Electric power iron tower remote detection fastener and detection method thereof

Abstract

The invention provides an electric power iron tower remote detection fastener and a detection method thereof, and relates to the technical field of electric power facility steel constructions.The electric power iron tower remote detection fastener comprises a bolt, a nut is in threaded connection with the bolt, and an electric power iron tower steel structure is arranged on a bolt rod between the nut and a bolt cap of the bolt; grooved gaskets are arranged on the portions, between the bolt and the electric iron tower steel structure and between the nut and the electric iron tower steel structure, of the tightening latch, and label gaskets are arranged on the portions, between the bolt and the grooved gaskets and between the nut and the grooved gaskets, of the tightening latch. According to the invention, the fastening force change of the bolt and the nut is converted into a measurable signal of the electronic tag through the cooperation of a mechanical structure in the tag gasket and a circuit, and lossless and remote detection of the fastening state of the electric iron tower is realized.

Description

Technical Field

[0001] This invention relates to the field of steel structure technology for power facilities, and in particular to a remote inspection fastener for power transmission towers and its inspection method. Background Technology

[0002] Fasteners for power transmission towers mainly refer to components such as bolts and nuts used to connect and fix the steel structure of the tower. Their core function is to maintain the mechanical integrity and stability of the tower structure. The inspection of fasteners aims to assess whether their preload is maintained within a safe range in order to prevent structural failure due to loosening.

[0003] In existing technologies, traditional detection methods mainly rely on manual inspection, such as tapping, wrench inspection, or using torque wrenches. These methods are not only inefficient and dependent on personnel experience, but also difficult to detect and quantify early loosening risks in a timely manner. While some technologies based on resistance strain gauges or fiber optic sensors can improve accuracy, they often suffer from system complexity, high cost, or insufficient stability in complex electromagnetic environments, and cannot effectively meet the operation and maintenance needs of efficient, non-destructive, and real-time remote detection of large power tower groups. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology. This invention converts the changes in the tightening force of bolts and nuts into measurable signals of electronic tags through the synergy of the mechanical structure and circuit inside the tag washer, thereby realizing non-destructive and remote detection of the tightening status of power towers.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a remote detection fastener for power transmission towers, comprising a bolt, a nut threaded onto the bolt, a power transmission tower steel structure mounted on the bolt shank between the nut and the bolt head, grooved washers on the bolt shank between the bolt and the power transmission tower steel structure and between the nut and the power transmission tower steel structure, and label washers on the bolt shank between the bolt and the grooved washers and between the nut and the grooved washers; The label gasket includes an annular electronic tag. Side telescopic strips are fixedly adhered to both sides of the electronic tag. The two side telescopic strips are connected by a rigid connecting rod. The side of the side telescopic strip away from the electronic tag is fixedly adhered to the outer edge of the side gasket ring. The two side gasket rings are connected by an arc-shaped spring. A sleeve is fixedly welded to the side of one of the sleeves. An insert is slidably inserted into the sleeve. The ends of the sleeve and the insert are connected by an elastic support rod. A force-sensitive resistor is fixedly installed on the end of the insert away from the sleeve. The force-sensitive resistor is movably disposed in a movable groove. The movable groove extends through the other sleeve.

[0006] In a preferred embodiment, the number of force-sensitive resistors is four, and the four force-sensitive resistors are connected in series in the circuit of the electronic tag.

[0007] As a preferred embodiment, the distance between the two ends of the arc-shaped spring sheet when it is naturally unfolded is greater than the distance between the two ends of the elastic strut when it is naturally unfolded.

[0008] A method for remotely inspecting fasteners on power transmission towers includes a testing process using a label washer as the inspected component, specifically comprising the following steps: S1. Inspection Preparation: Construct a drone inspection route planning system, number each electronic tag, formulate a tiered alarm strategy, and before installing the tag gasket, detect the correlation between the pressure on the tag gasket and signal changes. Simultaneously, while detecting the correlation between the pressure on the tag gasket and signal changes, build a pressure-signal strength response model. During the construction of the pressure-signal strength response model: A gradient preload is applied to the label gasket using a pressure testing bench, and the reading signal strength of the electronic tag is measured simultaneously. The critical switching threshold between signal presence and absence when the label gasket is under pressure and without pressure is recorded. The direction of the preload applied to the label gasket by the pressure testing bench is the same as the radial direction of the label gasket. S2. Inspection Planning: The drone inspection route planning system calculates the risk coefficient of the drone inspection task based on historical inspection data, and generates a dynamic drone inspection route based on the RFID signal strength distribution and risk coefficient in the historical inspection data. S4. Inspection execution: Install RFID readers on the drone and collect data from 5 during the drone inspection process; S5. Inspection Data Analysis: Process the data read by RFID during the drone inspection process and perform predictive analysis through AI prediction models; S6. Inspection Result Analysis: Execute the corresponding hierarchical alarm strategy based on the results of the predictive analysis; S7, Inspection Optimization: Optimize the drone inspection route planning system based on inspection data.

