Cable distributed fault sensing method and system based on visual image

Abstract

The invention relates to the field of cable fault sensing, and discloses a cable distributed fault sensing method and system based on a visual image, and the method comprises the following steps: determining a key position of a cable line, installing an electric power data sensor at the key position of the cable line, and collecting a real-time image of the corresponding position for state analysis; and the wireless communication system for constructing nodes is used for constructing a visual image algorithm model, positioning a fault position of a target cable, introducing a neural network algorithm to carry out fault position real-time prediction, finally carrying out traceability analysis on a cable fault distribution position, and making a repair scheme for the cable fault position. According to the invention, fault position sensing of the cable is realized through visual images, real-time monitoring, accurate positioning and rapid response of cable line faults are realized, and a powerful guarantee is provided for safe and stable operation of a power grid.

Description

Technical Field

[0001] The present invention relates to the field of cable fault perception, and in particular to a method and system for distributed cable fault perception based on visualized images. Background Art

[0002] The types of distributed faults that may occur in cables mainly include short circuit faults and ground faults. A short circuit fault refers to direct contact between two or more conductors of a line, resulting in a significant increase in current, which may cause equipment damage or fire. A ground fault refers to contact between the line conductor and the ground, resulting in abnormal current flow to the ground. When a distributed fault occurs in a cable, arcs, smoke, sparks, etc. are easily generated at the cable fault site, which will cause the cable to burn, so it is necessary to locate the location of the distributed fault in the cable. Through visual images, operation and maintenance personnel can intuitively see the location and situation of the fault, and quickly understand the nature and severity of the fault. At the same time, the multi-dimensional data acquisition device can use built-in high-precision sensors to collect key parameters on the transmission line in real time, including but not limited to current, temperature, vibration and other physical quantities. At the same time, the image information of the line and the surrounding environment is captured to provide comprehensive and accurate data support for subsequent fault analysis. In summary, the use of visual images to perceive distributed faults in cables can significantly improve the efficiency and accuracy of fault identification, and provide a strong guarantee for the safe and stable operation of the power system. Therefore, a cable distributed fault perception method and system based on visual images are proposed. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a cable distributed fault perception method and system based on visualized images.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is: The first aspect of the present invention provides a cable distributed fault perception method based on a visualized image, comprising the following steps: S102: determining a fault-prone position of the cable, installing a power data sensor at the fault-prone position of the cable, collecting a real-time image of the fault-prone position of the cable, and performing image enhancement and image feature analysis on the obtained real-time image of the fault-prone position of the cable; S104: performing communication configuration and communication association on target cable nodes of different categories to build a qualified node wireless communication system; S106: Control the power communication frequency at all the first-class target cable nodes to configure sensors for real-time monitoring of power data, and build a cable visual fault location model in combination with the node surface abnormal state at the first-class target cable nodes and the qualified node wireless communication system to locate the fault position of the target cable, and introduce a neural network algorithm to predict the fault position in real time.

[0005] Furthermore, in a preferred embodiment of the present invention, the step S102 is specifically: The cable that needs to be fault-detected is marked as the target cable, and all the power equipment connected to the target cable is obtained and marked as the target power equipment; Obtaining a connection method between a target cable and a target power device, and obtaining a big data network at the same time, retrieving historical failure rates at different locations on the target cable based on the connection method between the target cable and the target power device based on the big data network, and presetting a maximum failure rate; Wherein, the position on the target cable is defined as the target cable node, and the historical failure rates at different positions on the target cable are the historical failure rates at different target cable nodes; Select a target cable node whose historical failure rate is greater than the maximum failure rate, mark it as a type of target cable node, and obtain a power data reference range of the target cable, wherein the power data reference range of the target cable includes a tolerable voltage range and a tolerable current range of the target cable; Based on the power data reference range of the target cable, determine a power data sensor whose data monitoring range is equal to the power data reference range of the target cable, calibrate it as the target power data sensor, and install the target power data sensor at all the first-class target cable nodes; A real-time image acquisition device is introduced, and all target power equipment is operated. During the operation of the target power equipment, images at a type of target cable node are acquired in real time by the real-time image acquisition device and calibrated as a type of real-time image; Image enhancement processing and image feature extraction are performed on a class of real-time images to determine image features of a class of enhanced real-time images.

[0006] Furthermore, in a preferred embodiment of the present invention, the image enhancement processing and image feature extraction are performed on a class of real-time images to determine the image features of a class of enhanced real-time images, specifically: A class of real-time images is grayed, a Gaussian filter device is introduced, a Gaussian filter kernel is determined in the Gaussian filter device, and the Gaussian filter kernel is traversed and slid in the class of real-time images after the graying process to implement Gaussian filtering of the class of real-time images after the graying process, thereby obtaining a class of filtered real-time images; The illumination component of the image is calculated in a class of filtered real-time images, and the illumination component of the image is logarithmically transformed to obtain the reflection component of the image; Based on the reflection component of the image, a histogram of a class of filtered real-time images is obtained, and the MSR algorithm is introduced to perform histogram equalization processing and single-scale image detail enhancement processing in the histogram of the class of filtered real-time images to obtain the enhanced reflection component of the image; The enhanced reflection component of the image is applied to a type of filtered real-time image to obtain an initial enhanced type of filtered real-time image, different pixel points on the initial enhanced type of filtered real-time image are obtained, and pixel values ​​are logarithmically transformed on different pixel points on the initial enhanced type of filtered real-time image to obtain a type of enhanced real-time image; The ORB algorithm is introduced to calculate the pixel feature vector gradient at a target cable node on a class of enhanced real-time images, and the standard gradient range of the feature vector is preset. In the class of enhanced real-time images, the pixel points whose feature vector gradient remains within the standard gradient range of the feature vector are selected and calibrated as qualified pixel points, and pixel reconstruction is performed on all qualified pixel points to obtain the image features of a class of enhanced real-time images. Feature analysis is performed on image features of a type of enhanced real-time images to determine abnormal node surface states at a type of target cable nodes, wherein the abnormal node surface states include surface arcs, surface sparks, and surface smoke.

