Monitoring Method for Wind-induced Galloping of Transmission Lines by Integrating Crosshair Laser and Monocular Vision

Abstract

The present invention discloses a monitoring method for wind-induced swing and galloping of transmission lines by integrating crosshair lasers and monocular vision, which relates to the technical field of power system monitoring and includes: projecting lasers onto the transmission lines through a crosshair laser, continuously collecting the laser profiles of the lines within a predetermined window by using a monocular camera to obtain a set of contour images; performing binarization processing and contour feature extraction on the set of contour images, and fusing to obtain the line contour features; performing wind-induced swing and galloping analysis based on the line contour features, and outputting wind-induced swing and galloping data; determining whether the wind deviation angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle meet the dynamic monitoring thresholds, and if so, giving an early warning of the wind-induced swing and galloping of the line. The present invention solves the technical problem of insufficient monitoring accuracy of the wind-induced swing and galloping of transmission lines in the prior art, and achieves the technical effect of improving the monitoring accuracy of the wind-induced swing and galloping of transmission lines.

Description

Technical Field

[0001] The present invention relates to the technical field of power system monitoring, and particularly relates to a monitoring method for wind-induced deflection and galloping of transmission lines by integrating cross-line laser and monocular vision. Background Art

[0002] With the continuous expansion of modern power transmission networks, the safe and stable operation of transmission lines has become one of the important issues in the power system. During long-distance power transmission, transmission lines are usually exposed to complex external environments, such as wind, ice and snow, temperature changes, etc. These environmental factors may cause wind-induced deflection and galloping of transmission lines. Wind-induced deflection and galloping refer to the swing or vibration of transmission lines under the action of factors such as wind. This phenomenon not only affects the operation stability of the lines, but may also cause mechanical fatigue of the lines and even lead to serious faults such as line breakage. Therefore, it is particularly important to monitor the wind-induced deflection and galloping of transmission lines in real time.

[0003] Most of the existing wind-induced deflection and galloping monitoring technologies rely on sensors installed on towers or lines for vibration detection. Although such technologies can obtain some monitoring data, they have limitations, resulting in insufficient monitoring accuracy for wind-induced deflection and galloping of transmission lines. Summary of the Invention

[0004] The present application provides a monitoring method for wind-induced deflection and galloping of transmission lines by integrating cross-line laser and monocular vision, which is used to solve the technical problem of insufficient monitoring accuracy of wind-induced deflection and galloping of transmission lines in the existing technology.

[0005] In view of the above problems, the present application provides a monitoring method for wind-induced deflection and galloping of transmission lines by integrating cross-line laser and monocular vision.

[0006] The present application provides a monitoring method for wind-induced deflection and galloping of transmission lines by integrating cross-line laser and monocular vision, and the method includes:

[0007] Project a laser onto the transmission line through a cross-line laser, continuously collect the laser contour of the line within a predetermined window by using a monocular camera to obtain a contour image set; perform binarization processing and contour feature extraction on the contour image set, and fuse to obtain the line contour feature; perform wind-induced deflection and galloping analysis based on the line contour feature, and output wind-induced deflection and galloping data, where the wind-induced deflection and galloping data includes wind-induced deflection angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude and elliptical tilt angle; determine whether the wind-induced deflection angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude and elliptical tilt angle meet the dynamic monitoring threshold, and if so, give a warning of wind-induced deflection and galloping of the line, where the dynamic monitoring threshold is adaptively adjusted based on the tower state.

[0008] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0009] In this application, a cross-line laser is used to project laser onto a transmission line, and a monocular camera is used to continuously collect the line laser contour within a predetermined window to obtain a contour image set; the contour image set is subjected to binary processing and contour feature extraction, and the line contour features are fused; wind deviation and galloping analysis is performed based on the line contour features, and wind deviation and galloping data is output, where the wind deviation and galloping data includes wind deviation angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle; it is determined whether the wind deviation angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle meet the dynamic monitoring threshold, and if so, a warning for line wind deviation and galloping is given, where the dynamic monitoring threshold is adaptively adjusted based on the tower state. The present invention solves the technical problem of insufficient monitoring accuracy of wind deviation and galloping of transmission lines in the prior art. By projecting with a cross-line laser and collecting line contour images in real time with a monocular camera, binary processing and feature extraction are performed, the wind deviation and galloping data of the line are analyzed, and it is determined whether parameters such as the wind deviation angle and galloping amplitude exceed the dynamic monitoring threshold, achieving the technical effect of improving the monitoring accuracy of wind deviation and galloping of transmission lines. Description of the Drawings

