A method and device for determining the danger of traction wire rope in overhead lines

Abstract

The present invention discloses a method for determining the danger of a traction wire rope in an overhead line. The method comprises: determining multiple indicators that affect the danger level of the traction wire rope; collecting sample indicator parameters related to the indicators and establishing a first sample set; constructing a danger zone identification model and training the danger zone identification model based on the first sample set; collecting sample images around the traction wire rope and establishing a second sample set; constructing a dangerous behavior analysis model and training the dangerous behavior analysis model based on the second sample set; collecting real-time indicator parameters related to the indicators and inputting them into the dangerous zone identification model to determine the range of the dangerous zone; based on the range of the dangerous zone, controlling an image acquisition device to steer within the dangerous zone to capture dangerous zone images; and inputting the dangerous zone images into the dangerous behavior analysis model to obtain dangerous behavior identification results. The present invention can improve the efficiency of determining safe zones and behaviors during the wire rope steering process.

Description

Technical Field

[0001] The present invention relates to the technical field of overhead lines, and in particular to a method and device for determining the danger of a traction wire rope in an overhead line. Background Art

[0002] With the acceleration of economic integration and the continuous expansion of modern urban construction, transmission lines have a vital economic position in the development of modern power construction. High-voltage transmission lines are directly critical to the reliable operation of the entire power grid. As an important form of sustainable energy, the installation of transmission and distribution lines has a significant impact on the stability and safety of power supply. How to ensure the quality of line construction through scientific line construction management is the primary issue facing modern power companies in power grid construction. However, the commonly used line construction management methods still have certain drawbacks, and there is still room for improvement in line construction management.

[0003] In prior art, during line stringing construction, patent application CN108599007A discloses a device and method for installing pulley insulation ropes on high-voltage transmission lines using a drone. The method involves a drone-assisted working robot system carrying a hook pulley device, which is then brought online. The working portion of the working robot system then uses the hook pulley to bring the pulley online, and then installs the pulley insulation rope. The device includes a working robot portion and a hook pulley portion. While this allows for the installation of pulley insulation ropes on high-voltage transmission lines under manual control, it eliminates the need for dedicated personnel to risk safety hazards during the line-up and downline installation process when using maintenance robots. It also eliminates the need for additional structures to assist with line-up and downline installation, and replaces the process of using a rope thrower to throw the insulation rope around the line to pull an insulation ladder online when workers need to climb a tower for work. However, when the wire rope of the traction conductor needs to turn, there is a certain range of dangerous areas around the stressed wire rope, and the method of checking whether there are people staying or passing through the dangerous area still requires manual observation and processing, which leads to technical problems of untimely processing and low efficiency.

[0004] Therefore, how to design a method for quickly and accurately identifying whether there is danger in the wire rope is a technical problem that needs to be solved urgently in this field. Summary of the Invention

[0005] This invention provides a method and device for determining danger in traction wire ropes on overhead lines. These devices can identify dangerous areas based on real-time acquisition of indicator parameters and issue early warnings based on the results of dangerous behavior identification. This intelligently processes, analyzes, and identifies dangerous area images, assisting personnel with observation and processing, thereby improving the efficiency of determining safe areas and behaviors during wire rope steering.

[0006] In a first aspect, the present invention provides a method for determining the danger of a traction wire rope in an overhead line, wherein the traction wire rope is diverted and pulled by a diverting pulley, and an image acquisition device is provided on the overhead line. The method comprises:

[0007] Determine multiple indicators that affect the danger level of traction wire ropes;

[0008] Collecting sample indicator parameters related to the indicator and establishing a first sample set;

[0009] Constructing a dangerous area recognition model, and training the dangerous area recognition model based on the first sample set;

[0010] Collecting sample images around the traction wire rope and establishing a second sample set;

[0011] Constructing a dangerous behavior analysis model, and training the dangerous behavior analysis model based on the second sample set;

[0012] Collecting real-time indicator parameters of the indicators and inputting them into the dangerous area identification model to determine the scope of the dangerous area;

[0013] According to the range of the dangerous area, controlling the image acquisition device to turn to the range of the dangerous area to acquire the dangerous area image;

[0014] The dangerous area image is input into the dangerous behavior analysis model to obtain a dangerous behavior recognition result.

