A power grid security assurance monitoring method and system

Abstract

The present application relates to the technical field of power transmission line anti-vibration, and particularly relates to a power grid safety guarantee monitoring method and system. The method comprises the following steps: continuously monitoring the wind vibration amplitude of the power transmission line, and evaluating the vertical vibration degree of the conductor; detecting the spacer bar relaxation degree according to the vertical vibration degree of the conductor; obtaining the image of the four-split power transmission line; inputting the image of the four-split power transmission line into a pre-trained deep learning model to determine the abrasion degree of the conductor strand; predicting the conductor strand breakage probability based on the abrasion degree of the conductor strand; determining the hidden danger point according to the spacer bar relaxation degree and the conductor strand breakage probability; performing the anti-vibration hammer slip analysis according to the hidden danger point to determine the offset degree; adjusting the installation position of the anti-vibration hammer based on the offset degree, calculating the slip friction, and adjusting the spacer bar tightening torque to perform the dynamic anti-vibration task of the power transmission line. The present application is based on the power transmission line anti-vibration technology, and improves the operation safety and fault prevention rate of the power transmission line.

Description

Technical Field

[0001] This invention relates to the field of vibration prevention technology for power transmission lines, and in particular to a power grid safety monitoring method and system. Background Technology

[0002] Current transmission line vibration damping mainly relies on fixed vibration damping devices and experience-based installation methods to suppress wind-induced vibration responses. However, high-voltage and ultra-high-voltage transmission lines have large spans, flexible conductors, and low damping, making them susceptible to various vibration modes such as conductor galloping, secondary span oscillations, and continuous wind vibrations. Traditional static or single-point vibration damping measures cannot simultaneously reflect the amplitude and stress distribution of conductor segments under different wind speeds and vibration conditions, making it difficult to accurately identify conductor fatigue risks or potential relaxation points. The installation of vibration damping devices and the maintenance of spacers largely depend on manual experience, failing to quantify the impact of conductor vibration on hardware relaxation and wear. This leads to problems such as conductor strand wear, spacer relaxation, or vibration damper slippage being easily overlooked or misjudged. Furthermore, the assessment of conductor damping performance and energy attenuation capacity often relies on theoretical calculations or single-test methods, lacking detailed analysis of friction between different conductor strands, the superposition of vibration modes, and high attenuation regions, making it difficult to accurately identify high-risk sections. In terms of dynamic vibration damping optimization, existing systems often employ fixed vibration damping device placement or single-position adjustment strategies, which cannot respond to changes in conductor vibration in real time. The effects of amplitude suppression and energy attenuation are difficult to quantify, and the accuracy of abnormal section identification and vibration damping device optimization placement is insufficient. Transmission line vibration damping systems lack automated, closed-loop vibration monitoring and optimization mechanisms. The adjustment of vibration damping devices and maintenance of spacers largely rely on manual operation, making it impossible to simultaneously achieve dynamic control accuracy, operational efficiency, and line safety. Summary of the Invention

[0003] Therefore, it is necessary for the present invention to provide a power grid security monitoring method and system to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a power grid security monitoring method includes the following steps:

[0005] Step S1: Continuously monitor the wind vibration amplitude of the transmission line and assess the degree of vertical vibration of the conductor; detect the slack of the spacer bar based on the degree of vertical vibration of the conductor.

[0006] Step S2: Obtain images of the four-split transmission line; input the images of the four-split transmission line into a pre-trained deep learning model to determine the degree of wear of the conductor strands; predict the probability of conductor strand breakage based on the degree of wear of the conductor strands;

[0007] Step S3: Determine the potential hazard points based on the slack of the spacer bar and the probability of conductor strand breakage; conduct vibration damper slippage analysis based on the potential hazard points to determine the degree of offset;

[0008] Step S4: Adjust the installation position of the vibration damper based on the degree of offset, calculate the sliding friction force, and adjust the tightening torque of the spacer bar to perform the dynamic vibration damping task of the transmission line.

[0009] Preferably, this specification also provides a power grid security monitoring system for executing the power grid security monitoring method described above, the power grid security monitoring system comprising:

[0010] The wind vibration amplitude monitoring module is used to continuously monitor the wind vibration amplitude of transmission lines and assess the degree of vertical vibration of conductors; it also detects the slack of spacers based on the degree of vertical vibration of conductors.

[0011] The conductor strand breakage probability prediction module is used to acquire images of four-split transmission lines; input the images of four-split transmission lines into a pre-trained deep learning model to determine the degree of wear of conductor strands; and predict the probability of conductor strand breakage based on the degree of wear of conductor strands.

[0012] The hazard point identification module is used to identify hazard points based on the slackness of the spacer bar and the probability of conductor strand breakage; and to perform vibration damper slippage analysis based on the hazard points to determine the degree of offset.

[0013] The sliding friction calculation module is used to adjust the installation position of the vibration damper based on the degree of offset, calculate the sliding friction, and adjust the tightness torque of the spacer bar to perform dynamic vibration damping tasks for transmission lines.

[0014] The beneficial effects of this invention are as follows:

[0015] (1) By real-time monitoring of the wind vibration amplitude of transmission lines and assessment of the vertical vibration of conductors, dynamic perception of the vibration state of the line can be realized, and conductor sections with large amplitude or in vortex-induced resonance can be accurately identified, providing data support for vibration prevention optimization.

[0016] (2) By adopting a comprehensive detection strategy that combines the vertical vibration degree of the conductor with the slack of the spacer, the gap of the spacer fixing bolts and the installation status of the vibration damping device are analyzed in detail. This can identify spacer problems ranging from slight to severe slack and provide a quantitative basis for adjusting the vibration damping device.

