video information in the substation in real time.

[0056] Specifically, the monitoring video information in the substation is obtained in real time by the inspection equipment configured in the substation. For details, please refer to the first embodiment, which will not be described here.

[0057] Step 102: using a preset fault identification model, identify the type of substation equipment and abnormal behavior that occurs in the monitoring video information.

[0058] Among them, the preset fault identification model is a deep learning neural network model based on AI recognition algorithm, which is used to dynamically identify smoke, flames, foreign objects, oil stains and human behavior in the substation.

[0059] The construction and training process of the above fault identification model may specifically include:

[0060] Step 1021: collect a large amount of video data with smoke, flames, foreign objects, oil stains, and illegal behaviors of personnel, and annotate them. The annotated data includes the location and time information of smoke and flames, the types of foreign objects, oil stains, illegal behaviors of personnel, etc.

[0061] Step 1022: Build a deep learning neural network model.

[0062] The embodiment of the present application uses a convolutional neural network as a basic model to construct a deep learning neural network model for fire identification, foreign object identification, oil stain identification, and violation identification.

[0063] Step 1023: Use the labeled video data to train the deep learning neural network model.

[0064] Step 1024, using the back propagation algorithm to continuously iteratively optimize the model parameters, so that it can accurately identify fire, foreign matter, oil stains, and violations, and generate a corresponding fault identification model.

[0065] The trained fault recognition model extracts the feature information of the video frames in the monitoring video information through multiple layers of convolution and pooling layers, and then classifies the feature information through the fully connected layer to identify whether there are abnormal behaviors such as substation equipment fire, foreign objects, top oil pollution, personnel violations, etc. in the monitoring video information, and determine the type of substation equipment where the abnormality occurs, so as to facilitate subsequent inspection or maintenance.

[0066] Step 103: If it is identified that the substation equipment is on fire, the equipment information and environmental information of each primary and secondary equipment located near the burning equipment are obtained.

[0067] In a specific application scenario, the panoramic camera device 1 performs preliminary identification as a trigger point, identifies that important equipment in the substation is on fire, and links the corresponding interval visible light camera device 2, dual-spectrum camera device 3, drone and other inspection equipment to carry out special inspections and temperature measurement, and perform multi-dimensional reuse. After comprehensive analysis and confirmation of abnormal conditions based on the fault identification module, the relevant primary and secondary equipment information and environmental information are triggered to upload, providing auxiliary decision-making for strategy generation.

[0068] Specifically, the equipment information of each primary and secondary equipment includes equipment basic information, fire detection information, location information and operation status information. The environmental information includes temperature information, smoke particle size information, gas type monitoring information and gas concentration monitoring information.

[0069] Step 104: Capture the dependency and interaction status between the ignition device and each primary and secondary device based on the graph neural network, and analyze the abnormal behavior of each primary and secondary device.

[0070] Specifically, the technical solution for substation equipment relationship modeling and abnormal behavior prediction based on graph neural network (GNN) can be refined as follows:

[0071] First, a graph structure of substation equipment is constructed, including primary and secondary equipment and the relationship between them. The nodes represent the various devices in the substation, and the edges represent the physical connection or logical relationship between the devices. A feature vector is defined for each node in the graph, including the type, status, historical maintenance record, operation log, etc. of the equipment. The above feature vector is used as the input of GNN to help the GNN model understand the characteristics and status of each device, and obtain a GNN model that can capture the dependencies between nodes and learn the topological structure and attribute information of the equipment. Then, the GNN model is used to learn the device behavior pattern under normal operating conditions. By comparing the monitoring video information with the learned normal pattern, the GNN model can identify abnormal behavior. For example, it can be detected that the behavior of a node (primary / secondary equipment) is significantly different from that of its neighboring nodes, which may indicate that the device is abnormal.

[0072] Step 105: Determine the corresponding substation inspection strategy, including key inspection targets and inspection tasks, based on the acquired abnormal behaviors of the primary and secondary equipment and the corresponding environmental information.

[0073] In some embodiments, providing auxiliary decision-making for strategy generation based on the acquired device information and environment information, generating a corresponding substation inspection strategy may include:

[0074] For equipment that displays abnormal basic information or abnormal operating status information, including equipment failure or performance degradation, fault diagnosis and necessary recovery operations are required. For areas where the temperature is too high, the smoke particle size exceeds the first preset threshold, or the concentration of harmful gases exceeds the standard, measures such as ventilation, isolation, or the use of fire extinguishing equipment are required. In addition, based on the above equipment information and environmental information, substation equipment with the above abnormal conditions can be identified as key inspection targets to facilitate subsequent key inspections of the above substation equipment.