[0009] In a preferred embodiment, the risk coefficient in step S2 is calculated as follows: Risk coefficient = weight value 1 * historical anomaly frequency + weight value 2 * wind data + weight value 3 * the floor number of the inspected location; In this calculation, only the values ​​of historical abnormal frequency, wind force data, and the layer number of the inspected location are substituted; the units are not substituted into the calculation.

[0010] In a preferred implementation, in step S2, when the UAV inspection route planning system generates a dynamic flight path, it first calculates the reading failure rate of all electronic tags in the last thirty inspections as historical anomaly frequency. Simultaneously, it accesses real-time meteorological data streams to obtain current wind data and calculates multi-gradient risk coefficients based on the different altitudes of the electronic tags. Then, it combines this with the spatial distribution heatmap of RFID signal strength recorded in historical inspections to divide risk areas into high-risk areas (risk coefficient greater than 0.7), medium-risk areas (risk coefficient greater than or equal to 0.3 and less than or equal to 0.7), and low-risk areas (risk coefficient less than 0.3). Different cruise strategies are formulated for different risk areas. Furthermore, the UAV inspection route planning system dynamically adjusts the flight path based on real-time wind data, optimizes the UAV antenna orientation using historical RFID signal strength distribution data, and employs an S-shaped enhanced scanning trajectory in weak signal areas. Finally, it solves a multi-objective optimization problem using a genetic algorithm, taking into account minimizing total inspection time, maximizing high-risk area coverage, and battery life constraints, outputting a real-time dynamic flight path and injecting it into the UAV flight control system.

[0011] In a preferred implementation, step S5 first employs a sliding window Kalman filter to eliminate multipath effects and UAV attitude fluctuation noise in the received signal strength indication sequence of each electronic tag. Invalid data caused by antenna polarization mismatch is then eliminated based on phase consistency detection. Next, the mean, variance, and attenuation slope of the received signal strength are extracted from the electronic tag signal in the time domain. The dominant frequency characteristics of the electronic tag signal fluctuation are analyzed using a fast Fourier transform. Simultaneously, an electronic tag signal correlation matrix is ​​constructed for the same layer of tags. Historical data from the UAV inspection route system are combined and input into an AI prediction model. This AI prediction model uses a hybrid architecture of deep temporal convolutional networks and long short-term memory networks. The temporal convolutional network extracts local mutation patterns, while the long short-term memory network captures long-term evolution patterns. Based on the nonlinear mapping between signal attenuation slope and historical abnormal frequencies, the AI ​​prediction model predicts the failure probability of a specific electronic tag within a short period. Furthermore, the AI ​​prediction model utilizes a graph neural network to analyze the spatial correlation between the mechanical transmission path and signal attenuation in the tower structure, predicting spatially correlated fault risks.

[0012] In a preferred implementation, in step S7, during the optimization process, the UAV inspection route planning system first inputs the data streams collected by the UAV, including the success rate of electronic tag reading, the spatiotemporal distribution of signal strength, and the actual route execution efficiency, into the data warehouse of the UAV inspection route planning system. It then uses an incremental learning algorithm to update the risk coefficient algorithm and optimizes the route generation strategy based on the signal strength attenuation mode. With battery life and weather disturbances as constraints, the reward function comprehensively considers the detection rate of high-risk electronic tags and the total inspection time to iteratively generate the Pareto optimal route. The generated Pareto optimal route is then input into the UAV inspection route planning system for training and optimization.

[0013] In a preferred embodiment, the tiered alarm policy in step S is formulated by dividing the response into three levels based on the number of electronic tags that have lost connection: 1) Level 1 alarm: When a single electronic tag cannot be read, a re-check mark is triggered; 2) Level 2 Alarm: When multiple electronic tags at the same node cannot be read, manual verification is initiated; 3) Level 3 alarm: If all electronic tags at the same node cannot be read, the emergency maintenance process will be initiated immediately.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: This invention converts the changes in the tightening force of bolts and nuts into measurable signals of electronic tags through the synergy of the mechanical structure and circuitry inside the tag gasket, thereby achieving non-destructive, remote detection of the tightening status of power transmission towers.

[0015] This invention utilizes a precise operation and maintenance mode that is automatically executed by drones and intelligently analyzed by data models. Through a pre-established pressure-signal model and unique electronic tag identification, the physical traceability and spatial accuracy of the detection data are ensured. Combined with dynamic flight path planning based on risk coefficients, advanced signal processing, and AI prediction models, the system can provide early warnings of the probability of failure of individual fasteners and identify spatially related risks, thereby triggering automated hierarchical responses. Finally, through a continuous learning and optimization mechanism, the entire system significantly reduces operation and maintenance costs and safety risks while improving detection efficiency and reliability, ensuring the long-term safe and stable operation of power transmission tower structures. Attached Figure Description

[0016] Figure 1 The flowchart of a method for remotely inspecting fasteners on power transmission towers is provided in this invention. Figure 2 This invention provides a structural schematic diagram of a fastener for remote detection of power transmission towers; Figure 3 This invention provides a schematic diagram of the label washer position for remotely inspecting fasteners on power transmission towers; Figure 4 A perspective view of a label washer for remote detection of fasteners on power transmission towers is provided for this invention. Figure 5 This invention provides a half-sectional schematic diagram of a label washer for remotely detecting fasteners on power transmission towers. Figure 6 This invention provides an installation diagram of a force-sensitive resistor for remotely detecting fasteners on power transmission towers.