[0015] Furthermore, the dangerous area identification model is constructed based on BP neural network.

[0016] Furthermore, the dangerous behavior analysis model is constructed based on a convolutional neural network.

[0017] Furthermore, the indicators include a first force indicator of the traction wire rope, a second force indicator of the steering pulley, and a usage time indicator of the traction wire rope.

[0018] Furthermore, collecting sample indicator parameters related to the indicator and establishing a first sample set includes:

[0019] Performing a stress analysis on multiple sets of traction steel ropes to obtain first traction forces of the multiple sets of traction steel ropes and angles between the directions of the first traction forces and the laying directions of the traction steel ropes. The first traction forces are graded according to their magnitudes, and the angles are classified. Each level of first traction force in each type of angle corresponds to a dangerous area range, which serves as a sample indicator parameter of the first stress indicator.

[0020] Performing force analysis on multiple sets of steering pulleys to obtain the magnitude and direction of the second traction force exerted on the multiple sets of steering pulleys. Classifying the pulleys according to the direction of the second traction force, with each level of traction force in each category corresponding to a dangerous area range, serves as a sample indicator parameter of the second force indicator;

[0021] The wear degree of the traction wire rope is graded according to the usage time, and the sample index parameters corresponding to the usage time index are obtained;

[0022] Integrating the sample index parameters of the first force index, the sample index parameters of the second force index, and the sample index parameters corresponding to the usage time index to obtain multiple groups of sample index parameters;

[0023] Dangerous area ranges are marked for the multiple groups of sample indicator parameters to obtain the first sample set.

[0024] Furthermore, the scope of the danger zone marking includes the area above, below, around and inside corners of the traction wire rope;

[0025] Identifying the range of dangerous areas for the multiple groups of sample indicator parameters includes:

[0026] According to the level corresponding to each group of data in the multiple groups of sample indicator parameters and the distance between the area and the traction wire rope, the upper area, lower area, surrounding area and inner corner area are graded and the size of the dangerous area corresponding to each level is marked.

[0027] Furthermore, the sample images around the traction wire rope include different dangerous behaviors;

[0028] Collect sample images around the traction wire rope and create a second sample set, including:

[0029] The dangerous behaviors in the sample images are marked and graded to a dangerous level to obtain the second sample set.

[0030] Furthermore, the method further comprises:

[0031] The first sample set is divided into a first training set, a first validation set and a first test set. The dangerous area recognition model is trained using the first training set, and the trained dangerous area recognition model is verified and tested according to the first validation set and the first test set, respectively, until the accuracy of the dangerous area recognition model meets the preset requirements.

[0032] Furthermore, the method further comprises:

[0033] The second sample set is divided into a second training set, a second validation set and a second test set. The dangerous behavior analysis model is trained using the second training set, and the trained dangerous behavior analysis model is verified and tested according to the second validation set and the second test set, respectively, until the accuracy of the dangerous behavior analysis model meets the preset requirements.

[0034] In a second aspect, the present invention further provides a device for determining the danger of a traction wire rope in an overhead line applied to the above method, comprising:

[0035] An index determination module, used to determine multiple indicators that affect the danger level of the traction wire rope;

[0036] A first collection module, configured to collect sample indicator parameters related to the indicator and establish a first sample set;

[0037] a first training module, configured to construct a dangerous area recognition model and train the dangerous area recognition model according to the first sample set;

[0038] A second acquisition module is used to acquire sample images around the traction wire rope and establish a second sample set;

[0039] a second training module, configured to construct a dangerous behavior analysis model and train the dangerous behavior analysis model based on the second sample set;

[0040] A third acquisition module is used to collect real-time indicator parameters of the indicator and input them into the dangerous area identification model to determine the scope of the dangerous area;

[0041] An image acquisition control module is used to control the image acquisition device to turn to the dangerous area to acquire dangerous area images according to the dangerous area range;

[0042] The recognition module is used to input the dangerous area image into the dangerous behavior analysis model to obtain a dangerous behavior recognition result.

[0043] Furthermore, the first training module constructs a dangerous area identification model based on a BP neural network.

[0044] Furthermore, the first training module constructs the dangerous behavior analysis model based on a convolutional neural network.

