Transformer substation fault detection method and system based on Internet of Things

Abstract

The invention provides a transformer substation fault detection method and system based on the Internet of Things, and belongs to the technical field of fault detection. The method comprises the following steps: acquiring an insulator operation state parameter and a narrow channel environment parameter to determine an insulator aging index, and if the insulator aging index exceeds a corresponding threshold value, marking insulator fault equipment. And meanwhile, a GPS positioning signal and a UWB positioning signal are obtained, so that coarse positioning of insulator fault equipment is realized. And matching model point coordinates of the corresponding insulator fault equipment in combination with the three-dimensional layout model of the transformer substation to adjust a GPS positioning signal and a UWB positioning signal so as to obtain accurate coordinates of the insulator fault equipment. Then, a space correlation graph structure is constructed based on the accurate coordinates, the influence range of the insulator fault equipment on adjacent equipment is clearly defined, a maintenance navigation path from a narrow channel entrance to the insulator fault equipment is planned in combination with the influence range, the maintenance process is greatly optimized, the time for maintenance personnel to search for fault points is shortened, the maintenance efficiency is remarkably improved, and the maintenance cost is reduced. And substation operation and maintenance cost is reduced.

Description

Technical Field

[0001] This disclosure relates to the field of fault detection technology, specifically to a substation fault detection method and system based on the Internet of Things. Background Technology

[0002] Currently, substations, as core nodes of the power system, bear the important responsibility of power distribution and transmission, and their operational stability directly affects the safety and reliability of the power grid. In recent years, with the rapid growth of electricity demand and the continuous expansion of the power grid, the number of substation devices has surged, and the operating environment has become increasingly complex. The timeliness and accuracy of fault detection and location have become crucial to ensuring power supply. Traditional fault detection methods mostly rely on manual inspections or single sensor monitoring, which are insufficient to cope with the complex and ever-changing working scenarios of modern substations, revealing deficiencies in rapid response and accurate location. Furthermore, traditional methods rely on sensor data from fixed locations, lacking a comprehensive understanding of the operating status of equipment within the substation. Especially when a fault occurs, it is difficult to accurately determine the spatial correlation between the faulty equipment and surrounding equipment, causing maintenance personnel to spend a significant amount of time searching for the fault point in complex environments, severely impacting work efficiency and increasing maintenance time and costs.

[0003] The information disclosed in the background section is only for enhancing the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] In view of this, this disclosure provides a substation fault detection method and system based on the Internet of Things, which can improve the accuracy of fault equipment location.

[0005] In a first aspect, embodiments of this application provide a substation fault detection method and system based on the Internet of Things (IoT). The method includes: acquiring operating status parameters of insulator equipment and environmental parameters of a narrow passage in the substation; the operating status parameters include the surface temperature of the insulator equipment, the surface humidity of the insulator equipment, and the amplitude of the insulator equipment; the environmental parameters of the narrow passage include the humidity and wind speed within the narrow passage; determining an insulator aging index based on the operating status parameters and the environmental parameters of the narrow passage; determining whether the insulator aging index exceeds a preset aging index threshold; if it exceeds the preset aging index threshold, marking it as a faulty insulator equipment; and acquiring the GPS positioning signal and UWB positioning signal of the faulty insulator equipment based on wireless positioning technology; acquiring the weights corresponding to the GPS positioning signal and the UWB positioning signal, respectively; and performing a fault detection based on the weights corresponding to the GPS positioning signal and the UWB positioning signal. The system calculates the initial coordinates of the faulty insulator device based on the GPS and UWB positioning signals, assigning appropriate weights. It then acquires a three-dimensional layout model of the substation and matches the model point coordinates corresponding to the faulty insulator device within the model. It determines whether the Euclidean distance between the model point coordinates and the initial coordinates exceeds a preset Euclidean distance threshold. If it does, the weights corresponding to the GPS and UWB positioning signals are adjusted. The precise coordinates of the faulty insulator device are determined based on these adjusted weights. A spatial correlation graph is constructed based on the precise coordinates of the faulty insulator device. The influence range of the faulty insulator device on adjacent devices is determined based on the spatial correlation graph. Finally, a maintenance navigation path from the narrow passage entrance to the faulty insulator device is determined based on the influence range.

[0006] Secondly, embodiments of this application provide an Internet of Things-based substation fault detection system, which includes: an acquisition module, a first determination module, a first judgment module, a second determination module, a matching module, a second judgment module, a third determination module, a construction module, a fourth determination module, and a fifth determination module. The system includes: an acquisition module for acquiring operating status parameters and narrow passage environmental parameters of insulators in the substation; the operating status parameters include the surface temperature, humidity, and vibration amplitude of the insulators; and the narrow passage environmental parameters include humidity and wind speed within the narrow passage. A first determination module is used to determine the insulator aging index based on the operating status parameters and the narrow passage environmental parameters. A first judgment module is used to determine whether the insulator aging index exceeds a preset aging index threshold. If it exceeds the preset aging index threshold, the insulator is marked as faulty, and its GPS and UWB positioning signals are acquired based on wireless positioning technology. A second determination module is used to acquire the weights corresponding to the GPS and UWB positioning signals, and determine the appropriate parameters based on these weights. The initial coordinates of the faulty insulator device; a matching module for establishing a three-dimensional layout model of the substation and matching the model point coordinates corresponding to the faulty insulator device in the three-dimensional layout model of the substation; a second judgment module for judging whether the Euclidean distance between the model point coordinates and the initial coordinates exceeds a preset Euclidean distance threshold. If the judgment exceeds the preset Euclidean distance threshold, the weights corresponding to the GPS positioning signal and the UWB positioning signal are adjusted respectively; a third determination module for determining the precise coordinates of the faulty insulator device based on the adjusted weights corresponding to the GPS positioning signal and the UWB positioning signal; a construction module for constructing a spatial association graph structure based on the precise coordinates of the faulty insulator device; a fourth determination module for determining the influence range of the faulty insulator device on adjacent devices based on the spatial association graph structure; and a fifth determination module for determining the maintenance navigation path from the narrow passage entrance to the faulty insulator device based on the influence range.

[0007] This application provides a substation fault detection method and system based on the Internet of Things (IoT). By acquiring insulator operating status parameters and narrow passage environmental parameters, the insulator aging index is determined and compared with a preset aging index threshold. If the threshold is exceeded, the insulator is marked as faulty. This overcomes the limitations of traditional manual inspection or single-sensor monitoring, accurately identifying insulator faults and improving the timeliness and accuracy of fault detection. Next, the GPS and UWB positioning signals of the faulty insulator are acquired to determine its initial coordinates, achieving coarse positioning. Then, the model point coordinates of the corresponding faulty insulator are matched using a three-dimensional substation layout model. If the Euclidean distance between the initial coordinates and the model point coordinates exceeds a threshold, the GPS and UWB positioning signals are adjusted to obtain the precise coordinates, achieving precise coordinate locking. This solves the shortcomings of traditional positioning technologies, such as low accuracy and the ability to provide only vague area information, overcoming the technical bottleneck of difficulty in pinpointing specific intervals or equipment numbers. Subsequently, a spatial correlation graph structure is constructed based on the precise coordinates, clearly defining the impact range of the faulty equipment on adjacent equipment, compensating for the shortcomings of traditional methods in effectively analyzing the spatial correlation of equipment. Finally, by combining the scope of impact, a maintenance navigation route from the entrance of the narrow passage to the faulty insulator is planned, which greatly optimizes the maintenance process, reduces the time for maintenance personnel to find the fault point, significantly improves maintenance efficiency, and reduces the substation operation and maintenance costs. At the same time, by clarifying the scope of the fault impact and ensuring maintenance safety through efficient navigation, it ultimately provides strong technical support for the stable operation of the substation and the reliable power supply of the power grid. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments or conventional technologies of this disclosure, the accompanying drawings used in the description of the embodiments or conventional technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating an exemplary embodiment of the Internet of Things-based substation fault detection method provided in this application.

[0010] Figure 2 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0011] Figure 3 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0012] Figure 4This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0013] Figure 5 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0014] Figure 6 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0015] Figure 7 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0016] Figure 8 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application.

[0017] Figure 9 This is a flowchart illustrating an IoT-based substation fault detection method provided in another exemplary embodiment of this application. Detailed Implementation

[0018] Currently, substations, as core nodes of the power system, bear the important responsibility of power distribution and transmission, and their operational stability directly affects the safety and reliability of the power grid. In recent years, with the rapid growth of electricity demand and the continuous expansion of the power grid, the number of substation devices has surged, and the operating environment has become increasingly complex. The timeliness and accuracy of fault detection and location have become crucial to ensuring power supply. Traditional fault detection methods mostly rely on manual inspections or single sensor monitoring, which are insufficient to cope with the complex and ever-changing working scenarios of modern substations, revealing deficiencies in rapid response and accurate location. Furthermore, traditional methods rely on sensor data from fixed locations, lacking a comprehensive understanding of the operating status of equipment within the substation. Especially when a fault occurs, it is difficult to accurately determine the spatial correlation between the faulty equipment and surrounding equipment, causing maintenance personnel to spend a significant amount of time searching for the fault point in complex environments, severely impacting work efficiency and increasing substation operation and maintenance time and costs.