[0017] (3) In the process of optimizing the arrangement of vibration damping devices, dynamic adjustments are made by combining the friction position of the vibration damper, the damping coefficient of the conductor and the energy attenuation. This can optimize the position and installation parameters of vibration damping devices for different vibration modes and high attenuation areas, thereby achieving effective control of the conductor vibration amplitude.

[0018] (4) By detecting wire wear, identifying micro-pits and classifying damage levels, a detailed analysis of local damage to conductors can be achieved. Closed-loop optimization can be carried out in combination with the adjustment of vibration damping devices and the maintenance of spacers to improve the safety and stability of transmission line operation. At the same time, it provides repeatable operational basis for the dynamic optimization of vibration damping strategies. Attached Figure Description

[0019] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0020] Figure 1 This is a flowchart illustrating the steps of the power grid security monitoring method of the present invention;

[0021] Figure 2 This is a schematic diagram of wire wear in this invention;

[0022] Figure 3 This is a schematic diagram of a broken strand in the conductor of this invention;

[0023] Figure 4 This is a schematic diagram of the spacer relaxation test results of the present invention;

[0024] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0025] The technical method of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention.

[0026] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0027] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0028] To achieve the above objectives, please refer to Figures 1 to 4This invention provides a power grid security monitoring method, the method comprising the following steps:

[0029] Step S1: Continuously monitor the wind vibration amplitude of the transmission line and assess the degree of vertical vibration of the conductor; detect the slack of the spacer bar based on the degree of vertical vibration of the conductor.

[0030] In one embodiment, a triaxial vibration sensor is installed on the transmission line conductor to continuously acquire vertical vibration signals of the conductor at a sampling frequency of 100Hz. The sensor's built-in acceleration measurement unit obtains the instantaneous acceleration vector, and the instantaneous vibration amplitude and direction of the conductor are calculated. The degree of vertical vibration of the conductor is calculated using the changes in vibration amplitude and direction. Based on the amplitude fluctuations of the vibration signal at the conductor support points, the relative displacement amplitude of each spacer during the vibration process is recorded. The displacement amplitude is compared with a preset relaxation threshold (e.g., 0.5–5 mm), and spacers exceeding the threshold are marked as suspected relaxation zones. The fixing bolts in the suspected relaxation zones are re-measured: bolt gaps ≤1 mm are defined as slight relaxation, 1–3 mm as moderate relaxation, and 3–5 mm as severe relaxation.

[0031] In another embodiment, it is assumed that the vertical vibration data of the conductor is collected continuously for 10 minutes, and the conductor amplitude is between 5 and 18 mm; 12 spacers are measured as suspected relaxation areas, of which 5 are slightly relaxed, 4 are moderately relaxed, and 3 are severely relaxed; the average displacement of the spacers is recorded as: slightly relaxed 2.1 mm, moderately relaxed 2.8 mm, and severely relaxed 4.2 mm.

[0032] Step S2: Obtain images of the four-split transmission line; input the images of the four-split transmission line into a pre-trained deep learning model to determine the degree of wear of the conductor strands; predict the probability of conductor strand breakage based on the degree of wear of the conductor strands;

[0033] In one embodiment, images of a four-split transmission line are captured using a 1080p industrial camera or a drone, and the images are input into a convolutional layer (...). ; ; , A deep learning network consisting of convolutional layers, fully connected layers, and a softmax output layer is used. Convolutional layers extract conductor texture features, pooling layers compress spatial information, and fully connected layers output conductor strand wear classification. Based on the wear classification, the conductor wear ratio is calculated, and the remaining load-bearing capacity is calculated in conjunction with the conductor's rated tensile strength; the probability of strand breakage is predicted based on the remaining load-bearing capacity.

[0034] In another embodiment, assuming 10 images of the conductor are acquired, the deep learning model outputs the following classification of wear levels: 20 strands with slight wear, 15 strands with moderate wear, and 5 strands with severe wear; the probability of strand breakage is calculated based on the remaining load-bearing capacity: slight 0.02, moderate 0.15, and severe 0.38.

[0035] Step S3: Determine the potential hazard points based on the slack of the spacer bar and the probability of conductor strand breakage; conduct vibration damper slippage analysis based on the potential hazard points to determine the degree of offset;

[0036] In one embodiment, the measured spacer slack (slight 1–2 mm, moderate 3–3.5 mm, severe 4–5 mm) and the conductor strand breakage probability calculated in step S2 (slight 0–0.05, moderate 0.1–0.2, severe >0.3) are mapped onto the three-dimensional geometric model of the transmission line to generate a list of risk points for each conductor and corresponding spacer. Vibration damper stress analysis is performed on each risk point, and the friction coefficient is calculated. to Sliding friction under ,in To prevent the normal force (in N) between the vibratory hammer and the conductor, the actual conductor tension is measured. (Unit: kN). Calculate the offset of the vibration damper relative to the conductor using a force analysis model. And record the maximum offset value. This will serve as a reference for adjusting the installation position of the vibration damper in the future.

[0037] In another embodiment, assuming six potential hazard points are identified along the transmission line, the corresponding vibration damper offsets (in mm) are 3.2, 4.5, 2.8, 5.0, 3.7, and 4.1, respectively. The four hazard points with offsets greater than 3.5 mm are marked as high-risk areas and require priority adjustment; the two points with offsets less than 3.5 mm are considered medium-risk and can be addressed within the regular maintenance cycle. The sliding friction force at each point is calculated, assuming a friction coefficient of 0.3 and a conductor normal force of 25 N, resulting in friction forces of 7.5, 7.5, 7.5, 7.5, 7.5, and 7.5 N (for internal reference and model verification).

[0038] Step S4: Adjust the installation position of the vibration damper based on the degree of offset, calculate the sliding friction force, and adjust the tightening torque of the spacer bar to perform the dynamic vibration damping task of the transmission line.