[0045] Furthermore, the indicators include a first force indicator of the traction wire rope, a second force indicator of the steering pulley, and a usage time indicator of the traction wire rope.

[0046] Furthermore, the first acquisition module is further configured to:

[0047] Performing a stress analysis on multiple sets of traction steel ropes to obtain first traction forces of the multiple sets of traction steel ropes and angles between the directions of the first traction forces and the laying directions of the traction steel ropes. The first traction forces are graded according to their magnitudes, and the angles are classified. Each level of first traction force in each type of angle corresponds to a dangerous area range, which serves as a sample indicator parameter of the first stress indicator.

[0048] Performing force analysis on multiple sets of steering pulleys to obtain the magnitude and direction of the second traction force exerted on the multiple sets of steering pulleys. Classifying the pulleys according to the direction of the second traction force, with each level of traction force in each category corresponding to a dangerous area range, serves as a sample indicator parameter of the second force indicator;

[0049] The wear degree of the traction wire rope is graded according to the usage time, and the sample index parameters corresponding to the usage time index are obtained;

[0050] Integrating the sample index parameters of the first force index, the sample index parameters of the second force index, and the sample index parameters corresponding to the usage time index to obtain multiple groups of sample index parameters;

[0051] Dangerous area ranges are marked for the multiple groups of sample indicator parameters to obtain the first sample set.

[0052] Furthermore, the marking range of the danger zone includes the area above, below, around and inside corners of the traction wire rope;

[0053] The first acquisition module is further configured to:

[0054] According to the level corresponding to each group of data in the multiple groups of sample indicator parameters and the distance between the area and the traction wire rope, the upper area, lower area, surrounding area and inner corner area are graded and the size of the dangerous area corresponding to each level is marked.

[0055] Furthermore, the sample images around the traction wire rope include different dangerous behaviors;

[0056] The second acquisition module is further used for:

[0057] The dangerous behaviors in the sample images are marked and graded to a dangerous level to obtain the second sample set.

[0058] Furthermore, the first training module is further configured to:

[0059] The first sample set is divided into a first training set, a first validation set and a first test set. The dangerous area recognition model is trained using the first training set, and the trained dangerous area recognition model is verified and tested according to the first validation set and the first test set, respectively, until the accuracy of the dangerous area recognition model meets the preset requirements.

[0060] Furthermore, the second training module is further used to:

[0061] The second sample set is divided into a second training set, a second validation set and a second test set. The dangerous behavior analysis model is trained using the second training set, and the trained dangerous behavior analysis model is verified and tested according to the second validation set and the second test set, respectively, until the accuracy of the dangerous behavior analysis model meets the preset requirements.

[0062] The present invention provides a method and device for determining the danger of a traction wire rope in an overhead line, which has at least the following beneficial effects:

[0063] (1) By marking the dangerous areas of the indicator parameters collected in real time, controlling the movement of the image acquisition device and issuing early warnings based on the dangerous behavior identification results, the dangerous area images can be intelligently processed, analyzed and marked, thereby assisting the staff in observation and processing, thereby improving the work efficiency of the safe area and behavior judgment during the wire rope turning process.

[0064] (2) By grading multiple indicators that affect the degree of danger of traction wire ropes and grading dangerous behaviors into dangerous levels, the accuracy of determining the dangerous area of traction wire ropes and the effect of accurately determining dangerous behaviors can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A flow chart of a method for determining danger of traction wire ropes in overhead lines provided by the present invention;

[0066] Figure 2 A flowchart of establishing a first sample set according to an embodiment of the present invention;

[0067] Figure 3 A schematic diagram of a device for determining danger of a traction wire rope used in an overhead line provided by the present invention;

[0068] Figure 4 This is a schematic diagram of a first acquisition module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0069] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0070] See also Figure 1 As shown, the present invention provides a method for determining the danger of a traction wire rope in an overhead line. The traction wire rope is diverted and pulled by a diverting pulley. An image acquisition device is provided on the overhead line. The method includes:

[0071] Step S100: determining multiple indicators that affect the danger level of the traction wire rope;

[0072] Step S200: Collect sample indicator parameters related to the indicator and establish a first sample set;