[0007] Furthermore, in a preferred embodiment of the present invention, the step S104 is specifically: Installing communication modules in target power data sensors at all target cable nodes of the first type to obtain power communication sensors, presetting standard communication frequencies, and configuring the communication frequencies of all power communication sensors to be the standard communication frequencies; The power communication sensor after the communication frequency is configured is calibrated as a power communication frequency configured sensor. Based on the data monitoring range corresponding to the power communication frequency configured sensor, a communication microcontroller that meets the data monitoring range corresponding to the power communication frequency configured sensor is determined and calibrated as a target communication microcontroller, and a radio frequency module is installed in all power communication frequency configured sensors; Perform RF configuration on the RF module based on the standard communication frequency, and perform mutual RF in sensors with different power communication frequency configurations, so that the power communication frequency configuration sensors receive RF signals from sensors with different power communication frequency configurations, and build a preliminary node wireless communication system; In the preliminary node wireless communication system, determining whether there is a power communication frequency configuration sensor that cannot receive the radio frequency signal, if not, marking the preliminary node wireless communication system as the node wireless communication system to be analyzed; If yes, the corresponding power communication frequency configuration sensor is calibrated as an abnormal power communication frequency configuration sensor, and the communication protocol of the abnormal power communication frequency configuration sensor is updated and the battery is replaced to obtain the wireless communication system of the node to be analyzed; In the wireless communication system of the node to be analyzed, a communication test time is preset, and it is determined whether the wireless communication system of the node to be analyzed has a power communication frequency configuration sensor to stop receiving radio frequency signals during the communication test time. If not, the wireless communication system of the node to be analyzed is calibrated as a qualified node wireless communication system; If so, the communication module and the radio frequency module are replaced in the power communication frequency configuration sensor until the power communication frequency configuration sensor stops receiving radio frequency signals when there is no wireless communication system of the node to be analyzed within the communication test time.

[0008] Furthermore, in a preferred embodiment of the present invention, the step S106 is specifically as follows: Based on the power communication frequency configuration sensor, when the target power equipment is running, the node real-time power data at all the first-class target cable nodes are monitored, wherein the node real-time power data includes the real-time voltage and real-time current at the first-class target cable nodes; If a type of target cable node surface has a node surface abnormal state, the corresponding type of target cable node is marked as a type of target cable to-be-analyzed node, and a type of target cable node without a node surface abnormal state is marked as a type of target cable normal node; Perform real-time analysis on the node real-time power data of the first-class target cable to be analyzed node, determine whether the real-time voltage and real-time current of the first-class target cable to be analyzed node are maintained within a preset threshold value, if not, mark the first-class target cable to be analyzed node as a first-class target cable power data abnormal node; If yes, then the target cable node to be analyzed is marked as a target cable environment data abnormal node; Introducing visual simulation software, constructing a visual simulation model of the target cable, marking a type of normal target cable node, a type of abnormal target cable power data node, and a type of abnormal target cable environmental data node in the visual simulation model of the target cable, and marking corresponding abnormal node surface states on all the nodes to be analyzed of the type one target cable, to obtain a real-time visual simulation model of the target cable; The real-time visualization simulation model of the target cable is data-fused with the qualified node wireless communication system to construct a cable visualization fault location model for locating the fault position of the target cable. The fault position of the target cable is located in real time based on the cable visualization fault location model, and a neural network algorithm is introduced to perform real-time prediction of the fault position.

[0009] Furthermore, in a preferred embodiment of the present invention, the cable visual fault location model is constructed to locate the fault position of the target cable, the fault position of the target cable is located in real time based on the cable visual fault location model, and a neural network algorithm is introduced to predict the fault position in real time, specifically: Obtaining model parameters of the target cable real-time visualization simulation model, and obtaining system parameters of the qualified node wireless communication system, determining the model parameter standard format, converting the parameter format of the model parameters of the target cable real-time visualization simulation model into the model parameter standard format, and converting the parameter format of the system parameters of the qualified node wireless communication system into the model parameter standard format; A principal component transformation algorithm is introduced. Based on the principal component transformation algorithm, the model parameters of the target cable real-time visualization simulation model whose parameter format is equal to the model parameter standard format and the system parameters of the qualified node wireless communication system are subjected to parameter principal component change fusion, and a cable visualization fault location model is constructed after the parameter principal component change fusion; The cable visual fault location model is used to locate the fault position in the target cable in a visual image manner, wherein the fault position in the target cable includes a type of target cable power data abnormal node and a type of target cable environmental data abnormal node; A convolutional neural network is introduced. In the cable visual fault location model, the real-time power data of a target cable node is convolutionally analyzed through the convolutional neural network. At the same time, the environmental parameters of a target cable node are obtained, and the environmental parameters of a target cable node are further convolutionally analyzed through the convolutional neural network. If, after the convolution analysis, the real-time power data of a target cable node or the environmental parameters of a target cable node are not maintained within the preset range, the corresponding target cable node will be calibrated as a target cable predicted abnormal node, and the target cable predicted abnormal node will be located in the cable visualization fault location model in a visualized image manner.