[0039] In one embodiment, the offset is plotted as a conductor offset curve, and the amplitude peaks (maximum displacement points) and troughs (minimum displacement points) in the curve are identified. The center point of the peak and trough is calculated as the anti-vibration hammer ... And combined with conductor tension Perform stress verification; use a torque wrench to adjust the tightening torque of the spacer bars. The damping hammer is set to the design value (e.g., 6.0 Nm) to ensure that it does not slip during vibration, thus completing the dynamic vibration damping task.

[0040] In another embodiment, the peak values ​​(in mm) of the offset curve are assumed to be 5.0, 4.2, 3.8, and 4.5, with valley values ​​of 1.2, 0.8, 1.0, and 1.5, respectively. After calculating the center point position based on the peaks and valleys, the vibration damper is slid to the center point for installation. The actual measured sliding friction forces are 8.2, 7.6, 6.9, and 7.8 N, respectively. The ideal tightening torque calculated based on the conductor tension and friction coefficient is 6.0 Nm. A torque wrench is used to precisely tighten the spacer bolts to complete the dynamic vibration damping task, ensuring that the conductor vibration amplitude is within the allowable range (≤5 mm) and reducing the risk of strand breakage.

[0041] Most importantly, the adjustment of the vibration damper installation position based on the degree of offset in step S4 is specifically as follows:

[0042] The degree of offset is plotted as a conductor offset curve. The maximum value point in the conductor offset curve is identified as the amplitude peak; the minimum value point in the conductor offset curve is identified as the lower amplitude valley.

[0043] In one embodiment, continuous displacement monitoring is performed on the transmission line conductors, collecting lateral offset data along the conductor's length. The collected offset data is plotted as a conductor offset curve to display the conductor's vibration trend. Through curve analysis, local maximum points in the curve are identified and marked as amplitude peaks, representing the highest points of conductor vibration; simultaneously, local minimum points are identified and marked as lower amplitude valleys, representing the lowest points of conductor vibration. These peaks and valleys provide quantitative data on the conductor's vibration characteristics and are used for subsequent installation and adjustment of vibration damping devices.

[0044] In another embodiment, assuming that during continuous monitoring of a section of conductor, the offset curve data has 120 sampling points, and after curve smoothing, five amplitude peaks are detected, with peak values ​​of 22mm, 18mm, 25mm, 20mm, and 23mm respectively; and five amplitude troughs are detected, with trough values ​​of -10mm, -12mm, -9mm, -11mm, and -8mm respectively. These peak and trough positions provide preliminary reference positions for adjusting the vibration damper.

[0045] Calculate the coordinates of the center point between the amplitude peak and the adjacent lower amplitude valley, and use the center point coordinates as the antinode position; slide the anti-vibration hammer to the antinode position to adjust the installation position of the anti-vibration hammer.

[0046] In one embodiment, the coordinates of the center point between the amplitude peak and the adjacent lower amplitude valley are calculated based on the amplitude peak and the valley, and this center point is used as the antinode position. The antinode position represents the intermediate equilibrium point of the conductor vibration and is a key location for the installation or adjustment of the vibration damper. The vibration damper is then slid along the conductor to the corresponding antinode position to complete the installation or fine-tuning, thereby optimizing the conductor vibration suppression effect. After the installation position of the vibration damper is fine-tuned, the conductor offset curve is measured again to ensure that the peak amplitude is effectively suppressed.

[0047] In another embodiment, assuming the measured amplitude peak of the first wave is 22mm and the first lower amplitude trough is -10mm, the corresponding antinode position is approximately 6mm; the second peak is 18mm, the trough is -12mm, and the antinode position is approximately 3mm; the third peak is 25mm, the trough is -9mm, and the antinode position is approximately 8mm. After sliding the vibration damper to these antinode positions, the conductor offset is remeasured, and it is found that the peak amplitude decreases by about 30%–40%, and the peak and trough positions shift by no more than 5mm. This adjustment result can serve as a reference for subsequent vibration damper installation optimization or multi-point vibration damping arrangement.

[0048] Preferably, step S1, which involves continuously monitoring the wind vibration amplitude of the transmission line and assessing the degree of vertical vibration of the conductor, includes:

[0049] Vibration signals of the transmission line in the vertical direction are continuously acquired at a sampling frequency of 100-120Hz; the triaxial acceleration of the transmission line is obtained by using the acceleration measurement unit integrated in the vibration sensor, and the instantaneous vibration vector is calculated;

[0050] In one embodiment, triaxial accelerometers are installed on key conductor sections of the transmission line. Each sensor has a range of ±50g and a sensitivity of 0.002g. The sampling frequency is set to 120Hz to continuously acquire the vibration signal of the conductor in the vertical direction, while simultaneously recording the acceleration in the horizontal direction and along the line. During the acquisition process, the sensors use a low-pass filter to remove high-frequency noise, obtaining a stable vibration signal. Subsequently, based on the triaxial acceleration information output by the sensors, the instantaneous vibration amplitude at each sampling moment is calculated and saved as time-series data for evaluating the conductor vibration amplitude and intensity.

[0051] In another embodiment, it is assumed that six triaxial accelerometers are deployed along the transmission line, with a sampling frequency of 100Hz, continuously collecting data for 10 seconds, and each sensor recording 1000 sampling points. The collected vertical amplitudes (in mm) are 3.2, 4.1, 2.8, 5.0, 3.5, and 4.2; the horizontal amplitudes are 1.2, 1.5, 1.0, 1.8, 1.3, and 1.6. These data are used as conductor vibration characteristics for subsequent vibration risk analysis and vibration damping design.

[0052] The vibration signal and instantaneous vibration vector are used as the wind vibration amplitude; the degree of vertical vibration of the conductor is evaluated based on the wind vibration amplitude.