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0011] Figure 1 It is a schematic flowchart of a method for monitoring wind deviation and galloping of a transmission line by fusing cross-line laser and monocular vision provided by an embodiment of this application;

[0012] Figure 2 It is a schematic flowchart of configuring a dynamic monitoring threshold in a method for monitoring wind deviation and galloping of a transmission line by fusing cross-line laser and monocular vision provided by an embodiment of this application. Detailed Embodiments

[0013] This application provides a method for monitoring wind deviation and galloping of a transmission line by fusing cross-line laser and monocular vision, which is used to solve the technical problem of insufficient monitoring accuracy of wind deviation and galloping of transmission lines in the prior art. By projecting with a cross-line laser and collecting line contour images in real time with a monocular camera, binary processing and feature extraction are performed, the wind deviation and galloping data of the line are analyzed, and it is determined whether parameters such as the wind deviation angle and galloping amplitude exceed the dynamic monitoring threshold, achieving the technical effect of improving the monitoring accuracy of wind deviation and galloping of transmission lines.

[0014] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the scope of protection of the present application.

[0015] It should be noted that any variations of the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products, or devices.

[0016] As Figure 1 shown, the present application provides a monitoring method for wind-induced galloping of transmission lines by integrating crosshair laser and monocular vision. The method includes:

[0017] Step S100: Project laser onto the transmission line through a crosshair laser, and continuously collect the line laser contour within a predetermined window using a monocular camera to obtain a contour image set.

[0018] In the embodiments of the present application, first, a crosshair laser projects laser onto the transmission line. When projecting the laser, the crosshair laser emits two mutually perpendicular laser lines, forming a cross-shaped light spot. This light spot produces a clear laser contour on the transmission line. Then, a monocular camera installed at a predetermined position continuously collects the contours after laser projection on the line. The camera captures these contour images within its predetermined observation range, and finally forms a contour image set containing multiple frames of images.

[0019] Step S200: Perform binarization processing and contour feature extraction on the contour image set, and fuse to obtain the line contour feature.

[0020] In the embodiments of the present application, the collected contour image set is subjected to binarization processing to convert the images into images with only two colors, black and white. After completing the binarization processing, edge features of the binary image set are extracted to obtain multiple contour line segments. Finally, fitting is performed according to the order of the contour line segments to obtain the complete line contour feature.

[0021] Furthermore, in the method provided by the embodiments of the application, performing binarization processing and contour feature extraction on the contour image set, and fusing to obtain the line contour feature, further includes:

[0022] Perform grayscale processing on the contour image set to obtain a grayscale contour image set; configure a grayscale judgment scalar, and perform binary conversion on the grayscale contour image set based on the grayscale judgment scalar to obtain a binary image set; sequentially perform edge feature extraction on the binary image set to obtain a plurality of contour line segments, fit the plurality of contour line segments in sequence, and output the line contour feature.

[0023] In the embodiment of the present application, first, an image grayscale algorithm is used. For example, each pixel is grayscaled based on the weighted average method. By the weighted average method, the values of the red, green, and blue channels in the color image are used to calculate the brightness value of each pixel according to a preset weight, thereby generating a single-channel grayscale image, which is called a grayscale contour image set. Next, a grayscale judgment scalar is configured according to a preset threshold, and all pixel points greater than the threshold are set to white (foreground), and pixel points less than the threshold are set to black (background). Then, based on the configured grayscale judgment scalar, the grayscale contour image set is binarized. By using a global binarization algorithm, the grayscale image is converted into a black and white binary image to obtain a binary image set.

[0024] Subsequently, edge feature extraction is performed on the binary image set. The image is Gaussian filtered through the Canny edge detection algorithm to smooth the image, and then by calculating the gradient magnitude and direction of the image, the mutated parts in the image are identified, and then a plurality of contour line segments are extracted.

[0025] Finally, a polyline fitting algorithm is used to fit the plurality of contour line segments, connect the extracted contour line segments in sequence to form a continuous contour line, and output the complete line contour feature.

[0026] Step S300: Perform wind swing analysis based on the line contour feature and output wind swing data, where the wind swing data includes a wind deviation angle, a swing amplitude, a vertical swing amplitude, a horizontal swing amplitude, and an ellipse tilt angle.

