Long-distance current collection line fault on-line positioning system

Abstract

The invention relates to the field of wind power plant current collection lines, in particular to a long-distance current collection line fault online positioning system. A long-distance current collection line fault on-line positioning system comprises a fault detection module, a traveling wave signal acquisition module, a fault classification module and a fault positioning module. According to the invention, the long-distance current collection line is divided into a plurality of sections, and the risk assessment is carried out on the sections according to the historical fault data, so that the refined recognition of the fault area is realized; the segmentation strategy not only improves the pertinence and accuracy of fault detection, but also helps to quickly narrow the fault search range, thereby significantly improving the efficiency of fault positioning. Through segmented management, monitoring and maintenance resources are reasonably distributed according to the risk level of each segment.

Description

Technical Field

[0001] The invention relates to the field of wind farm collector lines, and in particular to an online fault positioning system for long-distance collector lines. Background Art

[0002] In modern power systems, long-distance collector lines, as the key link between wind farms and power grids, undertake the important task of efficiently transmitting electric energy. However, due to their wide geographical span and complex and changeable environmental conditions, long-distance collector lines face many fault risks. These faults not only affect the continuity and stability of power supply, but may also cause damage to equipment and even cause safety accidents. Therefore, the development of an efficient and accurate online fault location system for long-distance collector lines is of great significance to ensuring the safe and reliable operation of power systems.

[0003] Traditional fault location technologies mainly include protection relay action, time domain reflectometry (TDR) and other methods; although these methods can realize fault detection and location to a certain extent, they have some limitations; in order to overcome these limitations, traveling wave location technology came into being; traveling wave location technology uses traveling wave signals propagated in the power system when a fault occurs, and determines the location of the fault by detecting and analyzing these signals; compared with traditional methods, traveling wave location technology has higher accuracy and faster response speed, and is especially suitable for fault location of long-distance collector lines. Summary of the invention

[0004] The present invention achieves refined identification of fault areas by dividing long-distance collector lines into multiple sections and performing risk assessment on these sections based on historical fault data; this segmentation strategy not only improves the pertinence and accuracy of fault detection, but also helps to quickly narrow the fault search scope, thereby significantly improving the efficiency of fault location; through segmented management, monitoring and maintenance resources can be reasonably allocated according to the risk level of each section.

[0005] A long-distance collector line fault online positioning system, comprising: A fault detection module is used to determine whether there is a fault in the current collector line according to the current value and voltage value of the detection node obtained; The traveling wave signal acquisition module is used to acquire the original traveling wave signals of all detection nodes and process the original traveling wave signals using adaptive filtering technology to acquire advanced traveling wave signals; A fault classification module is used to input the advanced traveling wave signal into the fault classification model to obtain fault type information; The fault location module is used to obtain the fault section where the fault point is located according to the time when each node collects the traveling wave signal, and calculate the location of the fault point through the arrival time difference of the original traveling wave signal of the nodes at both ends of the fault section; The invention also includes a long-distance collector line fault online positioning method applied to a long-distance collector line fault online positioning system, and the specific steps are as follows: S1: Set up detection nodes at the main line end of the wind farm collection line, the access point of all wind turbines, and the connection point between the wind turbine and the main line; regard the line between two adjacent monitoring nodes as a section, and obtain several line sections; divide all line sections into several high-risk sections and low-risk sections according to historical fault occurrence data; Obtain the current value and voltage value of the detection nodes at both ends of all high-risk sections at the current first monitoring time point; obtain the current value and voltage value of the remaining detection nodes at the current second monitoring time point; For any detection node, determine whether the current value and voltage value obtained by the node at the current time are both within the normal threshold range. If so, no operation is performed; if not, proceed to step S2; S2: For all detection nodes, the original traveling wave signal is obtained; the original traveling wave signal is processed by using adaptive filtering technology to obtain an advanced traveling wave signal; S3: inputting the advanced traveling wave signal into the fault classification model to obtain fault type information; S4: According to the time when each node collects the traveling wave signal, the fault section where the fault point is located is obtained; the location of the fault point is calculated and obtained through the arrival time difference of the original traveling wave signals at the nodes at both ends of the fault section.