[0053] In one embodiment, vibration amplitude and instantaneous vibration vector are used as indicators of wind vibration amplitude. Statistical analysis is performed on the vertical vibration amplitude of the conductors, and vibration levels are classified according to amplitude range: minor vibration is defined as no more than 3 mm, moderate vibration as 3–6 mm, and severe vibration as exceeding 6 mm. The evaluation results are mapped onto a three-dimensional model of the transmission line, marking the vibration level of each conductor segment, providing a basis for subsequent spacer slack detection and vibration damping measures.

[0054] In another embodiment, assuming the vertical amplitudes collected by the six sensors are 3.2, 4.1, 2.8, 5.0, 3.5, and 4.2 mm, respectively, the vibration levels are divided into: slight vibration segment 2 (2.8, 3.2 mm), moderate vibration segment 3 (3.5, 4.1, 4.2 mm), and severe vibration segment 1 (5.0 mm). Further analysis of the vibration direction reveals that the vibration direction recorded by sensor number 4 deviates from the vertical direction by approximately 15°, indicating a potential torsional vibration point requiring close monitoring.

[0055] Preferably, assessing the degree of vertical vibration of the conductor based on wind vibration amplitude includes:

[0056] The Karman vortex street frequency formed on the leeward side of the conductor is calculated based on the wind vibration amplitude; the Karman vortex street frequency is compared with the conductor's preset natural frequency to identify the vortex-induced resonance state and assess the degree of vertical vibration of the conductor.

[0057] In one embodiment, vibration sensors and wind speed sensors are deployed on key conductor sections of the transmission line to continuously collect wind vibration amplitude data and conductor response information. Based on the aerodynamic characteristics of the conductor and local wind speed, the Karman vortex street frequency that may occur on the leeward side is calculated and compared with the conductor's preset natural frequency in the vertical direction. When the Karman vortex street frequency is close to the conductor's natural frequency, it is determined to be a vortex-induced resonance state. Subsequently, the actual vertical vibration amplitude of the conductor is combined with the resonance determination result to generate a conductor vibration risk level report, which is used for vibration damping design and spacer installation optimization.

[0058] In another embodiment, assuming six monitoring points are arranged along the transmission line, the measured local wind vibration amplitudes of the conductor are 3.2, 4.0, 3.8, 5.1, 3.5, and 4.3 mm. The Karman vortex street frequencies calculated based on wind speed and conductor geometry are 1.2, 1.5, 1.3, 1.7, 1.4, and 1.6 Hz, respectively, while the conductor's preset vertical natural frequency is 1.5 Hz. Comparison shows that monitoring points 2, 4, and 6 exhibit vortex-induced resonance, with corresponding vertical amplitudes of 4.0, 5.1, and 4.3 mm, respectively, belonging to the medium to high risk area. Other points have lower vibration amplitudes and vortex street frequencies far from the natural frequency, belonging to the low-risk area. This analysis can serve as a reference for subsequent adjustments to the vibration damper installation position and optimization of the spacer bar tightening torque.

[0059] Preferably, identifying the vortex-induced resonance state and assessing the degree of vertical vibration of the conductor includes:

[0060] If the Karman vortex street frequency is close to 5%-10% of the conductor's preset natural frequency, it is judged to be in a vortex-induced resonance state. In the vortex-induced resonance state, the displacement time history signal of the conductor is extracted to assess the degree of vertical vibration of the conductor.

[0061] In one embodiment, vibration sensors and wind speed sensors are installed on key conductor sections of the transmission line to continuously collect data on wind vibration amplitude and conductor response. Based on the collected wind speed data, the Karman vortex street frequency that may occur on the leeward side of the conductor is calculated and compared with the conductor's preset vertical natural frequency. When the Karman vortex street frequency is close to 5%-10% of the natural frequency, it is determined to be a vortex-induced resonance state. Subsequently, the time history signal of the conductor's vertical displacement is obtained from the vibration sensors, and the signal is processed and analyzed to obtain the conductor's maximum vertical displacement, average amplitude, and vibration duration, thereby assessing the degree of vertical vibration of the conductor and providing a reference for subsequent vibration damping design or vibration damper installation.

[0062] In another embodiment, assuming five monitoring points are arranged along the transmission line, the Karman vortex street frequencies of the conductor are calculated to be 1.45, 1.52, 1.48, 1.55, and 1.50 Hz in the continuously collected vibration time histories. Since the conductor's preset vertical natural frequency is 1.50 Hz, the Karman vortex street frequencies at monitoring points 1, 2, and 5 are close to the natural frequency by 5%-10%, indicating vortex-induced resonance. Extracting the conductor's vertical displacement time histories at these points yields maximum displacements of 4.2, 5.0, and 4.5 mm, average amplitudes of 3.1, 3.7, and 3.4 mm, and vibration durations of 12, 15, and 13 s. The remaining monitoring points have smaller vibration amplitudes and vortex street frequencies deviating from the natural frequency, classifying them as low-risk areas. This analysis provides a quantitative basis for prioritizing vibration control measures.

[0063] Preferably, step S1, which involves detecting the slack of the spacer bar based on the degree of vertical vibration of the conductor, includes:

[0064] Record the displacement amplitude of the spacer relative to the conductor during the vibration process based on the vertical vibration degree of the conductor; compare the displacement amplitude with the preset relaxation threshold, and mark the positions of the spacers that exceed the relaxation threshold as suspected relaxation zones;

[0065] In one embodiment, a high-precision displacement sensor is installed on each spacer in the transmission line during conductor vibration. The vertical displacement amplitude of the spacer relative to the conductor is recorded at a frequency of 100Hz, and the amplitude is compared with a preset relaxation threshold (e.g., 5mm). The spacer positions with amplitudes exceeding the threshold are marked as suspected relaxation areas.