[0027] In the embodiment of the present application, a line twin space is constructed based on standard tower structure data and benchmark line state simulation, and the line contour feature is imported into the twin space for wind swing analysis. During the analysis process, the wind deviation angle is obtained through geometric calculation, the vertical and horizontal swings are analyzed separately to determine the swing amplitude, and the ellipse tilt angle is calculated by combining the historical line contour feature sequence. The ellipse tilt angle is obtained by fitting the ellipse of the swing trajectory and calculating the tilt angle of its major axis relative to the horizontal line. Through the above process, wind swing data is obtained, where the wind swing data includes a wind deviation angle, a swing amplitude, a vertical swing amplitude, a horizontal swing amplitude, and an ellipse tilt angle.

[0028] Further, in the method provided by the application embodiment, when performing galloping analysis based on the line profile features and outputting galloping data, it further includes:

[0029] Construct a line twin space based on standard tower structure data and benchmark line state simulation; import the line profile features into the line twin space for galloping analysis and output galloping data.

[0030] In the embodiment of the present application, first, the finite element analysis technology is used to construct a three-dimensional tower model by inputting standard tower structure data, and at the same time, the simulation data of the benchmark line state is combined as a reference standard for galloping analysis. Through this process, a line twin space including the tower and the benchmark line is constructed. Among them, the standard tower structure data includes the geometric structure data and material properties of the standard tower, which are pre-stored. The benchmark line state is the state of the line under no wind or slight wind, which is also pre-stored.

[0031] Then, the line profile features extracted from the actual monitoring are exported by an image processing method, such as Canny edge detection, and these profile features are imported into the constructed twin space using data mapping technology. After import, these profile features are matched with the benchmark line state to reflect the actual morphological changes of the transmission line under galloping conditions.

[0032] In the twin space, by comparing the differences between the actual line profile features and the benchmark state, the galloping angle of the line, that is, the deviation angle of the line relative to the benchmark state, is obtained through geometric calculation. At the same time, galloping analysis in the vertical and horizontal directions is performed to calculate the vertical galloping amplitude and the horizontal galloping amplitude, and the overall galloping amplitude is obtained through vector synthesis. Finally, based on historical data, an ellipse fitting algorithm is used to fit the galloping trajectory and calculate the ellipse tilt angle. Finally, the output galloping data includes the galloping angle, the galloping amplitude, the vertical galloping amplitude, the horizontal galloping amplitude, and the ellipse tilt angle.

[0033] Further, in the method provided by the application embodiment, when outputting galloping data, it further includes:

[0034] In the line twin space, perform offset analysis on the line profile features and the benchmark line state, and obtain the galloping angle through geometric calculation; perform vertical galloping and horizontal galloping analysis based on the line profile features and the benchmark line state respectively to determine the vertical galloping amplitude and the horizontal galloping amplitude, and combine the vertical galloping amplitude and the horizontal galloping amplitude to obtain the galloping amplitude; call the historical line profile feature sequence within a preset period, and calculate the ellipse tilt angle based on the line profile features and the historical line profile feature sequence; compose the galloping data according to the galloping angle, the galloping amplitude, the vertical galloping amplitude, the horizontal galloping amplitude, and the ellipse tilt angle.

[0035] In the embodiment of the present application, in the line twin space, an offset analysis is performed on the line profile features and the reference line state. Specifically, in the twin space, the least squares method or the Euclidean distance calculation is used to analyze the geometric offset between the actual profile and the reference state. By calculating the displacement differences of each profile point in the vertical and horizontal directions relative to the reference state, the wind deflection angle of the transmission line under the action of wind is obtained, that is, the offset angle of the line relative to its reference state. For example, assuming that the inclination angle of the reference line is 0 degrees and the actually measured offset is 10 degrees, then the wind deflection angle is 10 degrees.

[0036] Next, based on the actual profile features, the FFT is used to analyze the vibration frequency of the line. The frequency components are extracted from the time series data through the FFT, so as to analyze the dancing amplitude and periodic changes of the line. After the vibration analysis, the vertical dancing amplitude and the horizontal dancing amplitude are calculated respectively. The vertical dancing amplitude is calculated by measuring the displacement range of the line in the vertical direction, while the horizontal dancing amplitude is calculated by measuring the displacement fluctuation range of the line in the horizontal direction. Assuming that the vertical dancing amplitude is 2 meters and the horizontal dancing amplitude is 1 meter, the vector synthesis method is used to combine the vertical and horizontal dancing amplitudes to obtain the total dancing amplitude of the line, which represents the comprehensive amplitude of the line in all directions.