[0006] Preferably, the specific operation of obtaining the fault section where the fault point is located is: Record the time when the original traveling wave signal arrives at each detection node, and obtain the time difference when the original traveling wave signal arrives at each detection node; calculate and obtain the theoretical time difference between each detection node based on the time difference of each detection node and the speed of the original traveling wave signal; for any line section, compare the time difference of the detection nodes at both ends of the line section with the theoretical time difference. If the difference exceeds the normal threshold, the line section is judged to be a fault section.

[0007] Preferably, the specific operation of calculating and obtaining the fault point location is: Obtain the distance between the detection nodes at both ends of the fault section; calculate the sum of the time it takes for the traveling wave signal to reach the detection nodes at both ends of the fault section to obtain the total time; take the detection node that first collects the traveling wave signal as the nearest node, and divide the time it takes for the traveling wave signal to reach the nearest node by the total time to obtain the distance ratio; multiply the distance between the detection nodes at both ends by the distance ratio to obtain the distance between the fault point and the nearest node; obtain the location of the fault point based on the distance between the fault point and the nearest node.

[0008] Preferably, the adaptive filtering technology adopts LMS algorithm.

[0009] Preferably, the specific operation of processing the original traveling wave signal to obtain the advanced traveling wave signal is: A1: Set the order M of the LMS filter, initialize the filter coefficient to zero, and set the learning rate; A2: For any detection node, the original traveling wave signal obtained by the detection node and the original traveling wave signals obtained by the first M-1 detection nodes are used as the input of the LMS filter, and the output value of the LMS filter is compared with the preset expected value to obtain the error value; the filter coefficient of the LMS filter is updated based on the error value; A3: Repeat the above operation until the preset maximum number of iterations is reached; for any detection node, the original traveling wave signal obtained by the detection node and the original traveling wave signals obtained by the first M-1 detection nodes are again used as the input of the LMS filter to obtain the high-level traveling wave signal.

[0010] Preferably, the fault classification model is established based on the CNN model, including an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer; the input layer is used to input high-level traveling wave signals; the convolution layer is used to extract the features of high-level traveling wave signals; the pooling layer is used to reduce the spatial dimension of the features; the fully connected layer is used to further learn the feature representation; the output layer is used to output the fault type information.

[0011] Preferably, for the training of the fault classification model, the specific operations are: Acquire several fault classification training samples, each of which contains a high-level traveling wave signal and corresponding fault type information; divide all acquired fault classification training samples into a fault classification training set and a fault classification test set; use the fault classification training set to train the fault classification model, and then input the fault classification test set into the fault classification model to obtain the test result; set the training conditions, and if the test result meets the training conditions, output the trained fault classification model; otherwise, continue to use the fault classification training set to train the fault classification model.

[0012] Preferably, all line sections are divided into several high-risk sections and low-risk sections. Obtain historical fault occurrence data, and calculate the historical fault occurrence frequency of each line section based on the historical fault occurrence data; set a risk assessment threshold, and classify the line sections with historical fault occurrence frequencies higher than the risk assessment threshold as high-risk sections; and classify the line sections with historical fault occurrence frequencies lower than the risk assessment threshold as low-risk sections.

[0013] The present invention has the following advantages: 1. The present invention divides the long-distance collector line into multiple sections and conducts risk assessment on these sections based on historical fault data, thereby realizing refined identification of the fault area; this segmentation strategy not only improves the pertinence and accuracy of fault detection, but also helps to quickly narrow the fault search scope, thereby significantly improving the efficiency of fault location; through segmentation management, it is possible to reasonably allocate monitoring and maintenance resources according to the risk level of each section.