[0073] Step S300: constructing a dangerous area recognition model and training the dangerous area recognition model based on the first sample set;

[0074] Step S400: collecting sample images around the traction wire rope and establishing a second sample set;

[0075] Step S500: constructing a dangerous behavior analysis model and training the dangerous behavior analysis model based on the second sample set;

[0076] Step S600: collecting real-time indicator parameters of the indicators and inputting them into the dangerous area identification model to determine the scope of the dangerous area;

[0077] Step S700: Control the image acquisition device to capture images of the dangerous area within the dangerous area according to the dangerous area range;

[0078] Step S800: Input the dangerous area image into the dangerous behavior analysis model to obtain the dangerous behavior recognition result.

[0079] The present invention uses a dangerous area identification model to identify the dangerous area range of indicator parameters collected in real time, controls the movement of an image acquisition device, and issues warnings based on the dangerous behavior identification results obtained by the dangerous behavior analysis model. This achieves intelligent processing, analysis, and identification of dangerous area images, thereby assisting personnel in observation and processing, thereby improving the efficiency of safe area and behavior determination during the wire rope steering process. The traction wire rope mentioned in the present invention is a rope composed of multiple fine steel wires twisted into strands and wound into a spiral around a rope core. It is steered and pulled by a pre-installed diverting pulley and is commonly used on construction sites for lifting, traction, and tensioning of lifting equipment. The diverting pulley is a circular wheel that can rotate around a central axis and has a groove on its circumference. When a wire rope is wrapped around the groove and either end of the wire rope is pulled, the friction between the wire rope and the circular wheel causes the circular wheel to rotate around the central axis. The diverting pulley's main function is to change direction.

[0080] In step S100, a usage time index is obtained based on the usage time of the wire rope, a force analysis of the wire rope during traction is performed to obtain a first force index, and a force analysis of the steering pulley during traction is performed to obtain a second force index. The usage time index of the traction wire rope, the first force index of the traction wire rope, and the second force index of the steering pulley are used as multiple indicators to be determined. By analyzing the force of the wire rope during traction, a preliminary understanding of the wire rope is achieved, laying the foundation for subsequent hazard identification based on the usage of the wire rope.

[0081] See also Figure 2 As shown, in step S200, sample indicator parameters of the indicator are collected and a first sample set is established, including:

[0082] Step S210: Performing force analysis on multiple groups of traction steel ropes to obtain first traction forces of the multiple groups of traction steel ropes and angles between the directions of the first traction forces and the laying directions of the traction steel ropes. The first traction forces are graded according to their magnitudes, and the angles are classified. Each level of first traction force in each type of angle corresponds to a dangerous area range, which serves as a sample indicator parameter of the first force indicator.

[0083] Step S220: Analyze the forces acting on the multiple steering pulleys to obtain the magnitudes and directions of the second traction forces acting on the multiple steering pulleys. Classify the pulleys according to the directions of the second traction forces. Each level of traction force in each direction corresponds to a dangerous area, which serves as a sample indicator parameter of the second force indicator.

[0084] Step S230: Classify the wear degree of the traction wire rope according to the usage time, and obtain sample index parameters corresponding to the usage time index;

[0085] Step S240: Integrate the sample index parameters of the first force index, the sample index parameters of the second force index, and the sample index parameters corresponding to the usage time index to obtain multiple groups of sample index parameters;

[0086] Step S250: Mark the range of dangerous areas for multiple groups of sample indicator parameters to obtain a first sample set.

[0087] In step S210, a force analysis is performed on the turning point of the wire rope to obtain the first traction force and friction force received by the wire rope at the turning point. Similarly, the magnitude of the first traction force received by the wire rope is graded according to the maximum force that the wire rope can withstand, and the first traction force is divided into four levels. It can be seen from the force analysis of the wire rope that the larger the angle between the first traction force and the direction of the traction force before turning (the first traction force direction and the vertical direction of the traction wire rope), the smaller the wear on the turning point of the wire rope. Therefore, the angle between the force direction of the wire rope at the turning point and the direction of the traction force before turning is graded, and the angle is also divided into four levels. The magnitude and angle of the first traction force received by the wire rope at the turning point are used as the first force index. According to the first force index, the sample index parameter of the first force index is obtained.