[0019] For example, substations have densely distributed equipment and complex spatial layouts. When a fault occurs, a single location technology is insufficient to accurately pinpoint the coordinates of specific equipment. Because some equipment may be located in high-density areas, traditional location methods may only provide vague regional information, failing to pinpoint specific intervals or equipment numbers. This inaccurate location further complicates the analysis of the correlation between faulty equipment and surrounding equipment. Spatial correlation analysis requires considering the relative positions, connections, and operating status of equipment; however, existing methods often lack efficient coordination mechanisms when integrating multi-source spatial information, resulting in an inability to provide maintenance personnel with clear equipment layout information and navigation guidance.

[0020] Therefore, how to quickly determine the precise geographical location of the faulty equipment when a fault occurs, and analyze its spatial correlation with surrounding equipment, so as to provide maintenance personnel with intuitive navigation and layout information, has become a technical problem that needs to be solved in the field of substation fault detection and location.

[0021] This disclosure provides a substation fault detection method based on the Internet of Things, such as... Figure 1 The illustrated method is an IoT-based substation fault detection method. This method may include the following steps: Step S110: Obtain the operating status parameters of the insulator equipment and the environmental parameters of the narrow passage in the substation. The operating status parameters include the surface temperature of the insulator equipment, the surface humidity of the insulator equipment, and the amplitude of the insulator equipment. The environmental parameters of the narrow passage include the humidity and wind speed in the narrow passage. Step S120: Determine the insulator aging index based on the operating status parameters of the insulator equipment and the environmental parameters of the narrow passage; Step S130: Determine whether the aging index of the insulator exceeds the preset aging index threshold. If it is determined that it exceeds the preset aging index threshold, mark it as an insulator fault device, and obtain the GPS positioning signal and UWB positioning signal of the insulator fault device based on wireless positioning technology. Step S140: Obtain the weights corresponding to the GPS positioning signal and the UWB positioning signal respectively, and determine the initial coordinates of the insulator fault device based on the weights corresponding to the GPS positioning signal and the UWB positioning signal. Step S150: Obtain the three-dimensional layout model of the substation and match the model point coordinates of the corresponding insulator fault equipment in the three-dimensional layout model of the substation; Step S160: Determine whether the Euclidean distance between the model point coordinates and the initial coordinates exceeds the preset Euclidean distance threshold. If it is determined that the preset Euclidean distance threshold is exceeded, adjust the weights corresponding to the GPS positioning signal and the UWB positioning signal respectively. Step S170: Determine the precise coordinates of the faulty insulator device based on the weights corresponding to the adjusted GPS positioning signal and the adjusted UWB positioning signal; Step S180: Construct a spatial correlation diagram structure based on the precise coordinates of the faulty insulator equipment; Step S190: Determine the impact range of the faulty insulator equipment on adjacent equipment based on the spatial correlation diagram structure; Step S195: Determine the maintenance navigation path from the entrance of the narrow passage to the faulty insulator based on the scope of impact.

[0022] In the above method, the aging index of the insulator is determined by acquiring the insulator's operating status parameters and the environmental parameters of the narrow passage, and compared with a preset aging index threshold. If the threshold is exceeded, the insulator is marked as faulty. This overcomes the limitations of traditional manual inspection or single-sensor monitoring, accurately identifies insulator faults, and improves the timeliness and accuracy of fault detection. Next, the GPS and UWB positioning signals of the faulty insulator are acquired to determine its initial coordinates, achieving coarse positioning. Then, the model point coordinates of the corresponding faulty insulator are matched using the substation's 3D layout model. If the Euclidean distance between the initial coordinates and the model point coordinates exceeds a threshold, the GPS and UWB positioning signals are adjusted to obtain the precise coordinates, achieving precise coordinate locking. This solves the shortcomings of traditional positioning technologies, such as low accuracy and the ability to provide only fuzzy area information, and overcomes the technical bottleneck of difficulty in pinpointing specific intervals or equipment numbers. Subsequently, a spatial correlation diagram is constructed based on the precise coordinates, clearly defining the impact range of the faulty equipment on adjacent equipment, thus compensating for the shortcomings of traditional methods in effectively analyzing the spatial correlation of equipment. Finally, by combining the scope of impact, a maintenance navigation route from the entrance of the narrow passage to the faulty insulator is planned, which greatly optimizes the maintenance process, reduces the time for maintenance personnel to find the fault point, significantly improves maintenance efficiency, and reduces the substation operation and maintenance costs. At the same time, by clarifying the scope of the fault impact and ensuring maintenance safety through efficient navigation, it ultimately provides strong technical support for the stable operation of the substation and the reliable power supply of the power grid.

[0023] The steps of the IoT-based substation fault detection method provided in this disclosure are described in detail below: In one embodiment of this disclosure, step S110 involves acquiring the operating status parameters of the insulator equipment and the environmental parameters of the narrow passage in the substation. The operating status parameters include the surface temperature of the insulator equipment, the surface humidity of the insulator equipment, and the amplitude of the insulator equipment's vibration. The environmental parameters of the narrow passage include the humidity and wind speed within the narrow passage. Specifically, a multi-source sensor network can be deployed within the substation. For the insulator equipment, three types of sensors are installed on its surface: a temperature sensor for collecting the surface temperature of the insulator equipment, a humidity sensor for collecting the surface humidity of the insulator equipment, and an acceleration sensor for collecting the amplitude of the insulator equipment. Simultaneously, environmental sensors are installed within the narrow passage of the substation to collect the humidity and wind speed within the narrow passage, forming a sensing network covering the insulator equipment itself and its surrounding environment. For example, specific parameters collected at a certain moment include: the surface temperature of the insulator equipment is 85.2℃, the surface humidity of the insulator equipment is 65%, the amplitude of the insulator equipment is 0.8mm, the humidity in the narrow passage is 72.5%, and the wind speed in the narrow passage is 1.2m / s. These data directly reflect the operating status of the insulator equipment and the environmental conditions of the narrow passage, providing a basis for subsequent fault diagnosis.

[0024] In one embodiment of this disclosure, step S120, determining the insulator aging index based on the operating status parameters of the insulator equipment and the narrow passage environment parameters, further includes the following steps: Figure 2 As shown, the specific content is as follows: Step S210: Calculate the deviation between the running status parameter and the preset threshold of the corresponding running status parameter, and use the calculation result as the first deviation value; Step S220: Calculate the deviation between the narrow passage environmental parameters and the preset threshold of the corresponding preset narrow passage environmental parameters, and use the calculation result as the second deviation value; Step S230: Obtain the weights corresponding to the running status parameters and the weights corresponding to the preset narrow channel environment parameters respectively; Step S240: Based on the weights corresponding to the operating status parameters and the weights corresponding to the preset narrow channel environment parameters, perform a weighted calculation on the first deviation value and the second deviation value, and use the weighted calculation result as the insulator aging index.

[0025] Specifically, for example, the preset threshold values ​​for the operating status parameters set by a certain substation are as follows: a preset threshold of 80℃ for the temperature of the insulator surface, a preset threshold of 60% for the humidity of the insulator surface, and a preset threshold of 2mm for the amplitude of the insulator. At a certain moment, the actual operating status parameters collected by the temperature sensor, humidity sensor, and vibration sensor deployed on the surface of the insulator are: a temperature of 85.2℃ for the insulator surface, a humidity of 65% for the insulator surface, and an amplitude of 2.8mm for the insulator. Then, the first deviation values ​​are calculated as follows: the first deviation value corresponding to the temperature of the insulator surface = 85.2℃ - 80℃ = 5.2℃ (exceeding the threshold; high temperature will accelerate the aging of the insulator insulation material and increase the risk of flashover); the first deviation value corresponding to the humidity of the insulator surface = 65% - 60% = 5% (exceeding the threshold; surface dampness can easily lead to increased leakage current and aggravate material corrosion); the first deviation value corresponding to the amplitude of the insulator = 2.8mm - 2mm = 0.8mm (exceeding the threshold; excessive vibration will cause the internal structure of the insulator to loosen and damage the insulation performance).

[0026] For example, a substation sets the preset thresholds for environmental parameters in a narrow passage as follows: humidity 70% and wind speed 1.5 m / s. At a certain moment, the actual environmental parameters collected by the environmental sensor in the narrow passage are: humidity 72.5% and wind speed 1.2 m / s. The second deviation values ​​are then calculated as follows: Humidity deviation = 72.5% - 70% = 2.5% (exceeding the threshold; high humidity prolongs the time the insulator surface remains wet, accelerating aging); Wind speed deviation = |1.2 m / s - 1.5 m / s| = 0.3 m / s (taking the absolute value; wind speeds below the threshold cause heat buildup in the passage, preventing timely removal of heat generated by the insulators and indirectly accelerating insulator aging).