[0010] The second aspect of the present invention further provides a cable distributed fault perception system based on visual images, the cable distributed fault perception system includes a memory and a processor, the memory stores a cable distributed fault perception method, and when the cable distributed fault perception method is executed by the processor, the following steps are implemented: Determine the fault-prone position of the cable, install a power data sensor at the fault-prone position of the cable, collect real-time images of the fault-prone position of the cable, and perform image enhancement and image feature analysis on the obtained real-time images of the fault-prone position of the cable; Perform communication configuration and communication association for different types of target cable nodes to build a qualified node wireless communication system; The power communication frequency is controlled at all Class I target cable nodes and sensors are configured to monitor power data in real time. Combined with the abnormal node surface status at the Class I target cable nodes and the qualified node wireless communication system, a cable visual fault location model is constructed to locate the fault position of the target cable. A neural network algorithm is introduced to predict the fault position in real time.

[0011] The present invention solves the technical defects existing in the background technology, and the present invention has the following beneficial effects: determining the key positions of the cable line, installing power data sensors at the key positions of the cable line, and collecting real-time images of the corresponding positions for status analysis; building a wireless communication system of nodes to build a visual image algorithm model, locate the fault position of the target cable, and introduce a neural network algorithm to predict the fault position in real time, and finally trace the distribution position of the cable fault and formulate a repair plan for the cable fault position. The present invention realizes the perception of the fault position of the cable through visual images, and now monitors, accurately locates and responds quickly to cable line faults in real time, providing strong guarantees for the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0013] Figure 1 A flow chart of a cable distributed fault perception method based on visual images is shown; Figure 2 The flowchart of constructing a cable visual fault location model, locating the fault location of the target cable, and introducing a neural network algorithm to perform real-time prediction of the fault location is shown; Figure 3 A program view of a cable distributed fault sensing system based on visualization images is shown. DETAILED DESCRIPTION

[0014] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0015] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0016] Figure 1 A flow chart of a cable distributed fault perception method based on a visualization image is shown, comprising the following steps: S102: determining a fault-prone position of the cable, installing a power data sensor at the fault-prone position of the cable, collecting a real-time image of the fault-prone position of the cable, and performing image enhancement and image feature analysis on the obtained real-time image of the fault-prone position of the cable; S104: performing communication configuration and communication association on target cable nodes of different categories to build a qualified node wireless communication system; S106: Control the power communication frequency at all the first-class target cable nodes to configure sensors for real-time monitoring of power data, and build a cable visual fault location model in combination with the node surface abnormal state at the first-class target cable nodes and the qualified node wireless communication system to locate the fault position of the target cable, and introduce a neural network algorithm to predict the fault position in real time.

[0017] Furthermore, in a preferred embodiment of the present invention, the step S102 is specifically: The cable that needs to be fault-detected is marked as the target cable, and all the power equipment connected to the target cable is obtained and marked as the target power equipment; Obtaining a connection method between a target cable and a target power device, and obtaining a big data network at the same time, retrieving historical failure rates at different locations on the target cable based on the connection method between the target cable and the target power device based on the big data network, and presetting a maximum failure rate; Wherein, the position on the target cable is defined as the target cable node, and the historical failure rates at different positions on the target cable are the historical failure rates at different target cable nodes; Select a target cable node whose historical failure rate is greater than the maximum failure rate, mark it as a type of target cable node, and obtain a power data reference range of the target cable, wherein the power data reference range of the target cable includes a tolerable voltage range and a tolerable current range of the target cable; Based on the power data reference range of the target cable, determine a power data sensor whose data monitoring range is equal to the power data reference range of the target cable, calibrate it as the target power data sensor, and install the target power data sensor at all the first-class target cable nodes; A real-time image acquisition device is introduced and all target power equipment is operated. During the operation of the target power equipment, images at a type of target cable nodes are acquired in real time by the real-time image acquisition device and calibrated as a type of real-time image.

[0018] It should be noted that the cable is prone to distributed faults during operation, and the points where the cable is prone to faults need to be focused on, so first determine the points with a greater possibility of faults and focus on them. A class of target cable nodes can be obtained by comparing the historical failure rates. Since power data needs to be monitored in different class I target cable nodes, the monitoring range of the detection sensor needs to meet the power data reference range at the class I target cable node during monitoring, otherwise the monitoring data is likely to be inaccurate. Determine the power data sensor whose data monitoring range is equal to the power data reference range of the target cable, calibrate it as the target power data sensor, and install the target power data sensor at all class I target cable nodes to realize real-time monitoring of power data. The real-time image acquisition device is used to collect the fault conditions of the cable surface at different nodes in real time. Because of the environment or the cable's own power data, it is easy to generate abnormal conditions such as sparks on the cable surface. It is necessary to judge whether it exists through real-time images, and finally obtain a class of real-time images.