[0017] Legend: 1. Bolts; 2. Nuts; 3. Steel structure of power transmission towers; 4. Grooved washers; 5. Label washers; 51. Electronic tag; 52. Side telescopic belt; 53. Rigid connecting rod; 54. Side washer ring; 55. Arc-shaped spring; 56. Sleeve; 57. Insert tube; 58. Elastic support rod; 59. Force-sensitive resistor; 510. Movable groove. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1 like Figure 2-6 As shown, the present invention provides a technical solution: a remote detection fastener for power transmission towers, including a bolt 1, a nut 2 threadedly connected to the bolt 1, and a power transmission tower steel structure 3 provided on the bolt rod between the nut 2 and the bolt head of the bolt 1, characterized in that: a grooved washer 4 is provided on the bolt rod between the bolt 1 and the power transmission tower steel structure 3 and between the nut 2 and the power transmission tower steel structure 3, and a label washer 5 is provided on the bolt rod between the bolt 1 and the grooved washer 4 and between the nut 2 and the grooved washer 4; In the above content, by setting the label washer 5, the label washer 5 can be subjected to pressure when the bolt 1 and nut 2 are tightened, thereby utilizing the built-in sensing mechanism of the label washer 5 to respond to the change of preload force and affect the signal characteristics of the electronic tag 51, realizing indirect detection of external pressure. By remotely reading the signal change of the label washer 5, the loosening or abnormal state of the fastener can be detected in real time, which can improve the safety detection efficiency and maintenance efficiency of the power tower structure. Specifically, the label gasket 5 includes an annular electronic tag 51. Side telescopic strips 52 are fixedly adhered to both sides of the electronic tag 51. The two side telescopic strips 52 are connected by a rigid connecting rod 53. The side of the side telescopic strip 52 away from the electronic tag 51 is fixedly adhered to the outer edge of a side gasket ring 54. The two side gasket rings 54 are connected by an arc-shaped spring piece 55. The telescopic creases on the side telescopic strips 52 maintain the connection between the electronic tag 51 and the side gasket rings 54 while providing sufficient clearance for spacing changes. A sleeve 56 is fixedly welded to the side of one of the sleeves 56. An insert 57 is slidably inserted into the sleeve 56. The ends of the tube 57 and the insert tube 58 are connected by an elastic support rod 58. The distance between the two ends of the elastic support rod 58 when it is naturally extended is less than the distance between the two ends of the arc-shaped spring piece 55 when it is naturally extended. A force-sensitive resistor 59 is fixedly installed on the end of the insert tube 57 away from the sleeve 56. A through groove is opened on the side of the insert tube 57. The wires of the force-sensitive resistor 59 can extend to the outside of the insert tube 57 through the internal space and the through groove to connect with the electronic tag 51. There are four force-sensitive resistors 59, and the four force-sensitive resistors 59 are connected in series in the circuit of the electronic tag 51. The force-sensitive resistors 59 are movably set in the movable cavity 510. The movable cavity 510 is opened through another sleeve 56.

[0020] In this embodiment, the label gasket 5, through the synergistic effect of its internal mechanical structure and circuitry, converts changes in physical pressure into detectable electrical signals. Specifically: First, during installation, the side washer rings 54 bear axial pressure from the bolt 1 and nut 2, forcing them to move closer together. This compresses the arc-shaped spring piece 55 between the two side washer rings 54. Since the natural length of the elastic support rod 58 is less than the natural length of the arc-shaped spring piece 55, the elastic support rod 58 will be under force after the arc-shaped spring piece 55. When the distance between the two side washer rings 54 is less than the natural length of the elastic support rod 58, the force-sensitive resistor 59 will be pushed out of the movable groove 510 by the elastic support rod 58 and press against the bolt cap 1 and nut 2. Under the continuous increase of axial pressure, the thrust exerted by the elastic support rod 58 on the force-sensitive resistor 59 will increase accordingly, thereby continuously increasing the pressure on the force-sensitive resistor 59 and continuously decreasing its resistance, thus directly modulating the circuit characteristics of the electronic tag 51. The side telescopic strip 52 and its telescopic creases provide the necessary deformation allowance during the overall deformation of the label gasket 5, ensuring that the electronic label 51 itself is not directly damaged and can maintain its connection with the side gasket 54. This design enables indirect, non-destructive testing of the preload of fasteners. When the bolt 1 connection becomes loose, the pressure acting on the tag washer 5 decreases, and the resistance of the force-sensitive resistor 59 changes, which in turn causes a corresponding change in the strength of the RFID signal returned by the electronic tag 51. This allows the overall fastening status of the fasteners to be detected and identified remotely.