[0066] In another embodiment, assuming there are 10 spacers along the conductor segment, the collected vertical displacement amplitude (in mm) is: Based on the threshold of 5mm, the 2nd, 4th, 6th, 9th, and 10th spacers are marked as suspected relaxation areas.

[0067] In the suspected slack area, the gap between the fixing bolts of the spacer is checked. If the gap between the fixing bolts is ≤1mm, it is defined as slight slack; if the gap between the fixing bolts is 1-3mm, it is defined as moderate slack; if the gap between the fixing bolts is 3-5mm, it is defined as severe slack. Slight slack, moderate slack, and severe slack are used as the slack of the spacer.

[0068] In one embodiment, the gap of the spacer fixing bolts is checked in the suspected slack area. A gap ≤1mm is defined as slight slack, a gap of 1–3mm is defined as moderate slack, and a gap of 3–5mm is defined as severe slack. Each level is recorded as the spacer slack degree for subsequent line maintenance and vibration risk assessment.

[0069] In another embodiment, the gaps between the fixing bolts of the 2nd, 4th, 6th, 9th, and 10th spacers are checked respectively. mm corresponds to slight relaxation, moderate relaxation, moderate relaxation, severe relaxation, and slight relaxation, ultimately forming the spacer bar relaxation distribution, providing accurate line health status information, which can be used for subsequent maintenance planning and vibration safety analysis.

[0070] Preferably, after detecting the slack of the spacer bar based on the degree of vertical vibration of the conductor, the method further includes:

[0071] If the loosening is slight, clean the threaded surface of the fixing bolt, apply rust-preventive lubricant, and tighten the fixing bolt to the preset torque using a torque wrench. Then, use a laser rangefinder to correct the elevation of the spacer bar.

[0072] In one embodiment, for spacer bars determined to be slightly loose, the threaded surface of the fixing bolts is first cleaned with a rust-removing brush, and then a rust-preventive lubricant is applied evenly. The fixing bolts are then tightened to a preset torque value (e.g., 50 N·m) using a torque wrench to ensure the spacer bars are securely fixed. Subsequently, a laser rangefinder is used to measure the elevation of the top of the spacer bars, and the bolt torque is fine-tuned if necessary to correct the elevation, ensuring that the vertical deviation of the spacer bars does not exceed ±2 mm.

[0073] In another embodiment, assuming that five slightly slack spacers are detected, the preset torque value is 50 N·m, and the laser rangefinder measures the elevation deviation of each spacer as follows: After fine-tuning the bolts, the final deviation was controlled within ±1mm. After this treatment, the stability of the slightly relaxed rod was essentially consistent with the original design, ensuring safety against conductor vibration.

[0074] If the spacing is moderately loose, add a rubber pad or plastic support plate to the bottom of the spacer and use a support rod to adjust the height of the spacer.

[0075] In one embodiment, for moderately relaxed spacers, a rubber pad or plastic support plate is added to the bottom of the spacer to fill the fixing gap. Then, the height of the spacer is adjusted using a support rod or adjusting bolt to match the design elevation. After the height adjustment is completed, the fixing bolt is tightened to a preset torque (e.g., 70 N·m), and the elevation and levelness are confirmed again using a laser rangefinder.

[0076] In another embodiment, assuming there are three moderately relaxed spacers with original bottom gaps of 1.5mm, 2.0mm, and 2.5mm respectively, after adjusting by adding a rubber pad with a thickness of 1–2mm and a support rod, the height deviations are respectively... mm, and the final fixing torque is controlled at 70 N·m to ensure the spacer bar is stable and eliminate the risk of vibration.

[0077] If the loosening is severe, remove the original fixing bolts, correct the level and elevation of the spacer bar, tighten it to the design torque, and adjust the conductor tension.

[0078] In one embodiment, for severely loose spacers, the original fixing bolts must first be removed, and the spacer's levelness must be corrected to the design elevation using a level and laser rangefinder. After correction, the fixing bolts are reinstalled and tightened to the design torque (e.g., 90 N·m), while simultaneously adjusting the conductor tension to maintain the conductor within the design tension range (e.g., 5–7 kN) between the spacers, ensuring that the conductor vibration amplitude meets the specifications.

[0079] In another embodiment, assuming there are two severely loose spacer bars, the horizontal deviations measured after removing the bolts are 5° and 4.5° respectively. After fine-tuning the support base and bolt installation position, the horizontal deviation is reduced to ≤0.5°. The bolts are tightened to 90 N·m, and the conductor tension is adjusted to 5.5 kN and 6.0 kN, restoring the overall uniformity of the conductor force and eliminating potential safety risks.

[0080] Preferably, step S2, which involves inputting the image of the four-split transmission line into a pre-trained deep learning model to determine the degree of wear on the conductor strands, includes:

[0081] Images of four-split transmission lines are input into a pre-trained deep learning model to determine the degree of wear on conductor strands; the deep learning model includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer;

[0082] In one embodiment, high-resolution images (e.g., 2048×2048 pixels) of a four-split transmission line are input into a pre-trained deep learning model. This model includes an input layer, convolutional layers, pooling layers, fully connected layers, and an output layer. Trained on previously acquired large-scale transmission line images, the model can identify surface wear textures, break points, and corrosion marks on the conductor strands. The model outputs a wear level (slight, moderate, severe) for each conductor strand, and subsequent algorithms can calculate the probability of conductor breakage.

[0083] In another embodiment, assuming the input image contains 4 wire strands, the model predicts the wear level of each wire strand as slight (15%), moderate (40%), severe (70%), and moderate (35%), with corresponding predicted strand breakage probabilities of 0.02, 0.12, 0.35, and 0.10, respectively, providing a basis for subsequent operation and maintenance decisions.

[0084] The input layer defines the size of the input image for the model and performs normalization. The convolutional layer uses the convolutional kernel to perform sliding calculations based on the normalized image output from the input layer, extracts texture features, and stacks them layer by layer to form a feature map of the transmission line.