[0088] In step S220, a force analysis is similarly performed on the target steering pulley, and the direction and magnitude of the traction force acting on the target steering pulley are graded, which serves as a second force index. Sample index parameters for the second force index are obtained based on the second force index. By analyzing the force acting on the wire rope during traction, a preliminary understanding of the wire rope is achieved, laying the foundation for subsequent hazard identification based on the wire rope's usage.

[0089] In step S230, the service life of the wire rope is obtained based on its service life and usage history, and the service life is graded according to the degree of wear. For example, a wire rope with a service life of five years is graded as Grade 1 due to good condition and low wear in the first three years. Grade 2 due to poor condition and some wear in the fourth year. Grade 3 due to severe wear and poor condition in the first half of the fifth year. Grade 4 due to the second half of the fifth year, which is nearing the point of being discarded and requiring replacement. This serves as the service life indicator. Sample indicator parameters for the service life indicator are obtained based on this service life indicator.

[0090] In step S250, the marking range of the dangerous area includes the upper area, the lower area, the surrounding area and the inner corner area of the traction wire rope.

[0091] Mark the range of dangerous areas for multiple groups of sample indicator parameters, including:

[0092] Based on the levels corresponding to each set of data in multiple groups of sample indicator parameters and the distance between the area and the traction wire rope, the upper area, lower area, surrounding area, and inner corner area are graded and the size of the corresponding dangerous area for each level is marked.

[0093] Specifically, the dangerous area is graded according to the maximum distance between the edge of the dangerous area and the wire rope, and is divided into four levels from near to far. The lower the level, the smaller the dangerous area, and the safer the corresponding wire rope. The obtained dangerous area range identification is matched with the corresponding wire rope, and the dangerous area is marked on the four sides, top, bottom and inner corners of the wire rope according to the dangerous area range identification. For example, according to the above classification results, for a wire rope with a usage time, force magnitude, and angle of level one, the wire rope is in a relatively new and less worn state. It is concluded that the dangerous area of the wire rope is small in all aspects. The dangerous area identification of the four sides, top, bottom and inner corners of the wire rope is all level one. In this way, the surrounding area, top area, bottom area and inner corner area of the wire rope are obtained. These dangerous areas are merged to obtain the dangerous area of the wire rope. This solves the problem of identifying dangerous areas and realizes the division of dangerous areas around the wire rope.

[0094] The hazard zone identification process is based on the assessment of multiple sample indicators by technical experts in the field of overhead line construction based on past accident experience. The assessment results are then integrated to determine the hazard zone size corresponding to each level. This allows for accurate identification of hazard zones based on historical incidents.

[0095] Step S300 may include:

[0096] The first sample set is divided into a first training set, a first validation set and a first test set. The hazardous area recognition model is trained using the first training set, and the trained hazardous area recognition model is verified and tested according to the first validation set and the first test set, respectively, until the accuracy of the hazardous area recognition model meets the preset requirements.

[0097] Specifically, the dangerous area identification model is constructed based on the BP neural network (back propagation, BP). The construction of the dangerous area identification model includes: first, using the first training set, the first validation set and the first test set as input training samples, relevant technical experts set multiple first mapping features according to business needs, and use the multiple first mapping features as prediction training samples, and normalize the training sample data and the prediction training samples. Then, the BP neural network is constructed, the number of training times, the learning rate and the minimum error of the training target are set, and the first training set, the first validation set and the first test set are used to train the model. Finally, the prediction results of the BP neural network are denormalized and the error is calculated. When the error is less than the first preset value, the construction of the dangerous area identification model is completed.

[0098] The hazard zone identification model includes an input layer, an implicit processing layer, and an output layer. After inputting the usage time, force magnitude, and angle of multiple indicator parameters through the input layer, the corresponding hazard zone range identification is output after processing by the processing layer. Based on the first mapping feature, the wear condition of the wire rope corresponding to the multiple indicator parameters or the corresponding hazard zone range identification can be obtained, and then the multiple indicator parameters can be sharded and distributed based on them. By training and constructing a hazard zone identification model, it is possible to obtain accurate corresponding hazard zone range identification based on multiple indicator parameters, and then identify and classify multiple indicator parameters to achieve the technical effect of accurately calculating output data.