[0021] Example 2 like Figure 1-6 As shown, based on the remote inspection fastener for power transmission towers provided in Embodiment 1, a method for inspecting the remote inspection fastener for power transmission towers is proposed, including an inspection process using a label washer 5 as the inspected component, specifically including the following steps: S1. Inspection Preparation: First, before installing the label washer 5, the pressure on the label washer 5 and the signal change are correlated. The quantitative correspondence between the physical pressure on the label washer 5 and the signal characteristics emitted by the electronic tag 51 is established in advance through experiments. This is achieved by applying a series of radial preloads of known size and direction consistent with the installation on the label washer 5 on a pressure test bench, and simultaneously measuring the reading signal strength of the electronic tag 51. In this way, the tightening force of the bolt 1 is converted into specific signal parameters that can be obtained remotely and non-contactly by the drone RFID reader. This allows users to infer whether the tightening force between the bolt 1 and the nut 2 is maintained within a safe range by detecting the change in the signal strength of the electronic tag 51. Simultaneously, a pressure-signal strength response model is constructed. During the construction process, a gradient preload is applied to the tag gasket 5 using a pressure test bench, and the reading signal strength of the electronic tag 51 is measured simultaneously. The critical transition threshold between the presence and absence of the signal when the tag gasket 5 is under pressure and without pressure is recorded. The direction of the preload applied to the tag gasket 5 by the pressure test bench is the same as the radial direction of the tag gasket 5. By creating a pressure-signal strength response model as a quantitative and repeatable reference standard, the abstract preload of the fastener is transformed into an electrical signal that the UAV inspection system can directly identify and measure. By establishing the critical transition threshold, a system for formulating UAV inspection routes is developed, risk coefficients are calculated, AI prediction models are used for failure probability analysis, and graded alarm strategies are triggered. This provides a physical basis and decision-making benchmark, ensuring the scientificity and reliability of the entire remote detection system from physical phenomena to data to maintenance. Furthermore, each electronic tag 51 is numbered, and the coding structure can include the power tower number, structural layer, node location, and installation sequence code. By using structured coding rules, a unique digital identity is created for each electronic tag 51, which can systematically integrate and digitally map the spatial location and time sequence information of the fasteners to which the electronic tag 51 is attached in the vast power grid assets. Then, a drone inspection route planning system is constructed, which binds the unique code of each electronic tag 51 with the precise geographic coordinates and enters them into the database, so that the data of the detection point can be accurately located in the digital space. At the same time, the three-dimensional geographic information module integrated in the geographic coordinates can process the spatial structure and relationship, and provide terrain and obstacle information for route planning. Then, based on electromagnetic wave theory, a radio frequency propagation model module is constructed to calculate the probability of the drone successfully reading the electronic tag 51 at a specific location under different flight altitudes, antenna orientations and other poses, so as to predict the quality of the communication link in the planning stage. At the same time, a tiered alarm strategy was formulated, dividing the response into three levels based on the number of lost connections of electronic tag 51: Level 1 alarm: If a single electronic tag 51 cannot be read, a re-check mark is triggered; Level 2 Alarm: When multiple electronic tags 51 cannot be read at the same node, manual verification is initiated; 3. Level 3 alarm: If all electronic tags 51 at the same node cannot be read, the emergency maintenance procedure will be initiated immediately; S2, Inspection Planning: First, the drone inspection route planning system calculates the risk coefficient of the drone inspection task based on historical detection data. The risk coefficient is calculated as follows: Risk coefficient = weight value 1 * historical anomaly frequency + weight value 2 * wind data + weight value 3 * layer number of the inspected location. In this calculation, only the values ​​of historical anomaly frequency, wind data, and layer number of the inspected location are substituted for calculation, and the units are not substituted for calculation. This calculation enables the drone inspection route planning system to divide the detection area into high-risk, medium-risk, and low-risk areas according to the risk coefficient, and dynamically adjust the drone's patrol strategy accordingly. Then, based on the RFID signal strength distribution and risk coefficient in historical inspection data, the reading failure rate of all electronic tags 51 in the last thirty inspections was calculated as the historical anomaly frequency. At the same time, the current wind force data was obtained by accessing the real-time meteorological data stream, and multi-gradient risk coefficients were calculated in combination with the different heights of the electronic tags 51. Combined with the RFID signal strength spatial distribution heat map recorded in historical inspections, the risk areas were divided into high-risk areas with a risk coefficient greater than 0.7, medium-risk areas with a risk coefficient greater than or equal to 0.3 and less than or equal to 0.7, and low-risk areas with a risk coefficient less than 0.3. Different patrol strategies were formulated for different risk areas. For high-risk areas, dense waypoint planning was adopted, with the waypoint spacing reduced to 5 meters and the hovering time extended to 3 seconds. For medium-risk areas, waypoints were deployed at an 8-meter spacing. For low-risk areas, adaptive sampling detection was adopted, with the waypoint spacing expanded to 15 meters. By implementing differentiated and refined patrol strategies for areas with different risk levels, the precise focus and rapid response to potential fault points were achieved with limited resources, improving the intelligent level of power tower fastener condition detection and the efficiency of operation and maintenance resource utilization. Meanwhile, the drone inspection route planning system dynamically adjusts the flight path based on real-time wind data. When the wind force is greater than level 5, it automatically reduces the flight altitude to 3 meters and avoids the windward path. It optimizes the drone antenna orientation using historical RFID signal strength distribution data and adopts an S-shaped enhanced scanning trajectory in areas with weak signals. By adaptively avoiding wind interference and optimizing the radio frequency communication link, it ensures the safety of the inspection mission and the integrity of the detection data in strong wind environments, thereby ensuring the robustness and continuous effectiveness of the entire power tower fastener remote detection system in complex natural environments. Finally, a multi-objective optimization problem is solved using a genetic algorithm. The genetic algorithm uses the fitness function as the evaluation criterion and takes minimizing the total inspection time, maximizing the coverage of high-risk areas, and battery life constraints as conditions. By simulating the selection, crossover, and mutation operations in natural selection, it iteratively searches in a huge space of possible route solutions and finally converges to an optimal solution that achieves a balance among multiple competing objectives. The real-time dynamic route is output and injected into the UAV flight control system to generate a dynamic UAV inspection route. S3, Inspection Execution: Install the RFID reader on the drone and collect data on 5 during the drone inspection process. By mounting the RFID reader on the drone platform, the label washer 5 installed between the bolt 1 and the nut 2 is collected in a non-contact and automated manner while the drone is flying along the preset dynamic route. S4. Inspection Data Analysis: First, the RFID data read during the UAV inspection process is processed using a sliding window Kalman filter. The state of the received signal strength indication sequence of electronic tag 51 is recursively estimated, and the prediction and update steps are dynamically adjusted to eliminate multipath effects and UAV attitude fluctuation noise in the received signal strength indication sequence of each electronic tag 51, thereby improving the stability and accuracy of the signal sequence. Based on phase consistency, the alignment of the signal phase is analyzed, and invalid data caused by antenna polarization mismatch is detected and eliminated. Then, the mean value reflecting the average signal level, the variance representing the signal fluctuation range, and the lower limit of the indicated signal strength are extracted from the time domain of the electronic tag 51 signal. The data characteristics of the decay slope of the decay rate are used to capture the basic statistical characteristics of the fastener preload change. The main frequency characteristics of the electronic tag 51 signal fluctuation are analyzed by fast Fourier transform to identify potential periodic anomalies or vibration modes. At the same time, the electronic tag 51 signal correlation matrix of the same layer tag group is constructed. By analyzing the spatial correlation between multiple electronic tag 51 signals, combined with historical data in the UAV inspection route formulation system, and jointly input into the AI ​​prediction model, the deep temporal convolutional network and long short-term memory network architecture are used to learn the local mutations and long-term laws of the electronic tag 51 signal evolution, so as to accurately predict the failure probability and spatial correlation risk of fasteners. Then, predictive analysis is performed using an AI prediction model. The AI ​​prediction model adopts a hybrid architecture of deep temporal convolutional network and long short-term memory network. The temporal convolutional network extracts local mutation patterns, while the long short-term memory network captures long-term evolution patterns. Through the design of the hybrid architecture, the AI ​​prediction model can comprehensively handle short-term fluctuations and long-term dependencies. Based on the nonlinear mapping between signal attenuation slope and historical abnormal frequency, the AI ​​prediction model predicts the failure probability of a specific electronic tag 51 within a certain period of time, thereby achieving early warning of fastener loosening risks. Furthermore, the AI ​​prediction model uses graph neural networks to analyze the spatial correlation between mechanical transmission paths and signal attenuation in the tower structure, modeling the tower nodes as graph structures. By capturing the propagation of mechanical influence through the connection relationships between nodes, it predicts spatial correlation failure risks, ultimately improving the accuracy and reliability of detection, reducing maintenance blind spots, and optimizing resource allocation. S5. Inspection Result Analysis: Based on the results of the predictive analysis, corresponding hierarchical alarm strategies are executed. Specifically: When the AI ​​prediction model predicts that a single electronic tag 51 has failed, the server automatically generates a review task work order containing its unique code and geographic coordinates, triggering a level 1 alarm and marking the fastener as pending review; when the model predicts that multiple electronic tags 51 on the same tower structure node will fail simultaneously, it will automatically escalate to a level 2 alarm, initiate a manual on-site verification process, and highlight the risk node in the drone inspection route planning system; when the model predicts that all electronic tags 51 on the same node will fail, a level 3 alarm will be triggered immediately, automatically initiating an emergency maintenance process, and simultaneously pushing detailed early warning information, risk location data, and suggested measures to the maintenance team, thereby achieving closed-loop automated management from data prediction to maintenance action. S6, Inspection Optimization: First, the data streams collected by the UAV, including the success rate of reading the electronic tag 51, the spatiotemporal distribution of signal strength, and the actual flight path execution efficiency, are input into the data warehouse of the UAV inspection route planning system. The risk coefficient algorithm is updated using an incremental learning algorithm. The statistical window is dynamically slid based on the historical anomaly frequency, and the most recent 50 inspection records are retained to ensure that the model training data always represents the latest operating status. Then, the objective function is optimized iteratively using the gradient descent method, and the distribution ratio between the weight values ​​is continuously adjusted. By calculating the error gradient between the predicted risk value and the actual fault state and updating the weight parameters along the negative gradient direction, the prediction error is minimized, so that the risk coefficient algorithm can adaptively approximate the real risk situation. Finally, the UAV inspection route planning system achieves self-evolution and accuracy improvement, ensuring that its ability to identify and warn of risk areas of power tower fasteners is continuously optimized over time. Meanwhile, the route generation strategy is optimized based on the signal strength attenuation pattern. By analyzing the spatiotemporal characteristics of signal attenuation, areas of sustained high stability are identified. In these areas, the waypoint spacing is automatically increased to 20 meters and the detection frequency is reduced to decrease resource consumption. For newly identified risk clusters, waypoints are narrowed to within 5 meters. With battery life and weather disturbances as constraints, the reward function comprehensively considers the detection rate of high-risk electronic tags 51 and the total inspection time to iteratively generate Pareto optimal routes. The generated Pareto optimal routes are then input into the UAV inspection route planning system for training and optimization. This allows the system to adaptively adjust the inspection strategy, improve the efficiency and reliability of electronic tag 51 detection, extend UAV endurance, and reduce operating costs.