[0075] Specifically, real-time fire monitoring can be performed on key inspection targets. After the key inspection target is detected to be on fire, the environmental information corresponding to the above key inspection target is compared with the preset fire information to determine the type of fire and the corresponding danger level. The determined information is uploaded to the background system. The corresponding inspection tasks are generated by the background system, including but not limited to appearance inspection, performance testing and necessary maintenance operations on key inspection targets. For example, high-definition cameras and infrared thermal imaging cameras are used for appearance and temperature detection. Use a robot patrol system for automated inspections, including data collection and analysis of micro-meteorological stations. Monitor and analyze the voiceprint data of the equipment through a voiceprint monitoring device.

[0076] Through the above substation inspection strategy, the safe operation of the substation is ensured, and abnormal conditions of equipment and environment are discovered and handled in time.

[0077] Step 106: If oil pollution is identified on the top of the substation equipment, determine the starting point of the oil leakage.

[0078] Specifically, to identify the top oil pollution of substation equipment, we must first extract the image data of the top of the substation equipment from the monitoring video information. After obtaining the above image data, perform preprocessing operations such as denoising and enhancement on the image. Segment the suspected oil pollution area from the preprocessed image, and extract the color, texture, shape and other features of the suspected oil pollution area. Use machine learning or deep learning models to determine whether it is top oil pollution based on the extracted features. After confirming the oil pollution, the starting point or leakage point of the oil pollution can also be determined based on the shape of the oil pollution (such as area, diffusion trend) and location information.

[0079] Step 107: call the historical video information of the oil pollution equipment within a preset time period to identify the oil leakage and obtain the oil leakage monitoring result.

[0080] Specifically, when oil pollution is detected on the top of the substation equipment, the historical video data of the equipment within a preset time period is automatically called. The historical video data provided is further used to obtain the dynamic change information of the oil pollution through continuous frame difference or other video analysis technology, and then the specific location, time information and oil leakage speed of the oil leakage are obtained to form a detailed oil leakage monitoring result.

[0081] Step 108: Determine a corresponding risk monitoring strategy based on the oil leakage monitoring result, including oil leakage risk monitoring and risk warning for oil pollution equipment.

[0082] In this embodiment, the oil leakage risk level of the oil pollution equipment is evaluated based on the oil leakage monitoring results. For example, the risk of oil pollution equipment is divided into low risk, medium risk and high risk according to factors such as the amount of oil leakage, speed and duration. Optionally, the classification standard of each level needs to be formulated according to industry norms and specific circumstances, and the embodiments of this application do not make specific limitations. For example, low risk refers to occasional minor leakage that is easy to control. High risk involves frequent and large-scale leakage, which may have a serious impact on the environment.

[0083] After determining the oil leakage risk level of the oil-contaminated equipment, a reasonable risk monitoring cycle is set accordingly. For low-risk equipment, a longer monitoring interval can be set, such as once a quarter or once a month. For high-risk equipment, more frequent inspections are required, such as once a week or even real-time monitoring. This ensures that problems are discovered and handled in a timely manner, while also ensuring the rational use of resources to avoid excessive monitoring. Optionally, the risk monitoring cycle can be formulated according to specific circumstances, and the embodiments of the present application do not make specific limitations.

[0084] During the monitoring process, if the oil pollution range is found to exceed the second preset threshold, it means that there is a serious leakage problem in the oil pollution equipment. At this time, the risk warning mechanism should be activated immediately, including but not limited to sending warning information to relevant personnel to remind the oil pollution equipment of abnormality, recording the details of the incident for subsequent investigation, and taking emergency measures to stop the operation of the equipment to prevent further pollution spread. Alternatively, the site can be cleaned up according to the emergency plan, and it can be evaluated whether the working status of the equipment needs to be adjusted or repaired and replaced to reduce the possibility of similar situations in the future.

[0085] Compared with the prior art, the technical solution provided in the second embodiment of the present application has the following beneficial effects:

[0086] By acquiring the monitoring video information in the substation in real time, the operating status of the substation equipment can be continuously monitored, abnormal situations can be discovered in time, and the monitoring efficiency and response speed can be improved. By using the preset fault identification model, abnormal situations in the monitoring video information can be automatically identified, reducing the workload of manual inspections and improving the accuracy of identification. When a fire is detected in the substation equipment, the system will obtain the primary and secondary equipment information and environmental information near the burning equipment, so as to determine a refined inspection strategy, including key inspection targets and inspection tasks, to improve the pertinence and effectiveness of the inspection. For equipment with oil pollution on the top, the starting point of the oil leakage can be determined, and the oil leakage can be identified by calling the historical video information within the preset time period, which is helpful for early detection of oil leakage problems and preventing potential safety risks. According to the oil leakage monitoring results, the risk monitoring strategy can be determined, and the oil leakage risk monitoring and risk alarm can be carried out on the oil pollution equipment, which is helpful for taking preventive measures in time to avoid accidents. Through the analysis of equipment information and environmental information, data-driven decision support can be provided to help operation and maintenance personnel formulate more scientific and reasonable inspection strategies and risk monitoring plans.