[0019] Furthermore, in a preferred embodiment of the present invention, the image enhancement processing and image feature extraction are performed on a class of real-time images to determine the image features of a class of enhanced real-time images, specifically: A class of real-time images is grayed, a Gaussian filter device is introduced, a Gaussian filter kernel is determined in the Gaussian filter device, and the Gaussian filter kernel is traversed and slid in the class of real-time images after the graying process to implement Gaussian filtering of the class of real-time images after the graying process, thereby obtaining a class of filtered real-time images; The illumination component of the image is calculated in a class of filtered real-time images, and the illumination component of the image is logarithmically transformed to obtain the reflection component of the image; Based on the reflection component of the image, a histogram of a class of filtered real-time images is obtained, and the MSR algorithm is introduced to perform histogram equalization processing and single-scale image detail enhancement processing in the histogram of the class of filtered real-time images to obtain the enhanced reflection component of the image; The enhanced reflection component of the image is applied to a type of filtered real-time image to obtain an initial enhanced type of filtered real-time image, different pixel points on the initial enhanced type of filtered real-time image are obtained, and pixel values ​​are logarithmically transformed on different pixel points on the initial enhanced type of filtered real-time image to obtain a type of enhanced real-time image; The ORB algorithm is introduced to calculate the pixel feature vector gradient at a target cable node on a class of enhanced real-time images, and the standard gradient range of the feature vector is preset. In the class of enhanced real-time images, the pixel points whose feature vector gradient remains within the standard gradient range of the feature vector are selected and calibrated as qualified pixel points, and pixel reconstruction is performed on all qualified pixel points to obtain the image features of a class of enhanced real-time images. Feature analysis is performed on image features of a type of enhanced real-time images to determine abnormal node surface states at a type of target cable nodes, wherein the abnormal node surface states include surface arcs, surface sparks, and surface smoke.

[0020] It should be noted that image feature extraction is performed in a class of real-time images to determine whether the surface state of the image is abnormal. Before image feature extraction, the image needs to be preprocessed, including grayscale processing and filtering processing. Therefore, the Gaussian filter kernel is traversed and slid in the class of real-time images after grayscale processing. During the sliding process of the Gaussian filter kernel, the kernel performs Gaussian weighted average processing of pixel values ​​on the covered pixels to achieve the purpose of filtering. When traversing and sliding, it proves that all pixels of the image are filtered. After obtaining the filtered image, since the image may be blurred after filtering, image enhancement is required to extract features more accurately. The MSR algorithm simulates the human visual system and enhances details and colors by separating the illumination and reflection components of the image. Before using the MSR algorithm, the illumination component of the image is calculated to extract the reflection component. This step determines the color and contrast of the final image. The extracted reflection component R is further enhanced to improve the brightness and contrast of the image. This can be achieved through histogram equalization, and the image is enhanced at a single scale at different scales through the MSR algorithm, and the results are accumulated and averaged to obtain the final enhanced image. This can better preserve the image details while avoiding over-enhancement or under-enhancement that may occur at a single scale, and finally obtain a class of enhanced real-time images. After image enhancement, the image needs to be feature extracted. The ORB algorithm first searches for special areas (key points) in the image, then calculates the corresponding feature vector for each key point, and obtains qualified pixels based on the corresponding feature vector to construct the image features of a class of enhanced real-time images. By analyzing the image features of a class of enhanced real-time images, it can be determined whether there is an abnormal state at the node position.

[0021] Furthermore, in a preferred embodiment of the present invention, the step S104 is specifically: Installing communication modules in target power data sensors at all target cable nodes of the first type to obtain power communication sensors, presetting standard communication frequencies, and configuring the communication frequencies of all power communication sensors to be the standard communication frequencies; The power communication sensor after the communication frequency is configured is calibrated as a power communication frequency configured sensor. Based on the data monitoring range corresponding to the power communication frequency configured sensor, a communication microcontroller that meets the data monitoring range corresponding to the power communication frequency configured sensor is determined and calibrated as a target communication microcontroller, and a radio frequency module is installed in all power communication frequency configured sensors; Perform RF configuration on the RF module based on the standard communication frequency, and perform mutual RF in sensors with different power communication frequency configurations, so that the power communication frequency configuration sensors receive RF signals from sensors with different power communication frequency configurations, and build a preliminary node wireless communication system; In the preliminary node wireless communication system, determining whether there is a power communication frequency configuration sensor that cannot receive the radio frequency signal, if not, marking the preliminary node wireless communication system as the node wireless communication system to be analyzed; If yes, the corresponding power communication frequency configuration sensor is calibrated as an abnormal power communication frequency configuration sensor, and the communication protocol of the abnormal power communication frequency configuration sensor is updated and the battery is replaced to obtain the wireless communication system of the node to be analyzed; In the wireless communication system of the node to be analyzed, a communication test time is preset, and it is determined whether the wireless communication system of the node to be analyzed has a power communication frequency configuration sensor to stop receiving radio frequency signals during the communication test time. If not, the wireless communication system of the node to be analyzed is calibrated as a qualified node wireless communication system; If so, the communication module and the radio frequency module are replaced in the power communication frequency configuration sensor until the power communication frequency configuration sensor stops receiving radio frequency signals when there is no wireless communication system of the node to be analyzed within the communication test time.