[0022] In this embodiment, during the inspection preparation stage, a gradient preload is applied to the tag gasket 5 using a pressure test bench, and the reading signal strength of the electronic tag 51 is measured simultaneously. This establishes a quantitative correspondence between physical pressure and the signal characteristics of the electronic tag 51, as well as a critical conversion threshold. This transforms the difficult-to-detect tightening force between the bolt 1 and the nut 2 into remotely detectable electrical signal parameters, providing a physical basis for subsequent analysis. Simultaneously, by structurally encoding each electronic tag 51 and binding the unique code with precise geographic coordinates into the UAV inspection route planning system, accurate mapping of the detection points in digital space is achieved. Combined with three-dimensional geographic information and radio frequency propagation models, this provides a spatial structure and communication link quality prediction basis for route planning. Furthermore, a three-level hierarchical alarm strategy based on the number of electronic tag 51 disconnections is formulated to form a response mechanism. During the inspection planning phase, the UAV inspection route formulation system calculates multi-gradient risk values ​​using a risk coefficient formula and formulates differentiated cruise strategies based on the risk area division results. At the same time, it dynamically adjusts the flight altitude and path based on real-time wind data and uses a genetic algorithm to generate real-time dynamic routes with constraints of minimizing total inspection time, maximizing coverage of high-risk areas, and battery life. This achieves efficient allocation of inspection resources and adaptability to tasks in complex environments. During the inspection execution phase, an RFID reader mounted on a drone is used to automatically collect data from the tag gasket 5, providing hardware support for non-contact testing. During the inspection data analysis phase, sliding window Kalman filtering and phase consistency detection were used to reduce noise and remove invalid data from the electronic tag 51 signal, and time domain features and frequency domain main frequency features were extracted. Then, the signal correlation matrix of the tag group in the same layer and historical data were input into the AI ​​prediction model to achieve accurate prediction of the failure probability and spatial correlation fault risk of a specific electronic tag 51. During the inspection results analysis phase, a tiered alarm strategy is automatically triggered based on AI prediction results, forming a closed-loop automated management system from prediction to action. During the inspection and optimization phase, data streams such as the success rate of reading the electronic tag 51 are input into the system's data warehouse. Incremental learning algorithms and dynamic sliding windows are used to update the weight allocation of the risk coefficient algorithm through gradient descent to minimize prediction errors. At the same time, waypoint spacing is adjusted according to signal strength attenuation patterns, and Pareto optimal routes are generated through multi-objective optimization with battery life and weather disturbances as constraints. Ultimately, this enables the entire system to evolve adaptively, continuously improving the efficiency, reliability, and risk warning capabilities of 51T electronic tag detection.