[0027] For example, based on the statistical data of insulator aging failures in a substation over the past 5 years (temperature exceeding the standard accounts for 40% of failures, surface humidity exceeding the standard accounts for 20%, amplitude exceeding the standard accounts for 30%, and environmental factors account for 10%), the following weights can be set: Temperature of the insulator surface corresponds to a weight of 0.4; humidity of the insulator surface corresponds to a weight of 0.2; amplitude of the insulator corresponds to a weight of 0.3; humidity in the narrow passage corresponds to a weight of 0.05; wind speed in the narrow passage corresponds to a weight of 0.05. Then, based on the above weights, the insulator aging index is calculated as follows: (First deviation value × weight of corresponding operating state parameter) + (Second deviation value × weight of corresponding preset narrow passage environmental parameter) = (5.2℃ × 0.4) + (5% × 0.2) + (0.8mm × 0.3) + (2.5% × 0.05) + (0.3m / s × 0.05) = 3.46 This method quantifies the deviation between the operating parameters of insulator equipment and the environmental parameters of narrow passages, achieving a comprehensive understanding of the factors influencing insulator aging and avoiding the limitations of single-dimensional assessment. Furthermore, through weight allocation, it accurately matches the actual impact of different dimensional parameters on aging (e.g., temperature has the highest weight, consistent with the historical pattern that temperature is the primary cause of faults). Finally, it transforms multi-dimensional parameters into a directly quantifiable insulator aging index, providing an objective and quantitative basis for subsequent labeling of faulty insulator equipment, rather than relying on subjective experience. This significantly improves the accuracy and reliability of insulator aging fault judgment, laying a precise foundation for fault identification in the entire IoT-based substation fault detection process.

[0028] In one embodiment of this disclosure, in step S130, it is determined whether the aging index of the insulator exceeds a preset aging index threshold. If it is determined that the aging index exceeds the preset aging index threshold, it is marked as an insulator fault device, and the GPS positioning signal and UWB positioning signal of the insulator fault device are obtained based on wireless positioning technology.

[0029] Specifically, for example, a substation sets a preset aging index threshold of 1.2 based on insulator fault data and equipment standards from the past 5 years. The aging index of a certain insulator is known to be 3.46. Comparing 3.46 with the preset aging index threshold of 1.2, 3.46 > 1.2, meeting the condition of exceeding the preset aging index threshold. Therefore, this insulator is marked as faulty, and a unique fault equipment identifier (such as INS-027) is generated. The marking time and current aging index value are simultaneously recorded and stored in the substation's fault equipment database, providing equipment identification for subsequent location and maintenance.

[0030] For example, a GPS (Global Positioning System) module (supporting both BeiDou and GPS dual-mode, adaptable to both open outdoor areas and semi-enclosed narrow passages in substations) pre-installed on the faulty insulator can receive satellite signals in real time to obtain the GPS positioning signal of the faulty insulator and calculate its latitude and longitude coordinates. Because GPS signals are easily obstructed by metal equipment in the substation, the positioning accuracy is typically coarse, mainly used to determine the approximate location of the faulty device. For instance, if the output GPS positioning signal corresponds to coarse positioning coordinates of 116.45° longitude and 39.92° latitude, with a positioning accuracy of 5m, it indicates that the faulty insulator is roughly located on the west side of the No. 3 narrow passage in the substation.

[0031] For example, a UWB (Ultra-Wideband) anchor points (positioning base stations) can be deployed on both sides of a narrow passage in a substation to form a UWB positioning network. Simultaneously, a UWB tag is pre-installed on each insulator. This tag can transmit UWB signals in real time to communicate with surrounding anchor points and receive feedback signals from the anchor points, providing a signal source for positioning calculations and thus acquiring UWB positioning signals. UWB positioning can provide centimeter / decimeter-level precision, thus compensating for the accuracy deficiencies of GPS positioning. For instance, a UWB anchor point array can be arranged with four nodes spaced 2 meters apart on both sides of a narrow passage. The anchor points must uniformly cover the narrow passage area, ensuring that the UWB tag on the faulty insulator (installed on the surface of the insulator) can establish signal interaction with at least three anchor points, meeting the positioning algorithm's requirement for multi-node signals.

[0032] In the above method, on the one hand, by quantitatively comparing the insulator aging index with the preset aging index threshold, the subjectivity of manual judgment is avoided, ensuring the accuracy of insulator fault equipment marking and reducing the problem of ineffective operation and maintenance caused by mismarking normal equipment or the risk to the power grid caused by missing fault equipment. On the other hand, by simultaneously acquiring GPS positioning signals (coarse positioning, wide range) and UWB positioning signals (fine positioning, high precision) through wireless positioning technology, a coarse-fine complementary positioning signal system is formed. This not only breaks through the accuracy limitations of GPS in narrow channels and metal obstruction scenarios, but also makes up for the limited coverage of UWB anchor points. It provides high-quality raw data for subsequent steps to fuse the dual signals to calculate the initial coordinates of faulty equipment, significantly improving the response speed of substation fault detection and the reliability of early positioning data, and reducing the time cost for subsequent maintenance personnel to find fault points.

[0033] In one embodiment of this disclosure, step S140, which involves obtaining the weights corresponding to the GPS positioning signal and the UWB positioning signal respectively, and determining the initial coordinates of the insulator fault device based on the weights corresponding to the GPS positioning signal and the UWB positioning signal, further includes the following steps: Figure 3 As shown, the specific content is as follows: Step S310: Determine the first coordinates of the faulty insulator device based on the GPS positioning signal; Step S320: Determine the second coordinates of the faulty insulator device based on the UWB positioning signal; Step S330: Based on the TDOA algorithm, calculate the coordinate offset of the second coordinate relative to the first coordinate; Step S340: Correct the first coordinate based on the coordinate offset; Step S350: Obtain the weights corresponding to the GPS positioning signal and the UWB positioning signal respectively; Step S360: Based on the extended Kalman filter algorithm, determine the initial coordinates of the insulator fault device according to the corrected first coordinates, second coordinates, the weights corresponding to the GPS positioning signal and the weights corresponding to the UWB positioning signal.

[0034] Specifically, for example, for a tagged insulator fault device (identified as INS-027), its pre-installed GPS module, after calculation, outputs the following coordinates for the GPS positioning signal: longitude 116.45°, latitude 39.92°, altitude 12.5m. These coordinates are the first coordinates of the insulator fault device. Using these coordinates, it can be preliminarily determined that the faulty device is located in the area west of the narrow passage of substation No. 3, but it is impossible to accurately pinpoint the specific equipment interval or equipment number (because there are 3 insulator devices within the 5m accuracy coverage area).

[0035] For example, four UWB anchor points (numbered UWB-01 to UWB-04) are deployed on both sides of the narrow passage of substation No. 3. The UWB tag of the insulator fault device interacts with the UWB-02 anchor point. Using a bilateral ranging algorithm combined with the TDOA (Time Difference of Arrival) algorithm, the local offset of the UWB tag relative to the UWB-02 anchor point is calculated as follows: -0.32m east (i.e., 0.32m west), +0.18m north, and 0m height. Therefore, the second coordinate of the insulator fault device is: the coordinates of the UWB-02 anchor point superimposed with the local offset, i.e., the longitude deviation (1m corresponds to approximately 0.000009°) corresponding to longitude 116.4503° -0.32m ≈ 116.45027°, the latitude deviation corresponding to latitude 39.9202° +0.18m ≈ 39.9202016°, and the height is 12.5m. These coordinates can pinpoint the location of the faulty equipment to be 0.32m west and 0.18m north of the UWB-02 anchor point on the west side of the No. 3 narrow passage, corresponding to the second interval of the No. 3 passage.

[0036] Furthermore, the TDOA algorithm is used to establish a correlation between UWB local fine positioning (second coordinates) and GPS global coarse positioning (first coordinates), quantifying the relative positional differences between the two. The core of the TDOA algorithm is to calculate the difference between the second and first coordinates in the three dimensions of longitude, latitude, and altitude to obtain the coordinate offset. This offset reflects the direction and magnitude of the positional correction between UWB fine positioning and GPS coarse positioning. For example, the first coordinates are longitude 116.45°, latitude 39.92°, and altitude 12.5m; the second coordinates are longitude 116.45027°, latitude 39.9202016°, and altitude 12.5m. The coordinate offsets calculated using the TDOA algorithm are as follows: Longitude offset = Second coordinate longitude - First coordinate longitude = 116.45027° - 116.45° = +0.00027° (eastward offset); Latitude offset = Second coordinate latitude - First coordinate latitude = 39.9202016° - 39.92° = +0.0002016° (northward offset); Altitude offset = Second coordinate altitude - First coordinate altitude = 12.5m - 12.5m = 0m. Therefore, the final coordinate offset of the second coordinate relative to the first coordinate is: +0.00027° longitude, +0.0002016° latitude, 0m altitude. The first coordinate is then corrected based on the coordinate offset as follows: Corrected longitude = 116.45° + 0.00027° = 116.45027°; Corrected latitude = 39.92° + 0.0002016° = 39.9202016°; Corrected altitude = 12.5m + 0m = 12.5m.