[0085] In one embodiment, the input layer normalizes the original image size to the input size required by the model (e.g., 512×512×3) and normalizes the pixel values ​​(range 0–1). Convolutional layers use 3×3 convolutional kernels to slide across the image, extracting surface texture features of the conductors. Convolutional layers are stacked layer by layer to form a feature map, from low-level edge information to high-level conductor texture and strand breakage patterns.

[0086] In another embodiment, it is assumed that a total of 5 convolutional layers are used, with the number of channels in each convolutional kernel being 32, 64, 128, 256, and 512, respectively. After convolution processing, the input image of 512×512×3 generates feature maps with sizes of 512×512×32, 256×256×64, 128×128×128, 64×64×256, and 32×32×512, respectively, which are used for subsequent pooling processing.

[0087] The pooling layer performs max pooling or average pooling operations on the feature map of the transmission line to compress the spatial dimension and retain the conductor feature information.

[0088] In one embodiment, max pooling or average pooling (e.g., 2×2 windowing) is performed on the feature map output by the convolutional layer to compress the spatial dimension while preserving the significant texture information and local broken strand features of the conductor strands. The pooled feature map reduces computation and enhances the model's robustness to noise and image shifts.

[0089] In another embodiment, assuming the convolutional feature map size is 128×128×128, a 64×64×128 feature map is obtained after 2×2 max pooling. The pooling operation retains approximately 90% of the information of the main wear texture points on the wire surface, while reducing noise interference by approximately 30%, ensuring the accuracy of subsequent classification.

[0090] The fully connected layer expands the wire feature information output by the pooling layer into a one-dimensional vector and transmits it to the output layer to perform wire strand wear classification and determine the degree of wire strand wear; based on the degree of wire strand wear, the probability of wire strand breakage is predicted.

[0091] In one embodiment, the feature map output from the pooling layer is unfolded into a one-dimensional vector and input into a fully connected layer. The fully connected layer integrates all local feature information and transmits it to the output layer to perform conductor strand wear classification, outputting the wear level (slight, moderate, severe) and corresponding breakage probability of each conductor strand. Based on the prediction results, maintenance plans or safe operation warnings can be formulated.

[0092] In another embodiment, assuming the pooled feature map size is 64×64×128, it unfolds into a 524,288-dimensional one-dimensional vector and inputs it into a fully connected layer. The output of the fully connected layer has four wire strands with wear probabilities of being slight. ,medium: serious: Based on this, the probabilities of breakage are calculated to be 0.02, 0.12, 0.35, and 0.10, respectively, providing data support for transmission line maintenance decisions.

[0093] Most importantly, after the fully connected layer expands the wire feature information output by the pooling layer into a one-dimensional vector, it also includes:

[0094] A one-dimensional vector is input into a pre-defined softmax layer, and the softmax function is used for classification. The expression for the softmax function is:

[0095] ;

[0096] In the formula, For the output of the model, Input vector, For the number of categories, For input The first one obtained after linear transformation Scores for each category.

[0097] Preferably, predicting the probability of conductor strand breakage based on the degree of wear of the conductor strands includes:

[0098] The wear cross-sectional area is calculated based on the degree of wear of the conductor strands, and the wear ratio is determined; the wear ratio is compared with the preset tensile strength to determine the remaining load-bearing capacity of the conductor.

[0099] In one embodiment, the degree of wear on each conductor strand is quantified. First, based on preliminary image recognition results, the conductor strands are classified into slight, moderate, or severe wear levels. Then, the effective cross-sectional area of ​​the conductor strands is adjusted according to the degree of wear; for example, slightly worn strands retain most of their cross-sectional area, moderately worn strands retain approximately two-thirds, and severely worn strands retain approximately half. Finally, combined with the standard tensile properties of the conductor material, the remaining load-bearing capacity of the conductor is determined, and the load-bearing level of each conductor strand is recorded, providing basic data for subsequent strand breakage risk assessment.

[0100] In another embodiment, it is assumed that the slight wear of a conductor strand accounts for approximately 12% of the total cross-sectional area, moderate wear accounts for 25%, and severe wear accounts for 40%. After adjustment, the remaining carrying capacity of each conductor strand corresponds to three levels: low, medium, and high. Through statistical analysis, the slightly worn conductor strand can withstand normal transmission tension, the moderately worn conductor strand is close to the warning state, and the severely worn conductor strand has a significantly reduced carrying capacity, requiring priority inspection or replacement.

[0101] Calculate the critical tension for strand breakage based on the remaining carrying capacity of the conductor; detect the degree of surface damage of the strands based on the critical tension for strand breakage; predict the probability of strand breakage of the conductor based on the degree of surface damage of the strands.

[0102] In one embodiment, the remaining load-bearing capacity of each conductor strand is used to determine its safety under the current tension. For conductor strands with less wear, the tension they can withstand is significantly higher than the actual operating tension; for conductor strands with more wear, the tension they can withstand is close to or lower than the actual operating tension. Subsequently, the probability of conductor strand breakage is comprehensively judged based on the damage to the conductor strand surface, including surface cracks, corrosion, and notches, and is marked as high, medium, and low, providing a basis for prioritizing transmission line maintenance and inspection.

[0103] In another embodiment, it is assumed that the breakage probability of a slightly worn conductor strand is approximately 2%, that of a moderately worn conductor strand is approximately 12%, and that of a severely worn conductor strand is approximately 35%. Based on inspection data, the breakage probability of different conductor strands is plotted into a risk distribution table, facilitating the rapid identification of high-risk conductor strands and the scheduling of maintenance or replacement.

[0104] Preferably, the method of detecting the degree of surface damage of the strand based on the critical tension of the broken strand includes:

[0105] Identify the friction points of adjacent strands based on the critical tension of the broken strands; calculate the conductor damping coefficient based on the friction points of adjacent strands; determine the energy attenuation based on the conductor damping coefficient.