[0023] Working principle: like Figure 1-6 As shown, this invention utilizes the synergistic effect of the mechanical structure inside the tag washer 5, such as the side washer 54, the arc-shaped spring 55, the elastic support rod 58, the electronic tag 51, and the force-sensitive resistor 59. When the bolt 1 and the nut 2 are tightened, the axial pressure forces the side washer 54 to move closer to each other, compressing the arc-shaped spring 55. After the pressure reaches a certain level, the elastic support rod 58 pushes the force-sensitive resistor 59 out of the movable groove 510 to make it pressurized, causing its resistance to change, thereby modulating the circuit characteristics of the electronic tag 51. Meanwhile, the side telescopic strip 52 provides deformation margin to protect the electronic tag 51, thereby converting the physical preload into a remotely readable signal from the electronic tag 51, realizing indirect and non-destructive detection of the fastener status. Utilizing the pressure-signal strength correlation of tag gasket 5, a quantitative model of the pressure of tag gasket 5 and the signal of electronic tag 51 is established through experiments. Electronic tag 51 is uniquely encoded and bound to geographic coordinates. A drone inspection route planning system is constructed, and a tiered alarm strategy is developed, laying the foundation for subsequent steps. In inspection planning, the system dynamically assesses regional risks using a risk coefficient formula and generates optimized routes. Inspection execution involves automated data collection from tag gasket 5 via a drone equipped with an RFID reader. Inspection data analysis employs sliding window Kalman filtering and phase consistency detection to process the signals and extract time and frequency domain characteristics. The system inputs data into an AI prediction model to predict failure probability and spatial risk. Based on the analysis of inspection results, it automatically executes a tiered alarm strategy, achieving a closed loop from data to action. Inspection optimization continuously updates the risk model and inspection routes through incremental learning and multi-objective optimization. The overall effect is to achieve remote, automated, and real-time detection of the fastener status of power transmission towers. By converting physical fastening forces into measurable electrical signals and using intelligent algorithms for analysis and prediction, it significantly improves detection efficiency and accuracy, provides early warning of loosening risks, optimizes the allocation of maintenance resources, and ultimately enhances the safety of power transmission tower structures and the level of intelligent operation and maintenance.