[0037] For example, the weights corresponding to GPS positioning signals and UWB positioning signals can be obtained separately to avoid the influence of single positioning signal bias on the fusion result. The weights corresponding to GPS positioning signals are: The weights corresponding to the UWB positioning signals are: =0.39. The Extended Kalman Filter (EKF) algorithm can be used to integrate the corrected first and second coordinates along with their corresponding weights, reducing measurement noise interference and outputting more accurate initial coordinates. The EKF algorithm includes a prediction step and an update step. In the prediction step, the corrected first coordinate is used as the initial state, and the coordinates at the next time step are predicted using the state transition matrix in conjunction with the weights corresponding to the GPS positioning signal. In the update step, the second coordinate is used as the observation value, and the Kalman gain is calculated using the Jacobian matrix in conjunction with the weights corresponding to the UWB positioning signal. The predicted coordinates are iteratively updated to obtain stable initial coordinates.

[0038] In the above method, the correlation between GPS positioning coordinates and UWB positioning coordinates is established through the TDOA algorithm, so that the corrected first coordinates have both global and precision attributes. Dynamic weight allocation ensures that the more reliable positioning signal (e.g., a higher weight for a strong UWB signal) plays a dominant role in the fusion process. Finally, the initial coordinates of the faulty insulator are determined using the extended Kalman filter algorithm. This method meets the positioning requirements down to specific equipment intervals or equipment numbers, significantly reducing the time cost for subsequent maintenance personnel to locate the fault point and improving the efficiency and reliability of substation fault detection.

[0039] In one embodiment of this disclosure, step S150 involves obtaining a three-dimensional layout model of the substation and matching the model point coordinates of the corresponding faulty insulator device within the three-dimensional layout model. Specifically, a three-dimensional layout model of the substation can be constructed using BIM (Building Information Modeling). This model integrates information such as the substation's building structure (e.g., narrow passages, equipment bays), equipment parameters (e.g., insulator device numbers and dimensions), and spatial coordinates (e.g., the local three-dimensional coordinates of each device within the three-dimensional layout model). During the construction process, reference to substation design drawings, equipment installation location survey data, and maintenance history is required to ensure consistency between the three-dimensional layout model and the actual scenario. Next, by associating the initial coordinates of the faulty insulator device with the three-dimensional layout model of the substation (local coordinate system, such as x / y / z meter-level coordinates), the model point in the three-dimensional layout model that best matches the actual location of the faulty insulator device (i.e., the digital coordinates of the corresponding faulty device within the model) is found.

[0040] For example, a BIM-format 3D layout model of a 220kV substation can be retrieved from a cloud database. The local coordinate system of this model is set with the center point of the substation's south gate as the origin (0,0,0), the x-axis along the east (main passage direction), the y-axis along the north, and the z-axis along the vertical upward direction. The model stores a precise coordinate library of 120 insulator devices. The theoretical model point coordinates corresponding to INS-027 are (x=45.0m, y=67.5m, z=12.3m), belonging to the second bay of the third narrow passage. The initial coordinates of the faulty insulator device are known to be (longitude 116.45027°, latitude 39.9202016°, altitude 12.5m). Using a Gaussian projection transformation algorithm, these coordinates are converted into matching coordinates in the local coordinate system of the 3D layout model, resulting in matching coordinates of (x=45.1m, y=67.6m, z=12.4m). Based on the coordinates to be matched (x=45.1m, y=67.6m, z=12.4m), insulator model points within the range of "x∈[44.0-46.0m], y∈[66.5-68.5m], z∈[12.0-13.0m]" in the 3D layout model were filtered, resulting in 3 candidate points. Candidate point 1: Equipment number INS-026, coordinates (x=44.8m, y=67.3m, z=12.3m); Candidate point 2: Equipment number INS-027, coordinates (x=45.0m, y=67.5m, z=12.3m); Candidate point 3: Equipment number INS-028, coordinates (x=45.2m, y=67.7m, z=12.4m). The Euclidean distance between the coordinates to be matched (45.1, 67.6, 12.4) and the 3 candidate points can be calculated using the least squares method. The Euclidean distance between the coordinates to be matched and the coordinates of candidate point 1 is: The Euclidean distance between the coordinates of the point to be matched and the coordinates of candidate point 2 is: The Euclidean distance between the coordinates to be matched and the coordinates of candidate point 3 is: It can be seen that the Euclidean distance between the coordinates to be matched and candidate point 3 is the smallest. Therefore, the coordinates of candidate point 3 (x=45.2m, y=67.7m, z=12.4m) are confirmed as the model point coordinates that match the faulty insulator equipment.

[0041] In the above method, a complete spatial reference system for the substation is provided by the three-dimensional layout model of the substation in BIM format, which integrates key layout data such as equipment coordinates and equipment interval information. Then, through coordinate transformation and least squares matching, the initial coordinates of the actual positioning are mapped to the precise model point coordinates within the model, realizing the spatial association of faulty equipment from abstract coordinates to specific equipment numbers and intervals, laying the foundation for subsequent adjustment of GPS and UWB signal weights and improvement of positioning accuracy.

[0042] In one embodiment of this disclosure, in steps S160 to S170, it is determined whether the Euclidean distance between the model point coordinates and the initial coordinates exceeds a preset Euclidean distance threshold. If it is determined that the preset Euclidean distance threshold is exceeded, the weights corresponding to the GPS positioning signal and the UWB positioning signal are adjusted respectively. The precise coordinates of the insulator fault device are determined based on the adjusted weights corresponding to the GPS positioning signal and the UWB positioning signal.

[0043] Specifically, for example, the coordinates of the model points of the insulator fault device are: The initial coordinates of the insulator fault device are Real-time acquisition of UWB signal strength The preset Euclidean distance threshold is 0.5m. The calculated Euclidean distance between the model point coordinates and the initial coordinates is: It can be seen that 0.58 > 0.5, exceeding the preset Euclidean distance threshold. Therefore, the weights corresponding to the GPS positioning signal and the UWB positioning signal are adjusted respectively. The formula for adjusting the weight of the GPS positioning signal is as follows: in, The weights corresponding to the GPS positioning signals. For adjustment coefficients, This is the Euclidean distance between the model point coordinates and the initial coordinates.

[0044] The formula for calculating the weight adjustment of the UWB positioning signal is as follows: in, Weights corresponding to UWB positioning signals. For adjustment coefficients, This represents the real-time UWB signal strength.

[0045] For example, Take 2, Take 0.1, , -70 , If it is 0.58m, then =1 / (1+2.0×0.58)=0.46. Next, after normalization, we get: ; .

[0046] For example, based on the adjusted weights of the GPS positioning signal and the UWB positioning signal, the corrected first and second coordinates can be weighted and fused to generate fused coordinates. This fused coordinates are then generated by combining an extended Kalman filter algorithm to suppress noise, ultimately outputting the accurate coordinates of the faulty insulator. For example, based on the adjusted weights, the x, y, and z coordinates of the corrected first and second coordinates can be weighted and summed to obtain: ; ; .

[0047] For example, the weights corresponding to the adjusted GPS positioning signals The weights corresponding to the adjusted UWB positioning signals =0.46. The corrected first coordinates are (45.3m, 67.9m, 12.6m), and the second coordinates are (45.1m, 67.6m, 12.4m). Then, after weighted calculation, the fused coordinates are obtained as: x=45.208m, y=67.762m, z=15.508m. After one iteration of optimization using extended Kalman filtering, the fused coordinates are updated to (45.18m, 67.72m, 12.48m). The Euclidean distance between the optimized coordinates and the model point coordinates (45.0, 67.5, 12.3) is calculated: As can be seen, 0.34 < 0.5, which does not exceed the preset Euclidean distance threshold of 0.5. Therefore, the final output insulator fault device's precise coordinates are (45.18m, 67.72m, 12.48m), with a positioning accuracy of 0.34m, which meets the maintenance requirements for precision down to the specific device interval.

[0048] In the above method, the deviation between the initial coordinates and the model point coordinates is calculated using Euclidean distance. A preset Euclidean distance threshold is used as the standard to trigger weight adjustments for GPS and UWB signals, ensuring that the weight allocation matches the actual positioning accuracy requirements (e.g., reducing GPS weight and increasing UWB weight when the deviation is large, fully leveraging the advantages of UWB precise positioning). Then, through weighted fusion and extended Kalman filtering, the deviation between the precise coordinates and the model baseline is stabilized within a preset threshold, achieving a leap from coarse positioning to centimeter / decimeter-level precise positioning. Simultaneously, it provides accurate fault location information, significantly reducing the time cost for maintenance personnel to locate fault points and improving the accuracy of substation fault detection and operational efficiency.

[0049] In one embodiment of this disclosure, step S180, constructing a spatial correlation graph structure based on the precise coordinates of the faulty insulator device, includes the following steps: Figure 4 As shown, the specific content is as follows: Step S410: Based on the precise coordinates of the faulty insulator, obtain the associated device identifier and associated device operating status parameters of the devices associated with the faulty insulator; Step S420: Based on graph theory algorithms, construct a spatial association graph structure with the associated device identifiers as nodes, the electrical connection relationships between associated devices as edges, and the associated device operating status parameters as weights.