[0014] 2. The present invention uses adaptive filtering technology to perform fine processing on the original traveling wave signal, effectively filters out noise and interference, and significantly improves the signal-to-noise ratio of the signal, thereby providing a solid foundation for the accurate extraction of fault features; combined with convolutional neural networks, the system can automatically learn and identify complex fault features in traveling wave signals, and achieve high-precision classification of fault types; this intelligent fault identification mechanism not only improves the accuracy of fault detection, but also reduces misjudgments caused by human factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a structural schematic diagram of a long-distance collector line fault online locating system adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0017] Embodiment, a long distance collector line fault online positioning system, such as Figure 1 As shown, including: A fault detection module is used to determine whether there is a fault in the current collector line according to the current value and voltage value of the detection node obtained; The traveling wave signal acquisition module is used to acquire the original traveling wave signals of all detection nodes and process the original traveling wave signals using adaptive filtering technology to acquire advanced traveling wave signals; A fault classification module is used to input the advanced traveling wave signal into the fault classification model to obtain fault type information; The fault location module is used to obtain the fault section where the fault point is located according to the time when each node collects the traveling wave signal, and calculate the location of the fault point through the arrival time difference of the original traveling wave signal of the nodes at both ends of the fault section; The invention also includes a long-distance collector line fault online positioning method applied to a long-distance collector line fault online positioning system, and the specific steps are as follows: S1: Set up detection nodes at the main line end of the wind farm collection line, the access point of all wind turbines, and the connection point between the wind turbine and the main line; the setting of detection nodes is the prerequisite for fault location, and these locations are key nodes for power flow; the line between two adjacent monitoring nodes is regarded as a section, and several line sections are obtained; such division helps to decompose the long-distance collection line into multiple manageable parts, which is convenient for localized and refined fault detection and management; according to historical fault occurrence data, all line sections are divided into several high-risk sections and low-risk sections, which provides a scientific basis for fault prevention and resource optimization allocation; Obtain the current and voltage values ​​of the detection nodes at both ends of all high-risk sections at the current first monitoring time point; obtain the current and voltage values ​​of the remaining detection nodes at the current second monitoring time point; the current and voltage values ​​provide the system with real-time operating status data, which is the basis for fault detection and diagnosis; the interval between each two adjacent first monitoring time points is shorter than the interval between two adjacent second monitoring time points; according to the risk assessment results, the system sets different monitoring frequencies for high-risk sections and low-risk sections; the monitoring frequency of high-risk sections is higher to achieve closer monitoring, while low-risk sections use a relatively low monitoring frequency to optimize resource allocation; For any detection node, determine whether the current value and voltage value obtained by the node at the current time are both within the normal threshold range. If so, no operation is performed; if not, proceed to step S2; The system has a built-in real-time threshold judgment mechanism that can instantly analyze the current and voltage values ​​obtained by the detection nodes to determine whether they are within the preset normal threshold range. This mechanism is the first step in fault detection and can quickly identify potential abnormalities. If the current and voltage values ​​of the detection nodes are within the normal threshold range, the system will not take further action to avoid unnecessary intervention. This intelligent decision-making logic helps reduce false alarms and waste of resources. On the contrary, if an abnormality is detected, the system will automatically trigger the S2 step to enter a more in-depth fault diagnosis and location process. Due to the higher monitoring frequency of high-risk sections, the system can respond to fault events in these areas more quickly. This rapid response capability is crucial to preventing the expansion of faults and reducing potential losses. S2: For all detection nodes, the original traveling wave signals are obtained; the traveling wave signals are electromagnetic waves caused by fault events in the power system, which propagate in the power lines at nearly the speed of light; the system captures these traveling wave signals through all detection nodes as the basic data for fault detection and location; the original traveling wave signals are processed using adaptive filtering technology to obtain advanced traveling wave signals; since the traveling wave signals may be affected by noise interference and signal attenuation during transmission, adaptive filtering technology is required to process the original signals; the adaptive filter can automatically adjust its parameters according to the characteristics of the signal to achieve the best filtering effect; the fault location system requires real-time signal processing to ensure rapid response after a fault occurs; the adaptive filtering technology can quickly process signals to meet the system's real-time requirements; Advanced traveling wave signals refer to signals that have been processed by adaptive filtering. They have a higher signal-to-noise ratio and clearer fault characteristics. These signals provide a reliable data basis for subsequent fault feature extraction and classification. S3: Input the high-level traveling wave signal into the fault classification model to obtain fault type information; the high-level traveling wave signal processed in step S2 contains rich information generated when the fault occurs; these signals serve as input data and provide the necessary raw materials for the fault classification model; Fault classification has significant benefits. It can not only significantly improve the operating efficiency and safety of the power system, but also is crucial for rapid response and accurate handling of faults. Through accurate classification, operation and maintenance personnel can quickly identify the nature and severity of the fault, so as to take the most appropriate response measures and avoid the risks and losses that may be caused by blind operation. In addition, fault classification can also help optimize resource allocation, ensure that high-risk areas receive sufficient attention and maintenance, and reasonably allocate maintenance resources in low-risk areas to maximize cost-effectiveness. The classification results can also be used to accumulate fault data, provide support for future fault prediction and health status assessment, and further enhance the system's preventive maintenance capabilities. At the same time, accurate fault classification can also promote the intelligent development of power systems and provide a basis for automated fault handling and intelligent decision-making. S4: According to the time when each node collects the traveling wave signal, the fault section where the fault point is located is obtained; by comparing the time when each node receives the traveling wave signal, the approximate section where the fault occurs can be determined; if the time when two adjacent nodes receive the traveling wave signal is significantly different, this usually means that the fault occurs in the section between the two nodes; The location of the fault point is calculated by the arrival time difference of the original traveling wave signals at the two end nodes of the fault section; the specific location of the fault point can be calculated by using a simple geometric relationship based on the arrival time difference of the traveling wave signals at the two end nodes and the known propagation speed of the traveling wave in the power line; if the distance between the two nodes and the time required for the traveling wave to propagate from the fault point to the two nodes are known, the distance from the fault point to one of the nodes can be determined by the proportional relationship; Specific operations for obtaining the fault section where the fault point is located: Record the time when the original traveling wave signal arrives at each detection node, and obtain the time difference when the original traveling wave signal arrives at each detection node; calculate and obtain the theoretical time difference between each detection node based on the time difference of each detection node and the speed of the original traveling wave signal, combined with the physical distance between the detection nodes; for any line section, compare the time difference of the detection nodes at both ends of the line section with the theoretical time difference. If the difference between the two is not large, it means that no fault has occurred in the section; if the difference is significant and exceeds the preset normal threshold, the line section is judged to be a faulty section.