[0037] To further analyze the dynamic dancing characteristics of the line, the historical line profile feature sequence within a preset period is called. By comparing with the current line profile features, the ellipse fitting algorithm is used to analyze the dancing trajectory, and the elliptical dancing trajectory of the line is fitted. Through the ellipse fitting method, such as the least squares method or the RANSAC algorithm, the motion trajectory of the line is fitted. After the fitting is completed, the ellipse inclination angle is calculated, that is, the inclination angle of the major axis of the ellipse relative to the horizontal line. Assuming that the fitted ellipse inclination angle is 20 degrees, this indicates that the dancing of the line has an obvious directional offset, and the magnitude of the inclination angle reflects the directional change of the dancing trajectory of the line.

[0038] Through the above process, the wind deflection angle, the dancing amplitude, the vertical dancing amplitude, the horizontal dancing amplitude, and the ellipse inclination angle are obtained. These data together constitute the wind deflection and dancing data. Among them, the obtained wind deflection angle, such as 10 degrees, represents the offset angle of the line relative to the reference state under the action of wind. The dancing amplitude, such as 2 meters, represents the total amplitude of the line. The vertical dancing amplitude, such as 1.5 meters, represents the vibration range of the line in the vertical direction. The horizontal dancing amplitude, such as 1 meter, represents the vibration range of the line in the horizontal direction. The ellipse inclination angle, such as 15 degrees, represents the inclination angle of the major axis of the fitted ellipse relative to the horizontal line.

[0039] Furthermore, in the method provided by the application embodiment, calculating the ellipse inclination angle according to the line profile features and the historical line profile feature sequence further includes:

[0040] Perform ellipse fitting on the dancing trajectory based on the line profile feature and the historical line profile feature sequence, and output the fitted elliptical trajectory; calculate the tilt angle based on the fitted elliptical trajectory to obtain the elliptical tilt angle, where the elliptical tilt angle is the tilt angle of the major axis of the ellipse relative to the horizontal line.

[0041] In the embodiment of the present application, perform ellipse fitting on the dancing trajectory based on the acquired line profile feature and the historical line profile feature sequence stored in the historical database. By using the least squares method to process the current line profile feature and the historical line profile feature sequence, the dancing trajectory of the line is fitted. Specifically, the profile feature points at each moment are fitted into an elliptical trajectory, and an elliptical model is generated through this fitting process. The major axis and minor axis of the ellipse respectively represent the movement amplitudes of the line in the main vibration direction and the secondary vibration direction.

[0042] After completing the ellipse fitting, calculate the tilt angle. Adopt the geometric analysis method to obtain the tilt angle of the ellipse by calculating the angle between the major axis of the fitted ellipse and the horizontal line. The tilt angle indicates the directionality of the ellipse in space and reflects whether there is an obvious direction offset in the dancing trajectory of the transmission line. Through geometric calculation, assuming that the angle between the major axis and the horizontal line is 15 degrees, the obtained elliptical tilt angle is 15 degrees, which means that the vibration of the line under the action of wind has a directional offset.

[0043] Step S400: Determine whether the wind deviation angle, dancing amplitude, vertical dancing amplitude, horizontal dancing amplitude, and elliptical tilt angle meet the dynamic monitoring threshold. If so, perform a warning for line wind deviation dancing, where the dynamic monitoring threshold is adaptively adjusted based on the tower state.

[0044] In the embodiment of the present application, the dynamic monitoring threshold is adaptively adjusted according to the actual state of the tower. The tower state includes multiple factors, such as the structural condition of the tower, the looseness of bolts, the tilt degree of the tower, and the surrounding environmental conditions.

[0045] After obtaining the wind-sag galloping data, first compare the wind-sag angle with the dynamic threshold. The wind-sag angle is the deviation angle of the line relative to the reference state. Assuming the threshold is 15 degrees, if the wind-sag angle exceeds 15 degrees, it indicates that the deviation of the line has exceeded the safety limit. Then, compare the vertical galloping amplitude and the horizontal galloping amplitude with the corresponding dynamic monitoring thresholds respectively. For example, assuming the threshold of the vertical galloping amplitude is 2 meters and the threshold of the horizontal galloping amplitude is 1.5 meters, if the actual galloping amplitude exceeds these thresholds, it means that the vibration amplitude of the line is too large and there are potential risks. At the same time, judge the elliptical tilt angle. The elliptical tilt angle represents the directional change of the line vibration trajectory. By comparing it with the threshold, if the elliptical tilt angle exceeds the set safety range, it indicates that the galloping trajectory of the line may have too large a deviation, further exacerbating the operation risk of the transmission line.