[0024] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A fastener for remote inspection of power transmission towers and its inspection method, comprising a bolt (1), wherein a nut (2) is threaded onto the bolt (1), and a power transmission tower steel structure (3) is provided on the bolt rod between the nut (2) and the bolt head of the bolt (1), characterized in that: Grooved washers (4) are provided on the bolt (1) between the bolt and the steel structure of the power tower (3) and on the bolt rod between the nut (2) and the steel structure of the power tower (3). Label washers (5) are provided on the bolt rod between the bolt (1) and the grooved washer (4) and on the bolt rod between the nut (2) and the grooved washer (4). The label gasket (5) includes an annular electronic tag (51). Side telescopic strips (52) are fixedly adhered to both sides of the electronic tag (51). The two side telescopic strips (52) are connected by a rigid connecting rod (53). The side of the side telescopic strip (52) away from the electronic tag (51) is fixedly adhered to the outer edge of a side gasket ring (54). The two side gasket rings (54) are connected by an arc-shaped spring piece (55). One of the sleeves ( A sleeve (56) is fixedly welded to the side of the sleeve (56), and an insert (57) is slidably inserted into the sleeve (56). The ends of the sleeve (56) and the insert (57) that are far apart from each other are connected by an elastic support rod (58). A force-sensitive resistor (59) is fixedly installed on the end of the insert (57) that is far away from the sleeve (56). The force-sensitive resistor (59) is movably arranged in a movable groove (510). The movable groove (510) is opened through another sleeve (56).

2. The method for remotely inspecting fasteners for power transmission towers according to claim 1, characterized in that: The number of force-sensitive resistors (59) is four, and the four force-sensitive resistors (59) are connected in series in the circuit of the electronic tag (51).

3. The method for remotely inspecting fasteners for power transmission towers according to claim 1, characterized in that: Therefore, the distance between the two ends of the arc-shaped spring sheet (55) when it is naturally unfolded is greater than the distance between the two ends of the elastic strut (58) when it is naturally unfolded.

4. A method for detecting fasteners for remote inspection of power transmission towers, as described in any one of claims 1-3, characterized in that: The testing process, which includes the label gasket (5) as the tested element, specifically includes the following steps: S1. Inspection Preparation: Construct a drone inspection route planning system, number each electronic tag (51), formulate a graded alarm strategy, and before installing the tag gasket (5), detect the correlation between the pressure of the tag gasket (5) and the signal change. While detecting the correlation between the pressure of the tag gasket (5) and the signal change, simultaneously construct a pressure-signal strength response model. When constructing the pressure-signal strength response model: The pressure test bench applies a gradient preload to the label gasket (5), and simultaneously measures the reading signal strength of the electronic tag (51). The critical switching threshold between the presence / absence of the signal when the label gasket (5) is under pressure and under no pressure is recorded. The direction of the preload applied by the pressure test bench to the label gasket (5) is the same as the radial direction of the label gasket (5). S2. Inspection Planning: The drone inspection route planning system calculates the risk coefficient of the drone inspection task based on historical inspection data, and generates a dynamic drone inspection route based on the RFID signal strength distribution and risk coefficient in the historical inspection data. S3, Inspection Execution: Install RFID readers on the drone and collect data from 5 during the drone's inspection process; S4. Inspection Data Analysis: Process the data read by RFID during the drone inspection process and perform predictive analysis through AI prediction models; S5. Inspection Result Analysis: Execute the corresponding hierarchical alarm strategy based on the results of the predictive analysis. S6. Inspection Optimization: Optimize the drone inspection route planning system based on inspection data.