[0087] Embodiment 3

[0088] See also Figure 4 , is a flow chart of a method for generating a patrol strategy for a fire equipment in a substation provided in Example 3 of the present application. Example 3 of the present application describes the patrol strategy generation method provided in the present application in a specific application scenario.

[0089] Taking the fire of important equipment in a substation as an example, the panoramic camera device is used as the trigger point for preliminary identification, and the corresponding interval visible light camera device, dual-spectrum camera device, and drone are linked to carry out special patrols and temperature measurement, and multi-dimensional reuse is performed. After comprehensive analysis and confirmation of abnormal conditions, the audio system is used to carry out personnel evacuation, and relevant primary and secondary equipment information and image uploads are triggered to provide auxiliary decision-making for operation and maintenance personnel. The specific steps may include:

[0090] Use 360-degree panoramic cameras to preliminarily identify, alarm, and continuously monitor reactive equipment, large oil-filled equipment (main transformers, high-voltage reactors) and other fire equipment in the station. Automatically call N cameras around the fire interval to confirm identification and monitor. Call surrounding dual-spectrum cameras for special patrols, temperature measurement, and comprehensive analysis; at the same time, drones can also be called for special patrols and temperature measurement. Call the audio system for alarms and personnel evacuation, and call N cameras in adjacent intervals to conduct special patrols and monitor surrounding equipment. Call related primary and secondary equipment information and perception signals and analyze them. Finally, provide auxiliary decision-making information and images for operation and maintenance personnel.

[0091] Embodiment 4

[0092] See also Figure 5 , is a flow chart of a method for generating an inspection strategy for oil pollution on top of a transformer provided in Example 4 of the present application. Example 4 of the present application describes the inspection strategy generation method provided by the present application in a specific application scenario.

[0093] Taking the oil stains on the top of the transformer as an example, the visible light camera device detects the abnormality and initially identifies it as a trigger point, calls other cameras, panoramic cameras, and drones for multiplexing confirmation, finds the starting point of the oil leakage, calls suspicious video historical images and data, and provides auxiliary decision-making for operation and maintenance personnel. The specific steps may include:

[0094] Use the main transformer interval camera to preliminarily identify the oil stains on the ground, and automatically call the surrounding camera devices for confirmation and further search. Call the panoramic camera device to search and identify the high-altitude area, and call the drone for special patrol and identification. Call the historical images and data of the suspicious video for research and analysis. Finally, provide auxiliary decision-making information and images for operation and maintenance personnel.

[0095] Embodiment 5

[0096] See also Figure 6 , which is a flow chart of a method for generating an inspection strategy for operators' illegal crossing of fences provided in Example 5 of the present application.

[0097] The panoramic camera device is used as the trigger for preliminary identification, and continuous tracking is carried out. The surrounding visible light camera device is used in real time to take close-up photos and confirm the identity. Then the audio is used for alarm reminders. The safety measures and work conditions in the work area are monitored and prevented. Violations are further recorded and found, and errors are corrected in a timely manner. The following steps may be taken:

[0098] The panoramic camera device initially identifies and tracks the person who crosses the fence as a violation. The surrounding cameras are used to confirm the identity of the person and take close-up photos. The audio system is used to issue warnings and reminders. The surrounding cameras are used to focus on monitoring and patrolling the work site and fence. The violation records and detailed information are provided to the operation and maintenance personnel, and errors are corrected in a timely manner.

[0099] Compared with the prior art, the technical solutions provided in Examples 3 to 6 of the present application have the following beneficial effects:

[0100] By integrating high-definition cameras, drones, robots and other intelligent inspection devices, a monitoring network with comprehensive coverage and dynamic response can be built. Not only can it realize mutual calling and resource sharing between devices, but it can also enhance the ability to identify abnormal situations through multi-dimensional data reuse. For example, when an abnormality occurs in a certain area, the surrounding cameras can be mobilized immediately for detailed inspection, while drones can be used for supplementary observation from a high-altitude perspective, and robots can enter some difficult-to-reach places for close inspection. This joint monitoring mode not only improves the accuracy of problem identification, but also gains valuable response time for operation and maintenance personnel, thereby significantly improving the efficiency of fault handling.

[0101] In view of the common production and safety issues in substations, optimizing the usability of software applications is the key to improving the overall management level. The software interface should be designed to be intuitive and easy to operate, so that operation and maintenance personnel can easily access the required information and make decisions quickly based on actual conditions. In addition, by presenting the policy-making authority to operation and maintenance personnel in a more user-friendly way, emergency response strategies can be flexibly formulated or customized management under specific conditions can be carried out.