[0106] In one embodiment, continuous monitoring is performed on each conductor strand of a four-split transmission line. First, tension or strain sensors are used to determine the conductor's operating state, and the friction locations of adjacent strands are identified based on the critical tension at which the strand breaks. Then, conductor motion characteristics, including vibration amplitude and frequency, are collected at the friction locations, and the local damping coefficient of the conductor is calculated based on the conductor material properties. The energy attenuation of the conductor under normal operation or wind-load vibration is further determined based on the conductor damping coefficient, forming an attenuation distribution map for each conductor strand, providing a reference for subsequent damage detection.

[0107] In another embodiment, assuming a conductor strand under typical wind load conditions, three friction locations are identified: location A, location B, and location C. The measured damping coefficients are 0.12, 0.18, and 0.22, respectively, corresponding to energy attenuations of approximately 2.5 J, 3.8 J, and 4.1 J. Through statistical analysis, friction locations with attenuations greater than 3 J are marked as high-attenuation locations for subsequent image acquisition and micro-dimple detection.

[0108] Images of strands with high energy attenuation are acquired based on the energy attenuation amount; micro-dimples in the strands are identified based on the high energy attenuation images, and the depth of the micro-dimples is detected to determine the degree of surface damage to the strands.

[0109] In one embodiment, high-resolution industrial cameras (such as 5-megapixel line scan cameras) are used to acquire images of the wire strands at high-attenuation locations. The acquired images undergo local enhancement processing, including contrast enhancement and edge sharpening, to highlight the micro-dimple features on the strand surface. Subsequently, the location, number, and distribution of the micro-dimples are identified on the enhanced images, and the depth of the micro-dimples is detected through depth measurement or optical profile analysis. This determines the degree of surface damage to the strand, forming a strand damage level assessment, and providing a basis for predicting strand breakage risk.

[0110] In another embodiment, it is assumed that 12 micro-pits were detected in images acquired at high attenuation locations, distributed across different winding layers of the conductor strand. Optical profile analysis revealed that the depth of the micro-pits ranged from 0.3 mm to 1.2 mm, with 3 pits exceeding 1 mm in depth. Based on the depth and distribution of the micro-pits, the surface damage to the conductor strands was categorized into mild (depth ≤ 0.5 mm, 5 pits), moderate (0.5–1.0 mm, 4 pits), and severe (depth > 1.0 mm, 3 pits), providing a quantitative basis for subsequent strand breakage probability assessment and inspection prioritization.

[0111] Of particular importance is the identification of micro-dimples in stock lines based on high attenuation stock line images, including:

[0112] Color channel enhancement is performed on the high-attenuation strand image, and edge detection is performed on the enhanced high-attenuation strand image to identify the surface contour edge of the strand.

[0113] In one embodiment, for the acquired high-attenuation strand image, the three RGB color channels are first independently enhanced, with the enhancement magnitude adaptively adjusted according to the grayscale distribution of the strand surface to improve the contrast of the micro-dimples. Then, the Canny edge detection algorithm is applied to the enhanced image to detect the contour edges of the strand surface, while setting high and low thresholds of 120 and 60 respectively to suppress noise interference. After edge detection is completed, a pixelated boundary map of the strand contour is obtained, which can be used for subsequent dimple localization and depth analysis.

[0114] In another embodiment, assuming the acquired high-attenuation strand image has a resolution of 2048×1024 pixels, and the R / G / B average values ​​are increased by 15%, 12%, and 18% respectively after color channel enhancement, Canny edge detection yields a strand contour with a total length of approximately 13,500 pixels, consisting of 78 continuous contour segments, each ranging in length from 50 to 250 pixels. These contour segments provide preliminary location references for micro-dimples, preparing for subsequent precise depth measurements.

[0115] The surface contour edge of the strand is converted into binary information, the edge gaps are filled, and continuous pixel blocks are marked to determine the micro-dimples of the strand.

[0116] In one embodiment, the contour edge information is converted into a binary image, and morphological closing operations are performed to fill the gaps in the contour. Subsequently, connected region labeling is performed on consecutive pixel blocks, with each connected region considered a potential micro-pit. For each connected region, its center point coordinates, area, and aspect ratio are calculated to preliminarily determine the shape characteristics of the pit. Pixels within consecutive blocks that conform to a predetermined area range (e.g., ...) are considered potential micro-pits. Furthermore, areas with an aspect ratio of 0.8–1.2 are identified as valid micro-pits, and their spatial locations are recorded for surface damage analysis.

[0117] In another embodiment, assuming that after binarization and closing operations, a total of 92 connected regions were detected, of which 63 were effectively marked micro-pits, with an area distribution of [missing information]. The average area is approximately The aspect ratio is between 0.85 and 1.15. The center point coordinates of each micro-dimple are randomly distributed within a range of 50–900 pixels along the strand length in the image coordinate system. These markers allow for further calculation of the strand surface damage density, which can be used for conductor life assessment and protection optimization design.