5. The method for remotely inspecting fasteners on power transmission towers according to claim 4, characterized in that: In step S2, the risk coefficient is calculated as follows: Risk coefficient = weight value 1 * historical anomaly frequency + weight value 2 * wind data + weight value 3 * the floor number of the inspected location; In this calculation, only the values ​​of historical abnormal frequency, wind force data, and the layer number of the inspected location are substituted; the units are not substituted into the calculation.

6. The method for remotely inspecting fasteners on power transmission towers according to claim 5, characterized in that: In step S2, when the UAV inspection route planning system generates a dynamic flight path, it first calculates the reading failure rate of all electronic tags (51) in the last thirty inspections as the historical abnormal frequency. At the same time, it accesses the real-time meteorological data stream to obtain the current wind force data, and calculates the multi-gradient risk coefficients based on the different altitudes of the electronic tags (51). Then, it combines the RFID signal strength spatial distribution heat map recorded in the historical inspections to divide the risk areas into high-risk areas with a risk coefficient greater than 0.7, medium-risk areas with a risk coefficient greater than or equal to 0.3 and less than or equal to 0.7, and low-risk areas with a risk coefficient less than 0.

3. Different cruise strategies are formulated for different risk areas. The UAV inspection route planning system also dynamically adjusts the flight path based on the real-time wind force data, optimizes the UAV antenna orientation using the historical distribution data of RFID signal strength, and adopts an S-shaped enhanced scanning trajectory in the weak signal area. Finally, it solves the multi-objective optimization problem through a genetic algorithm, with the conditions of minimizing the total inspection time, maximizing the coverage of the high-risk area, and battery endurance constraints, outputs the real-time dynamic flight path and injects it into the UAV flight control system.

7. The method for remotely inspecting fasteners on power transmission towers according to claim 4, characterized in that: In step S4, a sliding window Kalman filter is first used to eliminate the multipath effect and UAV attitude fluctuation noise of the received signal strength indication sequence of each electronic tag (51). Invalid data caused by antenna polarization mismatch is eliminated based on phase consistency detection. Then, the data features of the received signal strength mean, variance and attenuation slope in the time domain are extracted from the electronic tag (51) signal. The main frequency characteristics of the electronic tag (51) signal fluctuation are analyzed by fast Fourier transform. At the same time, the electronic tag (51) signal correlation matrix of the same layer tag group is constructed. Combined with the historical data in the UAV inspection route formulation system, the data is input into the AI ​​prediction model. The AI ​​prediction model adopts a hybrid architecture of deep temporal convolutional network and long short-term memory network. The temporal convolutional network extracts local mutation patterns, and the long short-term memory network captures long-term evolution laws. The AI ​​prediction model predicts the failure probability of a specific electronic tag within a certain period of time based on the nonlinear mapping of signal attenuation slope and historical abnormal frequency. The AI ​​prediction model uses graph neural network to analyze the spatial correlation between mechanical transmission path and signal attenuation in the tower structure and predicts spatial correlation failure risk.

8. The method for remotely inspecting fasteners on power transmission towers according to claim 4, characterized in that: In step S6, during the optimization process, the UAV inspection route planning system first inputs the data streams of the UAV collected electronic tag (51) reading success rate, signal strength spatiotemporal distribution, and actual route execution efficiency into the data warehouse of the UAV inspection route planning system. It then uses the incremental learning algorithm to update the risk coefficient algorithm and optimizes the route generation strategy according to the signal strength attenuation mode. With battery life and weather disturbance as constraints, the reward function comprehensively considers the detection rate of high-risk electronic tags (51) and the total inspection time to iteratively generate the Pareto optimal route. The generated Pareto optimal route is then input into the UAV inspection route planning system for training and optimization.

9. The method for remotely inspecting fasteners on power transmission towers according to claim 4, characterized in that: The graded alarm policy in step S1 is formulated by dividing the response into three levels based on the number of disconnections of electronic tags (51): 1) Level 1 alarm: When a single electronic tag (51) cannot be read, a re-check mark is triggered; 2) Level 2 alarm: When multiple electronic tags (51) at the same node cannot be read, manual verification is initiated; 3) Level 3 alarm: If all electronic tags (51) at the same node cannot be read, the emergency maintenance process will be started immediately.

Images