[0022] It should be noted that the purpose of building a node wireless communication system is that each node has independent image acquisition and transmission capabilities. These nodes are connected to each other through a wireless communication network to form a distributed monitoring system. When a node detects a fault, it will immediately transmit relevant information to the central monitoring center to achieve rapid perception and positioning of the fault, so it is necessary to build a node wireless communication system. Communication modules need to be installed at different nodes to achieve the purpose of communication between different nodes, and RF modules are used to connect different nodes. The frequency of the communication module should be maintained at the same frequency, and a communication microcontroller with sufficient processing power and memory size should be selected to meet the data monitoring range corresponding to the power communication frequency configuration sensor. In the process of mutual reception of RF signals, if there is a sensor that cannot receive the RF signal, it is judged that it is abnormal, which may be due to backward communication protocol or insufficient power in the sensor, and needs to be replaced. After dealing with the problem of abnormal reception of RF signals, a node wireless communication system is constructed, and it is judged whether the power communication frequency configuration sensor of the node wireless communication system to be analyzed stops receiving RF signals during the communication test time. If it is proved that there is a process problem inside the RF module, that is, there is a fault, it needs to be replaced, and finally a qualified node wireless communication system is obtained.

[0023] Figure 2 The flowchart of constructing a cable visual fault location model, locating the fault location of the target cable, and introducing a neural network algorithm for real-time prediction of the fault location is shown, including the following steps: S202: Control the power communication frequency at all the first-class target cable nodes to configure sensors for real-time monitoring of power data, and build a cable visual fault location model based on the abnormal state of the node surface at the first-class target cable nodes and the qualified node wireless communication system to locate the fault position of the target cable, and introduce a neural network algorithm to predict the fault position in real time; S204: constructing a cable visual fault location model for locating the fault position of the target cable, locating the fault position of the target cable in real time based on the cable visual fault location model, and introducing a neural network algorithm for real-time prediction of the fault position.

[0024] Furthermore, in a preferred embodiment of the present invention, the step S202 is specifically: Based on the power communication frequency configuration sensor, when the target power equipment is running, the node real-time power data at all the first-class target cable nodes are monitored, wherein the node real-time power data includes the real-time voltage and real-time current at the first-class target cable nodes; If a type of target cable node surface has a node surface abnormal state, the corresponding type of target cable node is marked as a type of target cable to-be-analyzed node, and a type of target cable node without a node surface abnormal state is marked as a type of target cable normal node; Perform real-time analysis on the node real-time power data of the first-class target cable to be analyzed node, determine whether the real-time voltage and real-time current of the first-class target cable to be analyzed node are maintained within a preset threshold value, if not, mark the first-class target cable to be analyzed node as a first-class target cable power data abnormal node; If yes, then the target cable node to be analyzed is marked as a target cable environment data abnormal node; Introducing visual simulation software, constructing a visual simulation model of the target cable, marking a type of normal target cable node, a type of abnormal target cable power data node, and a type of abnormal target cable environmental data node in the visual simulation model of the target cable, and marking corresponding abnormal node surface states on all the nodes to be analyzed of the type one target cable, to obtain a real-time visual simulation model of the target cable; The real-time visualization simulation model of the target cable is data-fused with the qualified node wireless communication system to construct a cable visualization fault location model for locating the fault position of the target cable. The fault position of the target cable is located in real time based on the cable visualization fault location model, and a neural network algorithm is introduced to perform real-time prediction of the fault position.

[0025] It should be noted that the power communication frequency configuration sensor is used to monitor the real-time power data of all the nodes at the first-class target cable nodes. The real-time power data is the key data that affects whether the node is in an abnormal state, because the node may be affected by the power data or by conditions such as excessive temperature. Therefore, according to the abnormal state of the node surface, the first-class target cable nodes are classified to obtain a class of normal target cable nodes, a class of abnormal target cable power data nodes, and a class of abnormal target cable environmental data nodes. Visual simulation software is used to build a simulation model with visual images. The advantage of the model with visual images is that it can clearly know the location of the abnormality in the cable and highlight the abnormal state, which is clear at a glance. After building the real-time visual simulation model of the target cable, it is necessary to introduce a qualified node wireless communication system into the model to achieve the purpose of intelligent perception and positioning of the model. The qualified node wireless communication system uploads data to the panel end to clearly know the fault location.