[0050] Specifically, for example, the precise coordinates of the faulty insulator (identified as INS-027) are (x=45.18m, y=67.72m, z=12.48m). Using these precise coordinates as the center and a radius of 50 meters, a filtering range is set. Combined with electrical connection relationships, information on related equipment is extracted from the substation equipment database using SQL queries. The resulting identification of related equipment and corresponding operating status parameters are: Transformer T-003 (voltage 95kV, current load 80A, equipment temperature 42℃, normal operating status), Disconnecting switch QS-012 (voltage 95kV, current load 80A, equipment temperature 38℃, normal operating status), Current transformer CT-008 (voltage 95kV, current load 78A, equipment temperature 36℃, normal operating status), and surge arrester LA-005 (voltage 95kV, current load 0.5A, equipment temperature 35℃, normal operating status).

[0051] For example, a spatial association graph structure can be constructed based on the weighted undirected graph modeling algorithm in graph theory. First, the node set is determined as {INS-027, T-003, QS-012, CT-008, LA-005}. Then, edges are drawn according to electrical connections: INS-027 and QS-012 are set as edges because they are directly connected in series; QS-012 and T-003 are set as edges because they are directly connected in series; T-003 and CT-008 are set as edges because they are directly connected in parallel; T-003 and LA-005, and INS-027 and T-003 / CT-008 / LA-005 are not set as edges because they have no direct connection. Finally, the current load in the operating status parameters of the associated equipment is used as the weight. The weights of the edges are redistributed as follows: the weights of the edges INS-027 and QS-012 are both 80A current loads (80A), the weights of the edges QS-012 and T-003 are both 80A current loads (80A), and the weights of the edges T-003 and CT-008 are the average of their current loads (80A and 78A) (79A). Finally, the spatial association graph structure is stored in the form of an adjacency matrix. The rows and columns in the adjacency matrix are node identifiers, and the matrix values ​​are the weights of the corresponding edges (0 if there is no connection). In this way, the spatial association graph structure of the insulator fault equipment can be constructed, which can intuitively present the quantitative relationship between the electrical connections and operating status of each device.

[0052] In the above method, by using the precise coordinates of the faulty insulator, spatially adjacent and electrically related devices are accurately screened from the equipment database, ensuring the comprehensiveness and relevance of the related device data. Then, a graph theory algorithm is used to transform the device identifier-electrical connection-operating parameter into a weighted spatial association graph structure, converting the abstract device association relationships into a visualized and quantifiable structured model (such as an adjacency matrix). The presence or absence of edges clarifies the electrical connection relationships between devices, and the weight values ​​quantify the degree of influence of the device's operating status on related devices. This provides support for maintenance personnel to intuitively understand the equipment layout and connection logic around the faulty device, avoiding misjudgments of fault impact caused by information dispersion in traditional analysis. It significantly improves the accuracy and efficiency of substation fault association analysis, providing a scientific basis for subsequent maintenance decisions.

[0053] In one embodiment of this disclosure, step S190, determining the impact range of an insulator fault on adjacent equipment based on the spatial correlation diagram structure, includes the following steps: Figure 5 As shown, the specific content is as follows: Step S510: Based on the breadth-first search algorithm, traverse the spatial association graph structure and calculate the path length and weight accumulation between the faulty insulator device and each associated device. Step S520: Determine the fault correlation coefficient of each associated device based on path length and weight accumulation; Step S530: Determine whether the fault correlation coefficient exceeds the preset correlation coefficient threshold; Step S540: If it is determined that the preset correlation coefficient threshold is exceeded, the currently associated device is marked as a high-risk device, and the first influence range is determined based on the insulator fault device and the high-risk device; Step S550: Based on the depth-first search algorithm, starting from the high-risk device, traverse the spatial association graph structure until the faulty device associated with the high-risk device is traced. Step S560: Calculate the failure propagation probability of the faulty equipment associated with the high-risk equipment; Step S570: Determine whether the fault propagation probability is greater than the preset fault propagation probability threshold; Step S580: If the preset fault propagation probability threshold is exceeded, an alarm is activated; Step S590: If it is determined that the preset fault propagation probability threshold is not exceeded, the impact range of the faulty insulator on adjacent equipment is determined based on the faulty insulator equipment, high-risk equipment, and faulty equipment associated with the high-risk equipment.

[0054] Specifically, for example, in the spatial association graph structure constructed in the aforementioned steps, the insulator fault device is identified as INS-027, and the associated devices include disconnector QS-012, transformer T-003, and current transformer CT-008 (the edge weights are INS-027-QS-012=80A, QS-012-T-003=80A, and T-003-CT-008=79A, respectively). The surge arrester LA-005 has no associated path. Starting with INS-027, a breadth-first search algorithm was used to traverse the spatial correlation graph, resulting in the following path lengths: QS-012 = 1, cumulative weight = 80A; T-003 = 2, cumulative weight = 160A; and CT-008 = 3, cumulative weight = 239A. Based on the fault correlation coefficient (cumulative weight / path length), the correlation coefficients for QS-012, T-003, and CT-008 were calculated to be 80, 80, and 79.67, respectively. These results were compared to a preset correlation coefficient threshold of 75, and all three exceeded this threshold. QS-012, T-003, and CT-008 were then marked as high-risk devices. Based on the relationship between INS-027 and these three high-risk devices, the first influence range was determined to be {INS-027, QS-012, T-003, CT-008}. Starting with high-risk device T-003, a depth-first search algorithm is used to traverse the network, tracing back to the busbar device BUS-002 (edge ​​weight = 82A) associated with T-003 and the voltage transformer PT-004 (edge ​​weight = 82A) associated with BUS-002. Therefore, the faulty devices associated with the high-risk device are BUS-002 and PT-004. Next, the fault propagation probability of the associated faulty devices is calculated using the following formula: in, The probability of fault propagation in a faulty device. For equipment failure rate, Let be the duration of the fault. Then, calculate the fault propagation probability of device BUS-002. The system compares the fault propagation probability with a preset fault propagation probability threshold of 5%, and determines that it does not exceed this threshold. Based on INS-027, high-risk equipment (QS-012, T-003, CT-008), and associated faulty equipment (BUS-002, PT-004), the impact range of the insulator faulty equipment on adjacent equipment is determined to be: {INS-027 (insulator faulty equipment), QS-012 (high-risk equipment), T-003 (high-risk equipment), CT-008 (high-risk equipment), BUS-002 (associated faulty equipment), PT-004 (associated faulty equipment)}. If the propagation probability of an associated equipment is determined to be 6% > 5%, the system immediately pushes an alarm message to the substation maintenance terminal: "Alarm! The fault propagation probability of equipment BUS-002 is 6%, requiring urgent repair of INS-027 and associated high-risk equipment!"

[0055] In the above method, high-risk equipment directly affected is accurately screened by BFS traversal and fault correlation coefficient calculation, avoiding the waste of resources from indiscriminately identifying risky equipment. Then, DFS is used to trace indirectly related equipment, and the risk of diffusion is quantified by combining the fault propagation probability formula, which makes up for the limitation of only focusing on direct impact and ignoring indirect propagation. Finally, by using the dual judgment of preset correlation coefficient threshold and preset fault propagation probability threshold, high-risk equipment is accurately marked and alarms are triggered on demand. This avoids the interference of low-risk equipment not being alarmed and also prevents the omission of high-risk equipment not being alarmed. It significantly improves the accuracy of substation fault risk assessment and the scientific nature of operation and maintenance decision-making, and effectively reduces the risk of power grid outage caused by fault propagation.

[0056] In one embodiment of this disclosure, step S195, determining the maintenance navigation path from the narrow passage entrance to the insulator faulty device based on the scope of influence, includes the following steps: Figure 6 As shown, the specific content is as follows: Step S610: Within the affected area, obtain the coordinates of the narrow passage entrance and the coordinates of the faulty insulator device; Step S620: Based on the A* algorithm, determine the shortest path from the narrow passage entrance to the insulator fault device according to the coordinates of the narrow passage entrance and the insulator fault device; Step S630: Obtain the object height from the entrance of the narrow passage to the intermediate node traversed by the insulator fault device; Step S640: Determine whether the object's height exceeds a preset height threshold; Step S650: If it is determined that the preset height threshold is exceeded, the current intermediate node is replaced to determine the maintenance navigation path from the entrance of the narrow passage to the faulty insulator equipment.

[0057] Specifically, for example, the influence range of the determined insulator fault equipment on adjacent equipment is {INS-027 (insulator fault equipment), QS-012 (disconnecting switch), T-003 (transformer), CT-008 (current transformer), BUS-002 (busbar), PT-004 (voltage transformer)}, and the three-dimensional layout model of the substation adopts a local coordinate system with the center point of the south gate of the substation as the origin and z=0m as the ground. Within this influence range, the coordinates of the entrance of the No. 3 narrow passage are obtained from the passage attribute library of the three-dimensional layout model as (x=10.0m, y=20.0m, z=0.0m), and the precise coordinates of the insulator fault equipment are retrieved as (x=45.18m, y=67.72m, z=12.48m).