[0099] In step S400, the sample images around the traction wire rope include different dangerous behaviors;

[0100] Collect sample images around the traction wire rope and create a second sample set, including:

[0101] The dangerous behaviors in the sample images are marked and graded to a dangerous level to obtain a second sample set.

[0102] That is, image information of dangerous areas of multiple wire ropes under different conditions is collected to obtain multiple sample images, and dangerous behaviors in the multiple sample images are manually analyzed and marked to obtain multiple sample dangerous behavior recognition results to obtain a second sample set.

[0103] Specifically, multiple sample areas are set up, and different dangerous behaviors are set in each sample area, such as people being in a dangerous area, approaching a dangerous area, or working in a dangerous area. Sample images of dangerous behaviors are collected from these sample areas and integrated to form multiple sample images. During the acquisition process, multiple cameras are set up in different positions. Each camera adjusts its angle to obtain image information of multiple dangerous areas of the wire rope within the sample area, including the area above the wire rope, the area below the wire rope, the surrounding area, and the inner corner area. The images of these dangerous areas are then integrated to form a sample image.

[0104] A number of technical experts in the field of overhead line construction judge and analyze dangerous behaviors in multiple sample images based on past accident experience, set danger levels, mark them according to the danger levels, obtain dangerous behavior recognition results for multiple samples, and obtain a second sample set.

[0105] Step S500 may include:

[0106] The second sample set is divided into a second training set, a second validation set and a second test set. The dangerous behavior analysis model is trained using the second training set, and the trained dangerous behavior analysis model is verified and tested according to the second validation set and the second test set, respectively, until the accuracy of the dangerous behavior analysis model meets the preset requirements.

[0107] The division of the second sample set is specifically to perform data identification and division on multiple sample images and multiple sample dangerous behavior recognition results in the second sample set to obtain a second training set, a second validation set and a second test set, and to construct a dangerous behavior analysis model based on a convolutional neural network.

[0108] The dangerous behavior analysis model is built based on a convolutional neural network. This model collects sample images and matches the image feature capture results. The model is trained with a large amount of training data, giving it greater "experience" in image processing. A first convolutional feature is obtained, which is a convolutional module for feature comparison. Multiple sample images are obtained, and technical experts in the field manually analyze and label the dangerous behaviors within the multiple sample images. Feature extraction is performed on different dangerous behavior images. Based on the feature extraction results, a first feature is obtained, which is used as the first convolutional feature.

[0109] The first convolution feature is used as a prediction training sample, and multiple sample images and prediction training samples are normalized. Then, a convolutional neural network is constructed, and the number of training times, learning rate, and training target minimum error are set. The model is trained using the second training set, second validation set, and second test set. Finally, the prediction results of the convolutional neural network are denormalized and the error is calculated. When the error is less than a second preset value, the dangerous behavior analysis model is constructed. Through the construction of the dangerous behavior analysis model, intelligent processing, analysis, and identification of dangerous area images are achieved, thereby assisting personnel in image processing and observation, achieving the technical effect of improving the work efficiency of behavior determination in the safe area during the wire rope steering process.

[0110] In step S600, the real-time indicator parameters are used as input data into the danger zone identification model, and the danger zone is output. This achieves the determination of the danger zone based on historical accidents and achieves the effect of accurately judging the danger zone.

[0111] In step S800, the dangerous area image is used as input data into the dangerous behavior analysis model, and the dangerous behavior identification result is output. This enables the judgment of dangerous behaviors based on historical accidents and achieves the effect of accurately determining dangerous behaviors.

[0112] See also Figure 3 As shown, the present invention also provides a device for determining danger of traction wire rope in an overhead line applied to the above method, comprising:

[0113] An index determination module 1 is used to determine multiple indicators that affect the degree of danger of the traction wire rope;

[0114] A first collection module 2, configured to collect sample indicator parameters related to an indicator and establish a first sample set;

[0115] A first training module 3 is used to construct a dangerous area recognition model and train the dangerous area recognition model according to the first sample set;

[0116] The second acquisition module 4 is used to collect sample images around the traction wire rope and establish a second sample set;

[0117] The second training module 5 is used to build a dangerous behavior analysis model and train the dangerous behavior analysis model based on the second sample set;

[0118] The third acquisition module 6 is used to collect real-time indicator parameters of the indicators and input them into the danger zone identification model to determine the scope of the danger zone;

[0119] An image acquisition control module 7 is used to control the image acquisition device to move to the dangerous area to acquire dangerous area images according to the dangerous area range;

[0120] The recognition module 8 is used to input the dangerous area image into the dangerous behavior analysis model to obtain the dangerous behavior recognition result.