[0026] Furthermore, in a preferred embodiment of the present invention, the step S202 is specifically: Obtaining model parameters of the target cable real-time visualization simulation model, and obtaining system parameters of the qualified node wireless communication system, determining the model parameter standard format, converting the parameter format of the model parameters of the target cable real-time visualization simulation model into the model parameter standard format, and converting the parameter format of the system parameters of the qualified node wireless communication system into the model parameter standard format; A principal component transformation algorithm is introduced. Based on the principal component transformation algorithm, the model parameters of the target cable real-time visualization simulation model whose parameter format is equal to the model parameter standard format and the system parameters of the qualified node wireless communication system are subjected to parameter principal component change fusion, and a cable visualization fault location model is constructed after the parameter principal component change fusion; The cable visual fault location model is used to locate the fault position in the target cable in a visual image manner, wherein the fault position in the target cable includes a type of target cable power data abnormal node and a type of target cable environmental data abnormal node; A convolutional neural network is introduced. In the cable visual fault location model, the real-time power data of a target cable node is convolutionally analyzed through the convolutional neural network. At the same time, the environmental parameters of a target cable node are obtained, and the environmental parameters of a target cable node are further convolutionally analyzed through the convolutional neural network. If, after the convolution analysis, the real-time power data of a target cable node or the environmental parameters of a target cable node are not maintained within the preset range, the corresponding target cable node will be calibrated as a target cable predicted abnormal node, and the target cable predicted abnormal node will be located in the cable visualization fault location model in a visualized image manner.

[0027] It should be noted that the data fusion of the target cable real-time visualization simulation model and the qualified node wireless communication system requires the format of the underlying parameters of the two to be unified. After the format is unified, the role of the principal component transformation algorithm is to fuse the underlying parameters of the two after the format is unified, so as to realize the communication of the status between nodes in the target cable real-time visualization simulation model and realize the purpose of intelligent perception and positioning of the model, and finally obtain the cable visualization fault location model. In the cable visualization fault location model, the nodes are analyzed. If there are nodes that are currently normal and have no abnormalities, but if the current and voltage are abnormal, there may still be abnormal states in the future, which will affect the normal operation of the cable, so it is necessary to introduce a convolutional neural network model in the model. The abnormal prediction of the degree node is realized through the convolutional neural network model. According to the current environmental parameters and combined with the real-time power data, it can be judged whether the point will become a type of target cable prediction abnormal node in the future.

[0028] In addition, the cable distributed fault perception method based on visual images also includes the following steps: For a type of target cable power data abnormal node, obtain the equipment load of the target power equipment, and adjust the load value of the equipment load of the target power equipment to determine whether the real-time voltage and real-time current of the type of target cable power data abnormal node are maintained within a preset threshold after the load value adjustment; If not, the target power equipment is controlled to stop running, and an electric pen is used to monitor the copper wire short circuit in the target cable at a type-one target cable power data abnormal node to determine whether there is a copper wire short circuit in the target cable. If so, simultaneous connection processing is required in a type-one target cable power data abnormal node of the target cable to control the copper wire in the target cable to be free of short circuit; If there is no short circuit in the copper wire of the target cable, and the real-time voltage and real-time current of the abnormal power data node of the target cable are still not maintained within the preset threshold, the target cable shall be replaced; For a type of target cable environmental data abnormal node, based on the image recognition method, it is determined whether the target cable has surface damage at the type of target cable environmental data abnormal node. If so, the target cable is directly replaced; If not, then searching the big data network for qualified operating environment parameters of the target cable, and at the same time searching for and outputting a solution for maintaining the operating environment parameters of the target cable equal to the qualified operating environment parameters; For a class of target cable predicted abnormal nodes, an operation and control scheme of a class of target cable power data abnormal nodes and a class of target cable environmental data abnormal nodes is applied in the class of target cable predicted abnormal nodes.

[0029] It should be noted that a type of target cable power data abnormal node is a node abnormality caused by abnormal real-time power data, so it is judged that the load is too high, resulting in a short circuit risk at the node, and load regulation is required. After load regulation, due to the previous high load, the copper wire inside the node may be poor or broken, resulting in a short circuit, which causes abnormal real-time power data, so copper wire maintenance is required. If the real-time circuit data is still abnormal after the copper wire maintenance, the cable is directly replaced to prevent the cable from causing danger during operation. For a type of target cable environmental data abnormal node, the environmental data includes whether the cable surface is damaged, and whether the ambient temperature and humidity around the cable are maintained within the preset range. If it is damaged, it cannot be repaired, and the cable is directly replaced. If it is not damaged, the environmental data is regulated. For a class of target cable predicted abnormal nodes, since a class of target cable predicted abnormal nodes will later cause a class of target cable power data abnormal nodes and a class of target cable environmental data abnormal nodes, the cables at the class of target cable predicted abnormal nodes are processed in advance, and the processing methods are equal to the above two, namely, the operation and control plans for the class of target cable power data abnormal nodes and the class of target cable environmental data abnormal nodes.

[0030] like Figure 3 As shown, the second aspect of the present invention further provides a cable distributed fault perception system based on visual images, the cable distributed fault perception system includes a memory and a processor, the memory stores a cable distributed fault perception method, and when the cable distributed fault perception method is executed by the processor, the following steps are implemented: Determine the fault-prone position of the cable, install a power data sensor at the fault-prone position of the cable, collect real-time images of the fault-prone position of the cable, and perform image enhancement and image feature analysis on the obtained real-time images of the fault-prone position of the cable; Perform communication configuration and communication association for different types of target cable nodes to build a qualified node wireless communication system; The power communication frequency is controlled at all Class I target cable nodes and sensors are configured to monitor power data in real time. Combined with the abnormal node surface status at the Class I target cable nodes and the qualified node wireless communication system, a cable visual fault location model is constructed to locate the fault position of the target cable. A neural network algorithm is introduced to predict the fault position in real time.