[0058] For example, based on the A* algorithm, with a grid resolution of 0.5m, a heuristic function of Euclidean distance, and avoiding the physical occupied area of ​​T-003 (x∈[42.0-48.0m], y∈[65.0-70.0m]), the shortest maintenance navigation path node sequence from the narrow passage entrance to the insulator fault device can be planned as follows: starting point (10.0,20.0,0.0) → intermediate node 1 (20.5,35.2,0.0) → intermediate node 2 (32.8,50.6,0.0) → intermediate node 3 (40.1,60.3,12.48) → ending point (45.18,67.72,12.48). The total length of the maintenance navigation path is calculated to be 63.0m. The object heights at each intermediate node are obtained using a lidar with a sampling frequency of 10Hz and a measurement accuracy of ±0.05m. The intermediate node heights are 0.1m for node 1, 1.9m for node 2, and 0.0m for node 3. Each object height is compared to a preset height threshold of 1.8m, and it is determined that the object height of intermediate node 2 (1.9m) exceeds this threshold. Within the range of x∈[31.8-33.8m] and y∈[49.6-51.6m] around intermediate node 2 (32.8,50.6,0.0), a feasible node (33.0,51.5,0.0) with an object height of 0.5m is selected as the new intermediate node. The node sequence of the replanned maintenance navigation path is as follows: starting point (10.0,20.0,0.0) → intermediate node 1 (20.5,35.2,0.0) → new intermediate node 2 (33.0,51.5,0.0) → intermediate node 3 (40.1,60.3,12.48) → ending point (45.18,67.72,12.48). The total length of the new path is calculated to be 64.2m. It is verified that the object height of all nodes does not exceed the preset height threshold of 1.8m. Finally, this node sequence is determined to be the maintenance navigation path from the entrance of the narrow passage to the insulator fault equipment.

[0059] In the above method, the spatial boundary of the maintenance navigation path is locked by the influence range, ensuring that both the starting point and the ending point are within the fault-related area, thus preventing the path from deviating from the maintenance target. Then, the A* algorithm is used to plan an efficient initial maintenance navigation path, balancing shortest distance and spatial constraints. Subsequently, real-time sensors detect the height of objects at intermediate nodes, and a preset height threshold is used to determine passage safety, preventing path failure due to vertical obstacles. Finally, a node replacement mechanism dynamically optimizes the maintenance navigation path, ensuring that all nodes meet safe passage requirements. This makes the maintenance navigation path both "short" and "safe and feasible," providing maintenance personnel with precise spatial guidance, significantly reducing the pathfinding time from narrow passage entrances to faulty insulators, while avoiding accidental entry into high-risk areas or obstruction by obstacles. This improves the efficiency and safety of substation maintenance and serves as a crucial bridge connecting fault detection and actual operation and maintenance.

[0060] In one embodiment of this disclosure, after determining the maintenance navigation path from the narrow passage entrance to the insulator fault device based on the scope of influence in step S195, the method further includes the following steps: Figure 7 As shown, the specific content is as follows: Step S710: Determine if there are obstacles in the inspection navigation path; Step S720: If an obstacle is determined to exist, the type of obstacle is identified based on the image recognition algorithm, and the location information of the obstacle is output; Step S730: Based on the A* algorithm, redetermine the shortest path from the entrance of the narrow passage to the faulty insulator according to the location information of the obstacle.

[0061] Specifically, for example, the maintenance navigation path node sequence from the entrance of the narrow passage to the faulty insulator is as follows: starting point (10.0m, 20.0m, 0.0m) → intermediate node 1 (20.5m, 35.2m, 0.0m) → new intermediate node 2 (33.0m, 51.5m, 0.0m) → intermediate node 3 (40.1m, 60.3m, 12.48m) → ending point (45.18m, 67.72m, 12.48m). The maintenance navigation path can be monitored in real time using LiDAR and a high-definition industrial camera. An object with a height of 2.2m is detected within the x∈[32.5-33.5m] and y∈[51.0-52.0m] range surrounding the new intermediate node 2 (33.0m, 51.5m, 0.0m). This object exceeds the preset obstacle height range, indicating an obstacle on the maintenance navigation path. Based on the YOLOv5 object detection algorithm (with a preset confidence threshold of 0.85 and IoU threshold of 0.5), the obstacle area captured by the high-definition industrial camera is then analyzed. The obstacle is identified from the domain image and output as a maintenance box (confidence level 0.92). Simultaneously, combined with the LiDAR depth data, the center coordinates of the obstacle's local coordinate system are determined to be (33.1m, 51.6m, 0.0m), the bounding box coordinates of the image pixel coordinate system are (x1=850, y1=480, x2=1050, y2=720), and the obstacle boundary range is x∈[32.6-33.6m], y∈[51.1-52.1m]. This obstacle location information is then output to the path planning module. Based on the A* algorithm, with a grid resolution of 0.5m and Euclidean distance as the heuristic function, and by extending the obstacle boundary outward by 0.5m as the obstacle avoidance zone (with node costs set to infinity within this zone), the maintenance navigation path node sequence is replanned as follows: starting point (10.0m, 20.0m, 0.0m) → intermediate node 1 (20.5m, 35.2m, 0.0m) → new intermediate node 2. * (34.0m, 52.5m, 0.0m) → Intermediate node 3 (40.1m, 60.3m, 12.48m) → End point (45.18m, 67.72m, 12.48m). The total length of the updated maintenance navigation path is calculated to be 65.6m. Finally, this node sequence is determined as the replanned maintenance navigation path.

[0062] In the above method, multi-source sensor collaboration using LiDAR and cameras ensures timely detection of sudden obstacles on the maintenance navigation path. Then, the YOLOv5 image recognition algorithm accurately identifies the obstacle type and location, providing clear spatial constraints for obstacle avoidance planning and preventing blind detours. Finally, the A* algorithm is reused to maintain the shortest maintenance navigation path while avoiding obstacles, ensuring the new path remains within the fault's impact range and does not deviate from the maintenance target. This gives the maintenance navigation path dynamic adaptability, significantly improving the flexibility and safety of the substation maintenance process, reducing maintenance time losses due to sudden obstacles, and further improving the full-process technical support from fault detection to maintenance implementation, ensuring substation operation and maintenance efficiency.

[0063] In one embodiment of this disclosure, after determining the maintenance navigation path from the narrow passage entrance to the insulator fault device based on the scope of influence in step S195, the method further includes the following steps: Figure 8 As shown, the specific content is as follows: Step S810: Obtain historical fault data of the insulator faulty equipment, and determine the first fault feature vector based on the historical fault data of the insulator faulty equipment; Step S820: Obtain the operating status parameters of the faulty insulator equipment, and determine the second fault feature vector based on the operating status parameters of the faulty insulator equipment; Step S830: Calculate the cosine similarity between the first fault feature vector and the second fault feature vector; Step S840: Determine whether the cosine similarity exceeds the preset cosine similarity threshold; Step S850: If the determination exceeds the preset cosine similarity threshold, the historical fault data is clustered based on the K-means clustering algorithm to determine the fault mode corresponding to the current insulator fault device. The fault modes include flashover fault, aging fault, pollution fault, mechanical fault and lightning strike fault.

[0064] Specifically, for example, if the current faulty insulator model is FXBW4-220 / 100, 200 historical fault data points for this model of insulator over the past 3 years are obtained from the substation operation and maintenance database. Three typical aging fault samples are extracted from these samples, based on the following parameters: average surface temperature of the insulator (75℃, standard deviation 8℃); average surface humidity of the insulator (55%, standard deviation 10%); average amplitude of the insulator (1.5mm, standard deviation 0.5mm); and average humidity within narrow passages (65%). With a standardization rule of 8% for the mean wind speed in the narrow passage being 1.0 m / s and the standard deviation being 0.3 m / s, the following first fault feature vectors are obtained: 1 [0.875, 0.7, 1.0, 0.625, -0.667], 2 [0.625, 0.5, 0.6, 0.375, -0.333], and 3 [1.25, 1.0, 1.4, 0.875, -1.0], forming the first fault feature vector set.