[0121] See also Figure 4 As shown, the first acquisition module 2 may include:

[0122] A first indicator acquisition module 21 is configured to perform stress analysis on multiple sets of traction steel ropes, obtain a mapping relationship between a first traction force of the multiple sets of traction steel ropes and an angle between the direction of the first traction force and the vertical direction of the traction steel ropes, and classify the angles as sample indicator parameters of the first stress indicator;

[0123] A second index acquisition module 22 is used to perform force analysis on multiple sets of steering pulleys to obtain the magnitude and direction of the second traction force of the multiple sets of steering pulleys as sample index parameters of the second force index;

[0124] The time index acquisition module 23 is used to classify the wear degree of the traction wire rope according to the use time and obtain the sample index parameter corresponding to the use time index;

[0125] An indicator integration module 24 is configured to integrate the sample indicator parameters of the first stress indicator, the sample indicator parameters of the second stress indicator, and the sample indicator parameters corresponding to the usage time indicator to obtain multiple groups of sample indicator parameters;

[0126] The area identification module 25 is used to identify the range of dangerous areas for multiple groups of sample indicator parameters to obtain a first sample set.

[0127] The method and device for determining the danger of traction wire ropes in overhead lines provided in the above embodiments have at least the following beneficial effects:

[0128] (1) By marking the dangerous areas of the indicator parameters collected in real time, controlling the movement of the image acquisition device and issuing early warnings based on the dangerous behavior identification results, the dangerous area images can be intelligently processed, analyzed and marked, thereby assisting the staff in observation and processing, thereby improving the work efficiency of the safe area and behavior judgment during the wire rope turning process.

[0129] (2) By grading multiple indicators that affect the degree of danger of traction wire ropes and grading dangerous behaviors into dangerous levels, the accuracy of determining the dangerous area of traction wire ropes and the effect of accurately determining dangerous behaviors can be improved.

[0130] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the invention. Thus, the present invention is intended to include such changes and modifications as fall within the scope of the claims and their equivalents.

Claims

1. A method for determining the danger of traction wire ropes in overhead lines, characterized in that: The traction wire rope is steered and pulled by a steering pulley, and an image acquisition device is provided on the overhead line. The method includes: Determine multiple indicators that affect the danger level of traction wire ropes; Collecting sample indicator parameters related to the indicator and establishing a first sample set; Constructing a dangerous area recognition model, and training the dangerous area recognition model based on the first sample set; Collecting sample images around the traction wire rope and establishing a second sample set; Constructing a dangerous behavior analysis model, and training the dangerous behavior analysis model based on the second sample set; Collecting real-time indicator parameters of the indicators and inputting them into the dangerous area identification model to determine the scope of the dangerous area; According to the range of the dangerous area, controlling the image acquisition device to turn to the range of the dangerous area to acquire the dangerous area image; Inputting the dangerous area image into the dangerous behavior analysis model to obtain a dangerous behavior recognition result; The indicators include a first force index of the traction wire rope, a second force index of the diverting pulley, and a usage time index of the traction wire rope; Collecting sample indicator parameters related to the indicator and establishing a first sample set includes: Performing a stress analysis on multiple sets of traction steel ropes to obtain first traction forces of the multiple sets of traction steel ropes and angles between the directions of the first traction forces and the laying directions of the traction steel ropes. The first traction forces are graded according to their magnitudes, and the angles are classified. Each level of first traction force in each type of angle corresponds to a dangerous area range, which serves as a sample indicator parameter of the first stress indicator. Performing force analysis on multiple sets of steering pulleys to obtain the magnitude and direction of the second traction force exerted on the multiple sets of steering pulleys. Classifying the pulleys according to the direction of the second traction force, with each level of traction force in each category corresponding to a dangerous area range, serves as a sample indicator parameter of the second force indicator; The wear degree of the traction wire rope is graded according to the usage time, and the sample index parameters corresponding to the usage time index are obtained; Integrating the sample index parameters of the first force index, the sample index parameters of the second force index, and the sample index parameters corresponding to the usage time index to obtain multiple groups of sample index parameters; Dangerous area ranges are marked for the multiple groups of sample indicator parameters to obtain the first sample set.