[0031] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A cable distributed fault perception method based on visual images, characterized in that: The following steps are involved: S102: determining a fault-prone position of the cable, installing a power data sensor at the fault-prone position of the cable, collecting a real-time image of the fault-prone position of the cable, and performing image enhancement and image feature analysis on the obtained real-time image of the fault-prone position of the cable; S104: performing communication configuration and communication association on target cable nodes of different categories to build a qualified node wireless communication system; S106: Control the power communication frequency at all the first-class target cable nodes to configure sensors for real-time monitoring of power data, and build a cable visual fault location model in combination with the node surface abnormal state at the first-class target cable nodes and the qualified node wireless communication system to locate the fault position of the target cable, and introduce a neural network algorithm to predict the fault position in real time.

2. The cable distributed fault perception method based on visual images according to claim 1 is characterized in that: The S102 is specifically: The cable that needs to be fault-detected is marked as the target cable, and all the power equipment connected to the target cable is obtained and marked as the target power equipment; Obtaining a connection method between a target cable and a target power device, and obtaining a big data network at the same time, retrieving historical failure rates at different locations on the target cable based on the connection method between the target cable and the target power device based on the big data network, and presetting a maximum failure rate; Wherein, the position on the target cable is defined as the target cable node, and the historical failure rates at different positions on the target cable are the historical failure rates at different target cable nodes; Select a target cable node whose historical failure rate is greater than the maximum failure rate, mark it as a type of target cable node, and obtain a power data reference range of the target cable, wherein the power data reference range of the target cable includes a tolerable voltage range and a tolerable current range of the target cable; Based on the power data reference range of the target cable, determine a power data sensor whose data monitoring range is equal to the power data reference range of the target cable, calibrate it as the target power data sensor, and install the target power data sensor at all the first-class target cable nodes; A real-time image acquisition device is introduced, and all target power equipment is operated. During the operation of the target power equipment, images at a type of target cable node are acquired in real time by the real-time image acquisition device and calibrated as a type of real-time image; Image enhancement processing and image feature extraction are performed on a class of real-time images to determine image features of a class of enhanced real-time images.

3. The cable distributed fault perception method based on visual images according to claim 2 is characterized in that: The image enhancement processing and image feature extraction are performed on a class of real-time images to determine the image features of a class of enhanced real-time images, specifically: A class of real-time images is grayed, a Gaussian filter device is introduced, a Gaussian filter kernel is determined in the Gaussian filter device, and the Gaussian filter kernel is traversed and slid in the class of real-time images after the graying process to implement Gaussian filtering of the class of real-time images after the graying process, thereby obtaining a class of filtered real-time images; The illumination component of the image is calculated in a class of filtered real-time images, and the illumination component of the image is logarithmically transformed to obtain the reflection component of the image; Based on the reflection component of the image, a histogram of a class of filtered real-time images is obtained, and the MSR algorithm is introduced to perform histogram equalization processing and single-scale image detail enhancement processing in the histogram of the class of filtered real-time images to obtain the enhanced reflection component of the image; The enhanced reflection component of the image is applied to a type of filtered real-time image to obtain an initial enhanced type of filtered real-time image, different pixel points on the initial enhanced type of filtered real-time image are obtained, and pixel values ​​are logarithmically transformed on different pixel points on the initial enhanced type of filtered real-time image to obtain a type of enhanced real-time image; The ORB algorithm is introduced to calculate the pixel feature vector gradient at a target cable node on a class of enhanced real-time images, and the standard gradient range of the feature vector is preset. In the class of enhanced real-time images, the pixel points whose feature vector gradient remains within the standard gradient range of the feature vector are selected and calibrated as qualified pixel points, and pixel reconstruction is performed on all qualified pixel points to obtain the image features of a class of enhanced real-time images. Feature analysis is performed on image features of a type of enhanced real-time images to determine abnormal node surface states at a type of target cable nodes, wherein the abnormal node surface states include surface arcs, surface sparks, and surface smoke.