[0065] The operating status parameters of the faulty insulator (insulator surface temperature 83℃, insulator surface humidity 63%, insulator amplitude 1.9mm) and the environmental parameters of the narrow passage (narrow passage humidity 69%, narrow passage wind speed 0.85m / s) are obtained. The second fault feature vector is calculated as [1.0, 0.8, 0.8, 0.5, -0.5] using the same standardization rules. The cosine similarity between the first and second fault feature vectors is calculated to be 0.983, and the average cosine similarity between the second and first fault feature vector sets is 0.92. The average cosine similarity of 0.92 is compared to a preset cosine similarity threshold of 0.85, and it is determined that the threshold is exceeded. Based on the K-means clustering algorithm (K=5, corresponding to five fault modes: flashover fault, aging fault, pollution fault, mechanical fault, and lightning strike fault), the 200 first fault feature vectors are clustered to obtain the center vectors of 5 clusters (Cluster 1 flashover fault: [2.5,1.8,0.5,1.2,0.3], Cluster 2 aging fault: [1.0,0.7,0.9,0.6,-0.5], Cluster 3 pollution fault: [0.5,1.5,0.3,1.8,0.1], Cluster 4 mechanical fault: [0.3,0.2,2.0,0.4,0.2], Cluster 5 lightning strike fault: [3.0,0.8,1.2,0.5,2.0]). The Euclidean distance between the second fault feature vector and the center of each cluster is calculated, and the distance to the center of Cluster 2 aging fault is 0.17 (the minimum distance). Finally, the fault mode corresponding to the current insulator fault device is determined to be aging fault.

[0066] In one embodiment of this disclosure, step S850: if it is determined that the cosine similarity threshold is exceeded, then based on the K-means clustering algorithm, the historical fault data is clustered to determine the fault mode corresponding to the current insulator fault device. The fault modes include flashover fault, aging fault, pollution fault, mechanical fault, and lightning strike fault. The following steps are also included: Figure 9 As shown, the specific content is as follows: Step S910: Within the impact range of the insulator failure equipment on adjacent equipment, determine the predicted risk of each adjacent equipment; Step S920: Obtain historical fault data for each adjacent device, and determine the actual risk of the adjacent devices based on the historical fault data of the adjacent devices; Step S930: Based on the gradient descent algorithm, determine the distance threshold between the faulty insulator equipment and adjacent equipment according to the predicted risk and the actual risk; Step S940: Adjust the influence range of the faulty insulator on adjacent equipment according to the distance threshold between the faulty insulator and adjacent equipment.

[0067] Specifically, a risk prediction model can be constructed by acquiring three types of factors related to risk. These include: the failure mode influence coefficient (the intensity of the impact of different failure modes on surrounding equipment; for example, the influence coefficient of aging failure is lower than that of flashover failure, set to 0.8), the spatial distance between adjacent equipment and the faulty insulator (the closer the distance, the higher the risk), and the electrical connection weight between adjacent equipment and the faulty insulator (the edge weights of the spatial correlation graph; the larger the weight, the stronger the electrical correlation, and the higher the risk). Therefore, the predicted risk = failure mode influence coefficient × (1 / spatial distance between adjacent equipment and the faulty insulator) × electrical connection weight.

[0068] For example, given that the fault mode of the insulator faulted device (identified as INS-027) is aging fault (fault mode influence coefficient = 0.8), and its precise coordinates are (x = 45.18m, y = 67.72m, z = 12.48m), the spatial distances (straight-line distances calculated based on the precise coordinates) and electrical connection weights of each adjacent device are as follows: Disconnecting switch QS-012: Spatial distance = 10.0m, Electrical connection weight = 80A → Predicted risk = 0.8 × (1 / 10.0) × 80 = 6.4; Transformer T-003: Spatial distance = 15.0m, Electrical connection weight = 80A → Predicted risk = 0.8 × (1 / 15.0) × 80 ≈ 4.27; Current transformer CT-008: Spatial distance = 20.0m, electrical connection weight = 79A → Predicted risk = 0.8×(1 / 20.0)×79≈3.16; Busbar BUS-002: Spatial distance = 45.0m, Electrical connection weight = 82A → Predicted risk = 0.8 × (1 / 45.0) × 82 ≈ 1.46; Voltage transformer PT-004: Spatial distance = 48.0m, electrical connection weight = 82A → Predicted risk = 0.8×(1 / 48.0)×82≈1.37; The predicted risks for each adjacent device were obtained as follows: QS-012 (6.4), T-003 (4.27), CT-008 (3.16), BUS-002 (1.46), and PT-004 (1.37).

[0069] For example, the distance threshold between a faulty insulator and its adjacent devices (the spatial radius of the previously affected area, such as 50m) can be optimized using the gradient descent algorithm. The core idea is to minimize the error between the predicted and actual risks of adjacent devices, ensuring that the distance threshold covers both high-risk devices and excludes low-risk, distant devices. The mean squared error (MSE) needs to be used to measure this error, as shown in the formula: in, For loss function, The actual risk to adjacent equipment, Let n be the predicted risk of adjacent devices, and n be the number of adjacent devices included in the calculation. In the gradient descent algorithm, the independent variable is the distance threshold (denoted as D, unit: m) between the faulty insulator and its adjacent devices, with the initial value set to the radius of the influence range of 50m; the learning rate η = 0.01 (to control the adjustment range of each iteration and avoid oscillation); the iteration termination condition is that the change in the loss function L < 0.001, and the iteration is repeated until L converges, and the finally converged D is the optimized distance threshold.

[0070] For example, D 初始=50m, including all 5 adjacent devices, the loss function L=6.96 is calculated. After 300 iterations, the loss function L stabilizes at 1.2 (change < 0.001), at which point D=42.0m, that is, the distance threshold between the faulty insulator device and its adjacent devices is 42.0m. Then the spatial distance from each adjacent device to INS-027 is recalculated: QS-012 (10.0m)≤42.0m→retain; T-003 (15.0m)≤42.0m→retain; CT-008 (20.0m)≤42.0m→retain; BUS-002 (45.0m)>42.0m→remove; PT-004 (48.0m)>42.0m→remove. Therefore, the adjusted impact range of insulator fault equipment on adjacent equipment is {INS-027 (insulator fault equipment), QS-012 (disconnecting switch), T-003 (transformer), CT-008 (current transformer)}, which focuses more on high-risk equipment compared to the original range (including BUS-002 and PT-004).

[0071] In the above method, a predictive risk model is constructed based on failure mode and spatial correlation data to achieve preliminary risk quantification. Then, the actual risk is calculated using historical failure data of adjacent devices, providing an objective calibration basis for the predicted risk. Subsequently, the gradient descent algorithm is used to minimize the error between the predicted and actual risks, optimizing the distance threshold to fit the real-world scenario and avoiding bias from subjectively set thresholds. Finally, the impact range is adjusted based on the optimized threshold, eliminating low-risk, long-distance devices and retaining high-risk core devices. This transforms the failure impact range from static and fixed to dynamic and precise, significantly improving the accuracy of failure risk assessment.

[0072] This disclosure also provides an Internet of Things (IoT) based substation fault detection system. The system may include an acquisition module, a first determination module, a first judgment module, a second determination module, a matching module, a second judgment module, a third determination module, a construction module, a fourth determination module, and a fifth determination module. The acquisition module is used to acquire operating status parameters of insulators and narrow passage environmental parameters in the substation. The operating status parameters include the surface temperature of the insulators, the surface humidity of the insulators, and the amplitude of the insulators. The narrow passage environmental parameters include the humidity and wind speed within the narrow passage. The first determination module is used to determine the insulator aging index based on the operating status parameters and the narrow passage environmental parameters. The first judgment module is used to determine whether the insulator aging index exceeds a preset aging index threshold. If it exceeds the preset aging index threshold, it is marked as a faulty insulator, and the GPS positioning signal and UWB positioning signal of the faulty insulator are acquired based on wireless positioning technology. The second determination module is used to acquire the weights corresponding to the GPS positioning signal and the UWB positioning signal, and based on the weights corresponding to the GPS positioning signal and the UWB positioning signal, determine the aging index of the insulator. The system comprises five modules: a signal module to determine the initial coordinates of the faulty insulator; a matching module to establish a 3D layout model of the substation and match the model point coordinates of the corresponding faulty insulator within the model; a judgment module to determine whether the Euclidean distance between the model point coordinates and the initial coordinates exceeds a preset Euclidean distance threshold; if it does, the weights corresponding to the GPS positioning signal and the UWB positioning signal are adjusted accordingly; a determination module to determine the precise coordinates of the faulty insulator based on the adjusted weights of the GPS and UWB positioning signals; a construction module to construct a spatial correlation diagram structure based on the precise coordinates of the faulty insulator; a construction module to determine the impact range of the faulty insulator on adjacent equipment based on the spatial correlation diagram structure; and a final determination module to determine the maintenance navigation path from the narrow passage entrance to the faulty insulator based on the impact range.