[0046] Through these comparisons and judgments, if any parameter among the wind-sag angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude or elliptical tilt angle is detected to exceed the dynamic monitoring threshold, trigger the wind-sag galloping warning of the line and notify the operation personnel to pay attention to the abnormal situation of the line.

[0047] Furthermore, as Figure 2 shown, in the method provided by the application embodiment, configuring the dynamic monitoring threshold further includes:

[0048] Obtain the threshold perturbation factors, where the threshold perturbation factors at least include the characteristics of tower bolt loosening, the characteristics of tower structure damage, the characteristics of tower tilt and the characteristics of environmental impact; configure a factor monitoring network based on the threshold perturbation factors, and within the predetermined window, use the factor monitoring network to collect real-time perturbation factors; input the real-time perturbation factors into a perturbation impact analyzer for impact analysis, and output the wind-sag angle impact coefficient, the galloping amplitude impact coefficient, the vertical galloping amplitude impact coefficient, the horizontal galloping amplitude impact coefficient and the elliptical tilt angle impact coefficient; correct the initial monitoring threshold based on the wind-sag angle impact coefficient, the galloping amplitude impact coefficient, the vertical galloping amplitude impact coefficient, the horizontal galloping amplitude impact coefficient and the elliptical tilt angle impact coefficient, and output the dynamic monitoring threshold.

[0049] In the embodiments of the present application, first, preset threshold disturbance factors are read, including tower bolt loosening characteristics, tower structure damage characteristics, tower inclination characteristics, and environmental impact characteristics. Next, through sensors and environmental monitoring devices installed on the tower, the tower state and the surrounding environment are monitored in real time to obtain real-time disturbance factors. The sensors installed on the tower include strain gauges, inclinometers, and meteorological sensors. Among them, the strain gauges are used to detect bolt loosening and tower structure damage, the inclinometers are used to measure the tower inclination angle, and the meteorological sensors are used to collect environmental data such as wind speed and temperature. The sensors monitor the tower bolt loosening characteristics, structure damage characteristics, tower inclination characteristics, and environmental impact characteristics by continuously collecting data and store these data for subsequent analysis. For example, bolt loosening is detected by the abnormal change of the tensile stress detected by the strain gauge, and the tower inclination is measured by the specific change of the inclination angle measured by the inclinometer.

[0050] Next, through the Internet of Things technology, the sensors distributed on the tower are connected to a centralized monitoring platform to form a factor monitoring network. This monitoring network uses a wireless communication protocol to transmit the sensor data to the central control system in real time. The sensor network regularly collects and transmits the disturbance factor data within a predetermined window, that is, a set time period or a specific area, to obtain real-time disturbance factors.

[0051] After obtaining the real-time disturbance factors, they are input into a pre-constructed disturbance impact analyzer for analysis to obtain the wind deflection angle impact coefficient, the galloping amplitude impact coefficient, the vertical galloping amplitude impact coefficient, the horizontal galloping amplitude impact coefficient, and the elliptical tilt angle impact coefficient. Based on the obtained wind deflection angle impact coefficient, the galloping amplitude impact coefficient, the vertical galloping amplitude impact coefficient, the horizontal galloping amplitude impact coefficient, and the elliptical tilt angle impact coefficient, the preset initial monitoring thresholds are multiplied for correction. The initial monitoring thresholds are set according to the standard tower state and typical environmental conditions and are used for monitoring under normal circumstances. However, due to the changes in the tower state and environmental conditions, these thresholds need to be adjusted according to the actual situation. Specifically, the initial threshold of each wind galloping parameter is multiplied by its corresponding impact coefficient. For example, assume that the initial threshold of the wind deflection angle is 15 degrees, and through analysis, the impact coefficient of the wind deflection angle is 1.2. Then the corrected wind deflection angle monitoring threshold will be 15×1.2 = 18 degrees. Similarly, assume that the initial threshold of the galloping amplitude is 2 meters, and the impact coefficient of the galloping amplitude is 0.9. Then the corrected galloping amplitude monitoring threshold will be 2×0.9 = 1.8 meters. In this way, it is ensured that the monitoring thresholds can be adaptively adjusted according to the real-time tower and environmental states.