4. The cable distributed fault perception method based on visual images according to claim 1 is characterized in that: The S104 is specifically: Installing communication modules in target power data sensors at all target cable nodes of the first type to obtain power communication sensors, presetting standard communication frequencies, and configuring the communication frequencies of all power communication sensors to be the standard communication frequencies; The power communication sensor after the communication frequency is configured is calibrated as a power communication frequency configured sensor. Based on the data monitoring range corresponding to the power communication frequency configured sensor, a communication microcontroller that meets the data monitoring range corresponding to the power communication frequency configured sensor is determined and calibrated as a target communication microcontroller, and a radio frequency module is installed in all power communication frequency configured sensors; Perform RF configuration on the RF module based on the standard communication frequency, and perform mutual RF in sensors with different power communication frequency configurations, so that the power communication frequency configuration sensors receive RF signals from sensors with different power communication frequency configurations, and build a preliminary node wireless communication system; In the preliminary node wireless communication system, determining whether there is a power communication frequency configuration sensor that cannot receive the radio frequency signal, if not, marking the preliminary node wireless communication system as the node wireless communication system to be analyzed; If yes, the corresponding power communication frequency configuration sensor is calibrated as an abnormal power communication frequency configuration sensor, and the communication protocol of the abnormal power communication frequency configuration sensor is updated and the battery is replaced to obtain the wireless communication system of the node to be analyzed; In the wireless communication system of the node to be analyzed, a communication test time is preset, and it is determined whether the wireless communication system of the node to be analyzed has a power communication frequency configuration sensor to stop receiving radio frequency signals during the communication test time. If not, the wireless communication system of the node to be analyzed is calibrated as a qualified node wireless communication system; If so, the communication module and the radio frequency module are replaced in the power communication frequency configuration sensor until the power communication frequency configuration sensor stops receiving radio frequency signals when there is no wireless communication system of the node to be analyzed within the communication test time.

5. The cable distributed fault perception method based on visual images according to claim 1 is characterized in that: The S106 is specifically: Based on the power communication frequency configuration sensor, when the target power equipment is running, the node real-time power data at all the first-class target cable nodes are monitored, wherein the node real-time power data includes the real-time voltage and real-time current at the first-class target cable nodes; If a type of target cable node surface has a node surface abnormal state, the corresponding type of target cable node is marked as a type of target cable to-be-analyzed node, and a type of target cable node without a node surface abnormal state is marked as a type of target cable normal node; Perform real-time analysis on the node real-time power data of the first-class target cable to be analyzed node, determine whether the real-time voltage and real-time current of the first-class target cable to be analyzed node are maintained within a preset threshold value, if not, mark the first-class target cable to be analyzed node as a first-class target cable power data abnormal node; If yes, then the target cable node to be analyzed is marked as a target cable environment data abnormal node; Introducing visual simulation software, constructing a visual simulation model of the target cable, marking a type of normal target cable node, a type of abnormal target cable power data node, and a type of abnormal target cable environmental data node in the visual simulation model of the target cable, and marking corresponding abnormal node surface states on all the nodes to be analyzed of the type one target cable, to obtain a real-time visual simulation model of the target cable; The real-time visualization simulation model of the target cable is data-fused with the qualified node wireless communication system to construct a cable visualization fault location model for locating the fault position of the target cable. The fault position of the target cable is located in real time based on the cable visualization fault location model, and a neural network algorithm is introduced to perform real-time prediction of the fault position.

6. The cable distributed fault perception method based on visual images according to claim 5, characterized in that: The cable visual fault location model is constructed to locate the fault position of the target cable, the fault position of the target cable is located in real time based on the cable visual fault location model, and a neural network algorithm is introduced to predict the fault position in real time, specifically: Obtaining model parameters of the target cable real-time visualization simulation model, and obtaining system parameters of the qualified node wireless communication system, determining the model parameter standard format, converting the parameter format of the model parameters of the target cable real-time visualization simulation model into the model parameter standard format, and converting the parameter format of the system parameters of the qualified node wireless communication system into the model parameter standard format; A principal component transformation algorithm is introduced. Based on the principal component transformation algorithm, the model parameters of the target cable real-time visualization simulation model whose parameter format is equal to the model parameter standard format and the system parameters of the qualified node wireless communication system are subjected to parameter principal component change fusion, and a cable visualization fault location model is constructed after the parameter principal component change fusion; The cable visual fault location model is used to locate the fault position in the target cable in a visual image manner, wherein the fault position in the target cable includes a type of target cable power data abnormal node and a type of target cable environmental data abnormal node; A convolutional neural network is introduced. In the cable visual fault location model, the real-time power data of a target cable node is convolutionally analyzed through the convolutional neural network. At the same time, the environmental parameters of a target cable node are obtained, and the environmental parameters of a target cable node are further convolutionally analyzed through the convolutional neural network. If, after the convolution analysis, the real-time power data of a target cable node or the environmental parameters of a target cable node are not maintained within the preset range, the corresponding target cable node will be calibrated as a target cable predicted abnormal node, and the target cable predicted abnormal node will be located in the cable visualization fault location model in a visualized image manner.

7. The cable distributed fault perception system based on visual images is characterized by: The cable distributed fault sensing system includes a memory and a processor, wherein the memory stores a cable distributed fault sensing method program, and when the cable distributed fault sensing method program is executed by the processor, the steps of the cable distributed fault sensing method based on visual images as described in any one of claims 1 to 6 are implemented: Determine the fault-prone position of the cable, install a power data sensor at the fault-prone position of the cable, collect real-time images of the fault-prone position of the cable, and perform image enhancement and image feature analysis on the obtained real-time images of the fault-prone position of the cable; Perform communication configuration and communication association for different types of target cable nodes to build a qualified node wireless communication system; The power communication frequency is controlled at all Class I target cable nodes and sensors are configured to monitor power data in real time. Combined with the abnormal node surface status at the Class I target cable nodes and the qualified node wireless communication system, a cable visual fault location model is constructed to locate the fault position of the target cable. A neural network algorithm is introduced to predict the fault position in real time.

Images