[0102] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0103] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0104] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for generating a substation inspection strategy, characterized in that: include: Obtain monitoring video information in the substation in real time to identify abnormal substation equipment and abnormal behavior, such as fire or oil pollution on the top; If a fire is detected in a substation device, the device information and environmental information of each primary and secondary device located near the burning device are obtained; the dependency relationship and interaction status between the burning device and each primary and secondary device are captured based on a graph neural network, and the abnormal behavior of each primary and secondary device is analyzed; According to the abnormal behavior and environmental information, determine the corresponding substation inspection strategy, including key inspection targets and inspection tasks; If oil pollution is identified on the top of the substation equipment, the starting point of the oil leakage is determined, and the historical video data of the oil pollution equipment within the preset time period is called to identify the oil leakage and obtain the oil leakage monitoring results; according to the oil leakage monitoring results, the corresponding risk monitoring strategy is determined, including oil leakage risk monitoring and risk alarm for the oil pollution equipment.

2. A method for generating a substation inspection strategy according to claim 1, characterized in that: The monitoring video information in the substation is obtained by the inspection equipment configured in the substation, including panoramic camera devices, visible light camera devices, dual-spectrum camera devices, and main transformer interval camera devices.

3. A method for generating a substation inspection strategy according to claim 1, characterized in that: The abnormal substation equipment and abnormal behavior are identified based on a preset fault identification model, and the fault identification model is a deep learning neural network model based on an AI identification algorithm; Generating the fault identification model includes: Collect video data with smoke, flames, foreign objects, oil stains and personnel violations, and annotate the video data, including the location and time of smoke and flames, the type of foreign objects, oil stains, and personnel violations; Build deep learning neural network models; The deep learning neural network model is trained using the annotated video data, and the model parameters are iteratively optimized using a back propagation algorithm to generate the fault recognition model.

4. A method for generating a substation inspection strategy according to claim 1, characterized in that: Identify abnormal substation equipment and abnormal behavior in surveillance video information, including: Extracting feature information from monitoring video information through multi-layer convolution and pooling layers of the fault recognition model; The feature information is classified through the fully connected layer of the fault identification model to identify abnormal substation equipment and abnormal behaviors in the monitoring video information.

5. A method for generating a substation inspection strategy according to claim 1, characterized in that: The equipment information of each primary and secondary equipment corresponding to the fire equipment includes equipment basic information, fire detection information, location information and operation status information; the environmental information includes temperature information and smoke particle size information.

6. A method for generating a substation inspection strategy according to claim 5, characterized in that: According to the abnormal behavior and environmental information, a corresponding substation inspection strategy is determined, including: Based on the abnormal behavior and environmental information, determine key inspection targets, including primary and secondary equipment with abnormal operating status and / or smoke particle size exceeding a first preset threshold; The environmental information corresponding to the key inspection targets is compared with the preset fire information to determine the fire type and danger level, and uploaded to the background system to determine the corresponding inspection tasks, including appearance, temperature and voiceprint monitoring of key inspection targets.

7. A method for generating a substation inspection strategy according to claim 1, characterized in that: Identify oil contamination on the top of substation equipment, including: Collecting image data of the top of the substation equipment from the monitoring video information, and preprocessing the image data, including image denoising and image enhancement; Based on the pre-processed image data, the suspected oil pollution area is determined, and the color, texture and shape characteristics of the suspected oil pollution area are extracted to determine whether it is top oil pollution.

8. A method for generating a substation inspection strategy according to claim 7, characterized in that: Determine the starting point of oil leakage in substation equipment, including: According to the morphological information and position information of the top oil stain, the oil leakage starting point corresponding to the oil stain is determined; wherein the morphological information includes the oil stain area and diffusion trend.

9. A method for generating a substation inspection strategy according to claim 1, characterized in that: Call the historical video information of the oil pollution equipment within the preset time period to identify the oil leakage and obtain the oil leakage monitoring results, including: Through continuous frame difference technology, the historical video information of the oil pollution equipment within a preset time period is identified to obtain the dynamic change information of the oil pollution formation; The oil leakage monitoring result is determined according to the dynamic change information, including the oil leakage position, the oil leakage time information and the oil leakage speed.

10. A method for generating a substation inspection strategy according to claim 1, characterized in that: According to the oil leakage monitoring results, determine the corresponding risk monitoring strategy, including: Assess the oil leakage risk level of oil-contaminated equipment based on the oil leakage monitoring results; According to the oil leakage risk level, determine the risk monitoring cycle to regularly monitor the oil leakage status of oil pollution equipment; When the monitored oil pollution range exceeds the second preset threshold, the risk warning mechanism is triggered to indicate that there is an abnormality in the oil pollution equipment.

Images