[0052] After completing the correction of all monitoring thresholds, the final dynamic monitoring thresholds are output. Among them, the dynamic monitoring thresholds are updated in real time.

[0053] Further, in the method provided by the application embodiment, constructing a perturbation impact analyzer further includes:

[0054] Collecting a plurality of sample perturbation factor sets and a plurality of sample perturbation coefficient sets based on the threshold perturbation factors; configuring a plurality of perturbation impact analysis branches, using the sample perturbation factor sets as inputs and the sample perturbation coefficient sets as supervision, and respectively supervising and training the plurality of perturbation impact analysis branches by using the plurality of sample perturbation factor sets and the plurality of sample perturbation coefficient sets to obtain a plurality of perturbation impact analysis branches that meet convergence; and parallelly constructing the perturbation impact analyzer based on the plurality of perturbation impact analysis branches.

[0055] In the embodiment of the present application, first, a plurality of sample perturbation factor sets and a plurality of sample perturbation coefficient sets are obtained from the historical database based on the threshold perturbation factors. The sample perturbation factor sets are collected and stored by sensors over a past period of time, reflecting the tower bolt loosening, structural damage, inclination angle, and environmental conditions of the transmission line under various operating states. The sample perturbation coefficient sets are obtained by technical experts based on experiments or historical data analysis. The data in the sample perturbation factor sets and the sample perturbation coefficient sets correspond one by one, and each perturbation factor corresponds to an impact coefficient.

[0056] Next, a plurality of perturbation impact analysis branches are configured. Each analysis branch is an independent analysis model for processing different perturbation factors. These branches use the sample perturbation factor sets as inputs and the sample perturbation coefficient sets as supervision data. Each perturbation impact analysis branch is trained by a supervised learning method, and the goal is to let the model learn how to predict the impact on the line state based on the input perturbation factors. In each perturbation impact analysis branch, the sample perturbation factor set is input into the model, combined with the corresponding perturbation coefficient, and the parameters of the model are continuously iteratively optimized. The goal during the training process is to make the output of each model gradually approach the true perturbation coefficient until convergence is achieved, that is, the error of the model drops to an acceptable range. Once all the perturbation impact analysis branches reach the convergence state, a parallel perturbation impact analyzer is constructed based on these converged analysis branches.

[0057] Further, the method provided by the application embodiment further includes:

[0058] Randomly select a first analysis index from the wind deflection angle, the amplitude of galloping, the vertical amplitude of galloping, the horizontal amplitude of galloping, and the elliptical tilt angle, and construct N first analysis units according to the first analysis index; extract the multiple sets of sample perturbation coefficients based on the first analysis index to obtain multiple first sample perturbation coefficients, and construct a first training set in combination with the multiple sets of sample perturbation factors; divide the first training set into N equal parts, and train and verify the N first analysis units respectively to obtain N converged first analysis units, and construct a first perturbation influence analysis branch based on the N converged first analysis units, and add it to the multiple perturbation influence analysis branches, where the output of the first perturbation influence analysis branch is the mean of the output results of the N converged first analysis units.

[0059] In the embodiment of the present application, first, randomly select one of the five parameters of the wind deflection angle, the amplitude of galloping, the vertical amplitude of galloping, the horizontal amplitude of galloping, and the elliptical tilt angle as the first analysis index. For example, randomly select the "amplitude of galloping" as the first analysis index. Next, based on the selected first analysis index, extract the set of sample perturbation coefficients related to the selected first analysis index from the historical database. For example, if the "amplitude of galloping" is the first analysis index, then extract all the sample perturbation coefficients related to the amplitude of galloping, and call the extracted sample perturbation coefficients the first sample perturbation coefficients.

[0060] At the same time, combine the first sample perturbation coefficients extracted from the historical database with the set of sample perturbation factors to construct a first training set. This training set is the data set used for model training. Each sample in the training set includes a set of perturbation factors and the corresponding influence coefficient of the amplitude of galloping. To improve the robustness of training, divide the first training set into N equal parts. For example, divide the data set into 10 parts. Then create N first analysis units. Each analysis unit will independently receive a portion of the training data and perform independent model training by using a supervised learning algorithm such as a support vector machine. Each analysis unit is a small learning model that learns the relationship between the perturbation factors in the first training set and the influence coefficient of the amplitude of galloping.