[0118] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0119] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A power grid security monitoring method, characterized in that, Includes the following steps: Steps S1: Continuously monitor the wind vibration amplitude of the transmission line and assess the degree of vertical vibration of the conductor; The slack of the spacer bar is detected based on the degree of vertical vibration of the conductor. This process includes: if the slack is slight, cleaning the threaded surface of the fixing bolt, applying anti-rust lubricant, and tightening the bolt to the preset torque using a torque wrench; then, using a laser rangefinder to correct the spacer bar elevation. If the slack is moderate, adding a rubber pad or plastic support plate to the bottom of the spacer bar and adjusting its height using a support rod. If the slack is severe, removing the original fixing bolt, correcting the levelness and elevation of the spacer bar, tightening it to the design torque, and adjusting the conductor tension. Step S2: Acquire an image of the four-split transmission line; input the image of the four-split transmission line into a pre-trained deep learning model to determine the degree of conductor strand wear; predict the probability of conductor strand breakage based on the degree of conductor strand wear; the prediction of the probability of conductor strand breakage based on the degree of conductor strand wear includes: The wear cross-sectional area is calculated based on the degree of conductor strand wear, and the wear ratio is determined. The wear ratio is compared with the preset tensile strength to determine the remaining carrying capacity of the conductor. The critical tension for strand breakage is calculated based on the remaining carrying capacity of the conductor. The degree of surface damage of the strands is detected based on the critical tension for strand breakage. The probability of conductor strand breakage is predicted based on the degree of surface damage of the strands. Among them, the detection of the degree of surface damage of the strands based on the critical tension for strand breakage includes: identifying the friction position of adjacent strands based on the critical tension for strand breakage; calculating the conductor damping coefficient based on the friction position of adjacent strands; determining the energy attenuation based on the conductor damping coefficient; acquiring images of strands with high attenuation based on the energy attenuation; identifying strand micro-dimples based on the images of strands with high attenuation and detecting the depth of strand micro-dimples to determine the degree of surface damage of the strands; Step S3: determining the hidden danger points based on the slack of the spacer bar and the probability of conductor strand breakage; performing vibration damper slippage analysis based on the hidden danger points to determine the degree of offset; Step S4: adjusting the installation position of the vibration damper based on the degree of offset, calculating the slip friction force, and adjusting the tightening torque of the spacer bar to perform the dynamic vibration damping task of the transmission line.

2. The power grid security monitoring method according to claim 1, characterized in that, Step S1 involves continuously monitoring the wind vibration amplitude of the transmission line and assessing the degree of vertical vibration of the conductor, including: continuously acquiring the vertical vibration signal of the transmission line at a sampling frequency of 100-120Hz; obtaining the triaxial acceleration of the transmission line using the acceleration measurement unit integrated with the vibration sensor and calculating the instantaneous vibration vector; using the vibration signal and the instantaneous vibration vector as the wind vibration amplitude; and assessing the degree of vertical vibration of the conductor based on the wind vibration amplitude.

3. The power grid security monitoring method according to claim 2, characterized in that, Assessing the degree of vertical vibration of a conductor based on wind vibration amplitude includes: calculating the Karman vortex street frequency formed on the leeward side of the conductor based on wind vibration amplitude; comparing the Karman vortex street frequency with the conductor's preset natural frequency to identify the vortex-induced resonance state and assess the degree of vertical vibration of the conductor.

4. The power grid security monitoring method according to claim 3, characterized in that, Identifying vortex-induced resonance and assessing the degree of vertical vibration of the conductor includes: If the Karman vortex street frequency is close to 5%-10% of the conductor's preset natural frequency, it is judged to be in a vortex-induced resonance state. In the vortex-induced resonance state, the displacement time history signal of the conductor is extracted to assess the degree of vertical vibration of the conductor.

5. The power grid security monitoring method according to claim 1, characterized in that, Step S1, which involves detecting the slack of the spacer bar based on the degree of vertical vibration of the conductor, includes: Record the displacement amplitude of the spacer relative to the conductor during the vibration process based on the vertical vibration degree of the conductor; compare the displacement amplitude with the preset relaxation threshold, and mark the positions of the spacers that exceed the relaxation threshold as suspected relaxation zones; In the suspected slack area, the gap between the fixing bolts of the spacer is checked. If the gap between the fixing bolts is ≤1mm, it is defined as slight slack; if the gap between the fixing bolts is 1-3mm, it is defined as moderate slack; if the gap between the fixing bolts is 3-5mm, it is defined as severe slack. Slight slack, moderate slack, and severe slack are used as the slack of the spacer.

6. The power grid security monitoring method according to claim 1, characterized in that, Step S2, inputting the image of the four-split transmission line into a pre-trained deep learning model to determine the wear degree of the conductor strands, includes: inputting the image of the four-split transmission line into a pre-trained deep learning model to determine the wear degree of the conductor strands; wherein the deep learning model includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer; wherein the input layer is used to define the input image size of the model and perform normalization operations; the convolutional layer, based on the normalized image output by the input layer, uses the convolutional kernel to perform sliding calculations, extracts texture features, and stacks them layer by layer to form a transmission line feature map; the pooling layer performs max pooling or average pooling operations on the transmission line feature map to compress the spatial dimension and retain conductor feature information; the fully connected layer unfolds the conductor feature information output by the pooling layer into a one-dimensional vector and transmits it to the output layer to perform conductor strand wear classification and determine the wear degree of the conductor strands; and predicts the probability of conductor strand breakage based on the wear degree of the conductor strands.

7. A power grid security monitoring system, characterized in that, For performing the power grid security monitoring method as described in claim 1, the power grid security monitoring system includes: The wind vibration amplitude monitoring module is used to continuously monitor the wind vibration amplitude of transmission lines and assess the degree of vertical vibration of conductors; it also detects the slack of spacers based on the degree of vertical vibration of conductors. The conductor strand breakage probability prediction module is used to acquire images of four-split transmission lines; input the images of four-split transmission lines into a pre-trained deep learning model to determine the degree of wear of conductor strands; and predict the probability of conductor strand breakage based on the degree of wear of conductor strands. The hazard point identification module is used to identify hazard points based on the slackness of the spacer bar and the probability of conductor strand breakage; and to perform vibration damper slippage analysis based on the hazard points to determine the degree of offset. The sliding friction calculation module is used to adjust the installation position of the vibration damper based on the degree of offset, calculate the sliding friction, and adjust the tightness torque of the spacer bar to perform dynamic vibration damping tasks for transmission lines.

Images