Claims

1. A substation fault detection method based on Internet of Things, characterized in that, The method comprises the following steps: acquiring the operation state parameters of the insulator device in the transformer substation and the narrow channel environment parameters, the operation state parameters including the temperature of the surface of the insulator device, the humidity of the surface of the insulator device and the amplitude of the insulator device, and the narrow channel environment parameters including the humidity and the wind speed in the narrow channel; determining the insulator aging index according to the operation state parameters of the insulator device and the narrow channel environment parameters; judging whether the insulator aging index exceeds the preset aging index threshold value, and if it is judged that the preset aging index threshold value is exceeded, marking the insulator fault device, and acquiring the GPS positioning signal and the UWB positioning signal of the insulator fault device respectively based on the wireless positioning technology; acquiring the weight corresponding to the GPS positioning signal and the weight corresponding to the UWB positioning signal respectively, and determining the initial coordinates of the insulator fault device according to the GPS positioning signal and the UWB positioning signal based on the weight corresponding to the GPS positioning signal and the weight corresponding to the UWB positioning signal; acquiring the three-dimensional layout model of the transformer substation, and matching the model point coordinates corresponding to the insulator fault device in the three-dimensional layout model of the transformer substation; judging whether the Euclidean distance between the model point coordinates and the initial coordinates exceeds the preset Euclidean distance threshold value, and if it is judged that the preset Euclidean distance threshold value is exceeded, adjusting the weight corresponding to the GPS positioning signal and the weight corresponding to the UWB positioning signal respectively; determining the accurate coordinates of the insulator fault device according to the adjusted weight corresponding to the GPS positioning signal and the adjusted weight corresponding to the UWB positioning signal; constructing the spatial correlation graph structure according to the accurate coordinates of the insulator fault device; determining the influence range of the insulator fault device on the adjacent devices according to the spatial correlation graph structure; determining the maintenance navigation path from the narrow channel entrance to the insulator fault device according to the influence range. 2.The IoT-based substation fault detection method of claim 1, wherein, The method comprises the following steps: calculating the deviation between the operation state parameters and the preset threshold value corresponding to the operation state parameters respectively, and taking the calculation result as the first deviation value; calculating the deviation between the narrow channel environment parameters and the preset threshold value corresponding to the preset narrow channel environment parameters respectively, and taking the calculation result as the second deviation value; acquiring the weight corresponding to the operation state parameters and the weight corresponding to the preset narrow channel environment parameters respectively; performing weighted operation on the first deviation value and the second deviation value based on the weight corresponding to the operation state parameters and the weight corresponding to the preset narrow channel environment parameters, and taking the weighted operation result as the insulator aging index. 3.The IoT-based substation fault detection method of claim 1, wherein, The method comprises the following steps: determining a first coordinate of the insulator fault device according to the GPS positioning signal; determining a second coordinate of the insulator fault device according to the UWB positioning signal; calculating a coordinate offset of the second coordinate relative to the first coordinate based on a TDOA algorithm; correcting the first coordinate according to the coordinate offset; obtaining a weight corresponding to the GPS positioning signal and a weight corresponding to the UWB positioning signal respectively; determining an initial coordinate of the insulator fault device based on the corrected first coordinate, the second coordinate, the weight corresponding to the GPS positioning signal and the weight corresponding to the UWB positioning signal according to an extended Kalman filtering algorithm. 4.The IoT-based substation fault detection method of claim 1, wherein, The constructing a space correlation graph structure according to the accurate coordinate of the insulator fault device comprises: obtaining an associated device identifier and an associated device operating state parameter of an associated device related to the insulator fault device based on the accurate coordinate of the insulator fault device; constructing a space correlation graph structure based on a graph theory algorithm, taking the associated device identifier as a node, taking an electrical connection relationship between the associated devices as an edge, and taking the associated device operating state parameter as a weight. 5.The IoT-based substation fault detection method of claim 1, wherein, The determining an influence range of the insulator fault device on adjacent devices according to the space correlation graph structure comprises: traversing the space correlation graph structure based on a breadth-first search algorithm, and respectively calculating a path length and a weight accumulation between the insulator fault device and each associated device; determining a fault correlation coefficient of each associated device according to the path length and the weight accumulation; judging whether the fault correlation coefficient exceeds a preset correlation coefficient threshold; if it is judged that the preset correlation coefficient threshold is exceeded, marking the current associated device as a high-risk device, and determining a first influence range according to the insulator fault device and the high-risk device; traversing the space correlation graph structure from the high-risk device as a starting point until a fault device associated with the high-risk device is traced back based on a depth-first search algorithm; calculating a fault propagation probability of the fault device associated with the high-risk device; judging whether the fault propagation probability is greater than a preset fault propagation probability threshold; if it is judged that the preset fault propagation probability threshold is exceeded, starting an alarm; if it is judged that the preset fault propagation probability threshold is not exceeded, determining an influence range of the insulator fault device on adjacent devices according to the insulator fault device, the high-risk device and the fault device associated with the high-risk device. 6.The IoT-based substation fault detection method of claim 1, wherein, The determining a maintenance navigation path from a narrow channel entrance to the insulator fault device according to the influence range comprises: obtaining a coordinate of the narrow channel entrance and a coordinate of the insulator fault device within the influence range; determining a shortest path between the narrow channel entrance and the insulator fault device according to the coordinate of the narrow channel entrance and the coordinate of the insulator fault device based on an A-star algorithm; obtaining an object height of an intermediate node passed through from the narrow channel entrance to the insulator fault device; judging whether the object height exceeds a preset height threshold; If it is judged that the preset height threshold is exceeded, the current intermediate node is replaced to determine a maintenance navigation path from the narrow channel entrance to the insulator fault device. 7.The IoT-based substation fault detection method of claim 6, wherein, After the maintenance navigation path from the narrow channel entrance to the insulator fault device is determined according to the influence range, the method further includes: judging whether the maintenance navigation path has an obstacle; if it is judged that the obstacle exists, identifying the type of the obstacle based on an image recognition algorithm and outputting position information of the obstacle; redetermining a shortest path between the narrow channel entrance and the insulator fault device according to the position information of the obstacle based on an A-star algorithm. 8.The IoT-based substation fault detection method of claim 1, wherein, After the maintenance navigation path from the narrow channel entrance to the insulator fault device is determined according to the influence range, the method further includes: obtaining historical fault data of the insulator fault device and determining a first fault feature vector according to the historical fault data of the insulator fault device; obtaining an operating state parameter of the insulator fault device and determining a second fault feature vector according to the operating state parameter of the insulator fault device; calculating a cosine similarity between the first fault feature vector and the second fault feature vector; judging whether the cosine similarity exceeds a preset cosine similarity threshold; if it is judged that the preset cosine similarity threshold is exceeded, clustering the historical fault data based on a K-means clustering algorithm to determine a fault mode corresponding to the current insulator fault device, the fault mode including flashover fault, aging fault, pollution fault, mechanical fault and lightning strike fault. 9.The IoT-based substation fault detection method of claim 8, wherein, After the historical fault data is clustered based on the K-means clustering algorithm to determine the fault mode corresponding to the current insulator fault device if it is judged that the preset cosine similarity threshold is exceeded, the method further includes: determining a predicted risk of each of the adjacent devices within the influence range of the insulator fault device on the adjacent devices; obtaining historical fault data of each of the adjacent devices and determining an actual risk of the adjacent devices according to the historical fault data of the adjacent devices; determining a distance threshold between the insulator fault device and the adjacent devices based on a gradient descent algorithm according to the predicted risk and the actual risk; adjusting the influence range of the insulator fault device on the adjacent devices according to the distance threshold between the insulator fault device and the adjacent devices.

10. A substation fault detection system based on Internet of Things, characterized by, The method includes: an obtaining module, configured to obtain an operating state parameter of an insulator device in a substation and a narrow channel environment parameter, the operating state parameter including a temperature of a surface of the insulator device, a humidity of the surface of the insulator device and an amplitude of the insulator device, and the narrow channel environment parameter including a humidity and a wind speed in the narrow channel; a first determining module, configured to determine an insulator aging index according to the operating state parameter of the insulator device and the narrow channel environment parameter; The first judging module is configured to judge whether the insulator aging index exceeds a preset aging index threshold, and if it is judged that the insulator aging index exceeds the preset aging index threshold, mark the insulator as a fault device, and respectively acquire a GPS positioning signal and a UWB positioning signal of the insulator fault device based on a wireless positioning technology; The second determining module is configured to respectively acquire a weight corresponding to the GPS positioning signal and a weight corresponding to the UWB positioning signal, and determine an initial coordinate of the insulator fault device according to the GPS positioning signal and the UWB positioning signal based on the weight corresponding to the GPS positioning signal and the weight corresponding to the UWB positioning signal; The matching module is configured to establish a three-dimensional layout model of a substation, and match a model point coordinate corresponding to the insulator fault device in the three-dimensional layout model of the substation; The second judging module is configured to judge whether a Euclidean distance between the model point coordinate and the initial coordinate exceeds a preset Euclidean distance threshold, and if it is judged that the Euclidean distance exceeds the preset Euclidean distance threshold, respectively adjust the weight corresponding to the GPS positioning signal and the weight corresponding to the UWB positioning signal; The third determining module is configured to determine an accurate coordinate of the insulator fault device according to the adjusted weight corresponding to the GPS positioning signal and the adjusted weight corresponding to the UWB positioning signal; The constructing module is configured to construct a spatial correlation graph structure according to the accurate coordinate of the insulator fault device; The fourth determining module is configured to determine an influence range of the insulator fault device on a neighboring device according to the spatial correlation graph structure; The fifth determining module is configured to determine a maintenance navigation path from a narrow channel entrance to the insulator fault device according to the influence range.

Images