[0061] During the training process, perform independent training and verification on each analysis unit, and continuously monitor the change of the model error through a convergence determination algorithm such as the mean square error. When the error of the model no longer decreases significantly and the output result tends to be stable, it is considered to have converged. When all N first analysis units reach the preset convergence standard, it is considered that the training of these models is completed.

[0062] Next, based on these N convergent first analysis units, a first perturbation impact analysis branch is constructed. This branch generates the final output by synthesizing the prediction results of each analysis unit. Specifically, the prediction results of each convergent analysis unit are averaged. That is, after averaging the output results of all analysis units, a comprehensive galloping amplitude impact coefficient is generated. In this way, the output of the first perturbation impact analysis branch is the impact coefficient after averaging the results of multiple analysis units. Finally, the constructed first perturbation impact analysis branch is added to multiple perturbation impact analysis branches.

[0063] After that, the above process is repeated, and the same analysis is performed on the remaining four wind-induced galloping parameters in turn. Specifically, each time a new analysis index is selected, such as "wind deflection angle", and relevant sample perturbation coefficients and perturbation factors are extracted based on the wind deflection angle to construct a second training set. At this time, after the second training set is segmented, it is also used to train N second analysis units. These analysis units are trained through a supervised learning algorithm to learn the perturbation impact of the wind deflection angle until convergence. Finally, the output results of these convergent second analysis units are averaged to obtain an impact coefficient of the wind deflection angle, and a second perturbation impact analysis branch is constructed. Similarly, the same steps are performed on the vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle respectively. After each analysis index is selected, relevant sample data is extracted and multi-unit training and verification are carried out, and finally a perturbation impact analysis branch for each parameter is generated.

[0064] In the embodiment of the present application, in summary, the embodiment of the present application has at least the following technical effects:

[0065] In the present application, a cross-line laser is used to project a laser on a transmission line, and a monocular camera is used to continuously collect the line laser contour within a predetermined window to obtain a contour image set; the contour image set is subjected to binary processing and contour feature extraction, and the line contour features are fused; based on the line contour features, wind-induced galloping analysis is performed, and wind-induced galloping data is output, where the wind-induced galloping data includes wind deflection angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle; it is judged whether the wind deflection angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle meet the dynamic monitoring threshold. If so, a warning of line wind-induced galloping is issued, where the dynamic monitoring threshold is adaptively adjusted based on the tower state. The present invention solves the technical problem of insufficient monitoring accuracy of wind-induced galloping of transmission lines in the prior art. By projecting a cross-line laser and using a monocular camera to collect line contour images in real time, binary processing and feature extraction are performed, and the wind-induced galloping data of the line is analyzed to judge whether parameters such as the wind deflection angle and galloping amplitude exceed the dynamic monitoring threshold, achieving the technical effect of improving the monitoring accuracy of wind-induced galloping of transmission lines.

[0066] It should be noted that the above order of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above description of specific embodiments of this specification has been made. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0067] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0068] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A monitoring method for wind-induced galloping of transmission lines by integrating crosshair laser and monocular vision, characterized in that, Including: Projecting laser on the transmission line through a crosshair laser, and continuously collecting the line laser profile within a predetermined window by using a monocular camera to obtain a set of contour images; Performing binarization processing and contour feature extraction on the set of contour images, and fusing to obtain line contour features; Performing wind deviation and galloping analysis based on the line contour features, and outputting wind deviation and galloping data, where the wind deviation and galloping data include wind deviation angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle; Judging whether the wind deviation angle, galloping amplitude, vertical galloping amplitude, horizontal galloping amplitude, and elliptical tilt angle meet the dynamic monitoring threshold. If so, carry out line wind deviation and galloping early warning, where the dynamic monitoring threshold is adaptively adjusted based on the tower state; Performing wind deviation and galloping analysis based on the line contour features, and outputting wind deviation and galloping data, including: Constructing a line twin space based on standard tower structure data and benchmark line state simulation; Importing the line contour features into the line twin space for wind deviation and galloping analysis, and outputting wind deviation and galloping data; Configuring the dynamic monitoring threshold, including: Obtaining threshold perturbation factors, where the threshold perturbation factors at least include tower bolt loosening characteristics, tower structure damage characteristics, tower tilt characteristics, and environmental impact characteristics; Configuring a factor monitoring network based on the threshold perturbation factors, and collecting real-time perturbation factors within the predetermined window by using the factor monitoring network; Inputting the real-time perturbation factors into a perturbation impact analyzer for impact analysis, and outputting a wind deviation angle impact coefficient, a galloping amplitude impact coefficient, a vertical galloping amplitude impact coefficient, a horizontal galloping amplitude impact coefficient, and an elliptical tilt angle impact coefficient; Correcting the initial monitoring threshold based on the wind deviation angle impact coefficient, the galloping amplitude impact coefficient, the vertical galloping amplitude impact coefficient, the horizontal galloping amplitude impact coefficient, and the elliptical tilt angle impact coefficient, and outputting the dynamic monitoring threshold.