2. The method according to claim 1, characterized in that The dangerous area identification model is constructed based on BP neural network.

3. The method according to claim 1, characterized in that The dangerous behavior analysis model is constructed based on a convolutional neural network.

4. The method according to claim 1, wherein The marking range of the dangerous area includes the area above, below, around and inside corners of the traction wire rope; Identifying the range of dangerous areas for the multiple groups of sample indicator parameters includes: According to the level corresponding to each group of data in the multiple groups of sample indicator parameters and the distance between the area and the traction wire rope, the upper area, lower area, surrounding area and inner corner area are graded and the size of the dangerous area corresponding to each level is marked.

5. The method according to claim 1, wherein The sample images around the traction wire rope include different dangerous behaviors; Collect sample images around the traction wire rope and create a second sample set, including: The dangerous behaviors in the sample images are marked and graded to a dangerous level to obtain the second sample set.

6. The method according to claim 1, characterized in that The method further comprises: The first sample set is divided into a first training set, a first validation set and a first test set. The dangerous area recognition model is trained using the first training set, and the trained dangerous area recognition model is verified and tested according to the first validation set and the first test set, respectively, until the accuracy of the dangerous area recognition model meets the preset requirements.

7. The method according to claim 1, characterized in that The method further comprises: The second sample set is divided into a second training set, a second validation set and a second test set. The dangerous behavior analysis model is trained using the second training set, and the trained dangerous behavior analysis model is verified and tested according to the second validation set and the second test set, respectively, until the accuracy of the dangerous behavior analysis model meets the preset requirements.

8. A device for determining danger of traction wire rope in an overhead line applied to the method according to any one of claims 1 to 7, characterized in that: include: An index determination module, used to determine multiple indicators that affect the danger level of the traction wire rope; A first collection module, configured to collect sample indicator parameters related to the indicator and establish a first sample set; a first training module, configured to construct a dangerous area recognition model and train the dangerous area recognition model according to the first sample set; A second acquisition module is used to acquire sample images around the traction wire rope and establish a second sample set; a second training module, configured to construct a dangerous behavior analysis model and train the dangerous behavior analysis model based on the second sample set; A third acquisition module is used to collect real-time indicator parameters of the indicator and input them into the dangerous area identification model to determine the scope of the dangerous area; An image acquisition control module is used to control the image acquisition device to turn to the dangerous area to acquire dangerous area images according to the dangerous area range; an identification module, configured to input the dangerous area image into the dangerous behavior analysis model to obtain a dangerous behavior identification result; The indicators include a first force index of the traction wire rope, a second force index of the diverting pulley, and a usage time index of the traction wire rope; The first acquisition module is further configured to: Performing a stress analysis on multiple sets of traction steel ropes to obtain first traction forces of the multiple sets of traction steel ropes and angles between the directions of the first traction forces and the laying directions of the traction steel ropes. The first traction forces are graded according to their magnitudes, and the angles are classified. Each level of first traction force in each type of angle corresponds to a dangerous area range, which serves as a sample indicator parameter of the first stress indicator. Performing force analysis on multiple sets of steering pulleys to obtain the magnitude and direction of the second traction force exerted on the multiple sets of steering pulleys. Classifying the pulleys according to the direction of the second traction force, with each level of traction force in each category corresponding to a dangerous area range, serves as a sample indicator parameter of the second force indicator; The wear degree of the traction wire rope is graded according to the usage time, and the sample index parameters corresponding to the usage time index are obtained; Integrating the sample index parameters of the first force index, the sample index parameters of the second force index, and the sample index parameters corresponding to the usage time index to obtain multiple groups of sample index parameters; Dangerous area ranges are marked for the multiple groups of sample indicator parameters to obtain the first sample set.

Images