2. The method for monitoring the wind-induced galloping of a transmission line by fusing a crosshair laser and monocular vision according to claim 1, wherein, Performing binarization processing and contour feature extraction on the set of contour images, and fusing to obtain line contour features, including: Performing graying processing on the set of contour images to obtain a set of gray contour images; Configuring a gray judgment scalar, and performing binary conversion on the set of gray contour images based on the gray judgment scalar to obtain a set of binary images; Successively performing edge feature extraction on the set of binary images to obtain a plurality of contour line segments, and fitting the plurality of contour line segments in sequence to output the line contour features.

3. The method for monitoring the wind-induced swing and galloping of a transmission line by fusing a crosshair laser and monocular vision according to claim 1, wherein Outputting wind deviation and galloping data, including: Within the line twin space, performing offset analysis on the line contour features and the benchmark line state, and obtaining the wind deviation angle through geometric calculation; Performing vertical galloping and horizontal galloping analysis based on the line contour features and the benchmark line state respectively, determining the vertical galloping amplitude and the horizontal galloping amplitude, and obtaining the galloping amplitude by combining the vertical galloping amplitude and the horizontal galloping amplitude; Invoking the historical line contour feature sequence within a preset period, and calculating the elliptical tilt angle based on the line contour features and the historical line contour feature sequence; Composing the wind deviation and galloping data according to the wind deviation angle, the galloping amplitude, the vertical galloping amplitude, the horizontal galloping amplitude, and the elliptical tilt angle.

4. The power transmission line wind-induced deviation and galloping monitoring method integrating crosshair laser and monocular vision according to claim 3, characterized in that, Calculating the elliptical tilt angle based on the line profile feature and the historical line profile feature sequence, including: Performing dancing trajectory ellipse fitting based on the line profile feature and the historical line profile feature sequence, and outputting a fitted ellipse trajectory; Calculating the tilt angle based on the fitted ellipse trajectory to obtain the elliptical tilt angle, where the elliptical tilt angle is the tilt angle of the major axis of the ellipse relative to the horizontal line.

5. The method for monitoring the wind-induced galloping of a transmission line by fusing a crosshair laser and monocular vision according to claim 1, characterized in that, Constructing a perturbation impact analyzer, including: Collecting a plurality of sample perturbation factor sets and a plurality of sample perturbation coefficient sets based on the threshold perturbation factors; Configuring a plurality of perturbation impact analysis branches, using the sample perturbation factor sets as inputs and the sample perturbation coefficient sets as supervision, and respectively supervising and training the plurality of perturbation impact analysis branches with the plurality of sample perturbation factor sets and the plurality of sample perturbation coefficient sets to obtain a plurality of perturbation impact analysis branches that meet convergence; Parallelly building the perturbation impact analyzer based on the plurality of perturbation impact analysis branches.

6. The method for monitoring the wind-induced deviation and galloping of a transmission line by fusing crosshair laser and monocular vision according to claim 5, wherein, The method further includes: Randomly selecting a first analysis index from the wind deflection angle, dancing amplitude, vertical dancing amplitude, horizontal dancing amplitude, and elliptical tilt angle, and constructing N first analysis units according to the first analysis index; Extracting the plurality of sample perturbation coefficient sets based on the first analysis index to obtain a plurality of first sample perturbation coefficients, and constructing a first training set in combination with the plurality of sample perturbation factor sets; Dividing the first training set into N equal parts, respectively training and validating the N first analysis units to obtain N converged first analysis units, constructing a first perturbation impact analysis branch based on the N converged first analysis units, and adding it to the plurality of perturbation impact analysis branches, where the output of the first perturbation impact analysis branch is the mean of the output results of the N converged first analysis units.

Images