[0018] Specific operations for calculating the fault point location: Obtain the distance between the detection nodes at both ends of the fault section. This distance can be obtained through geographic information system (GIS) data or other measurement methods. Calculate the sum of the time it takes for the traveling wave signal to reach the detection nodes at both ends of the fault section to obtain the total time. When a fault occurs, the traveling wave signal will propagate in two directions at the same time, so there will be two arrival times. The detection node that first collects the traveling wave signal is taken as the nearest node because it is closer to the fault point. Use the time it takes for the traveling wave signal to reach the nearest node divided by the total time to obtain the distance ratio, which is a proportional factor of the position of the fault point relative to the detection nodes at both ends. Multiply the distance between the detection nodes at both ends by the distance ratio to obtain the distance between the fault point and the nearest node. According to the distance between the fault point and the nearest node, the location of the fault point is obtained, which provides important information for rapid repair and subsequent analysis of the fault.

[0019] Adaptive filtering technology uses the LMS algorithm, the least mean square algorithm, or LMS algorithm, which is an adaptive linear filter design method that is widely used in the field of signal processing, especially in application scenarios where filter coefficients need to be updated in real time; the LMS algorithm adjusts the filter coefficients by minimizing the mean square error between the filter output and the expected response. Its core advantages lie in the simplicity, real-time nature, and low demand for computing resources of the algorithm; in each iteration, the LMS algorithm uses the current input signal and the error signal to update the filter coefficients, and in this way gradually approaches the optimal filter to achieve accurate estimation or prediction of the signal; the key parameter of the algorithm is the learning rate, which determines the step size of the coefficient update and needs to be carefully selected to ensure the convergence and stability of the algorithm; the adaptive characteristics of the LMS algorithm enable it to process non-stationary signals and is easy to implement and maintain, which makes it widely used in a variety of applications such as communication systems, echo cancellation, noise suppression, and fault detection; The specific operations for processing the original traveling wave signal to obtain the advanced traveling wave signal are as follows: A1: Set the order M of the LMS filter. The order M of the filter determines the complexity and performance of the LMS filter. The higher the order, the better the filter can approximate the target signal, but the amount of calculation will also increase. Initialize the filter coefficient to zero, which is the starting point of filter learning. Set the learning rate. An appropriate learning rate can ensure the convergence and stability of the algorithm. A2: For any detection node, the original traveling wave signal obtained by the detection node and the original traveling wave signals obtained by the first M-1 detection nodes are used as the input of the LMS filter, and the output value of the LMS filter is compared with the preset expected value to obtain the error value; the filter coefficient of the LMS filter is updated based on the error value; A3: Repeat the above operation until the preset maximum number of iterations is reached. Each iteration will make the filter coefficient closer to the optimal value, thereby improving the filtering effect. For any detection node, the original traveling wave signal obtained by the detection node and the original traveling wave signal obtained by the first M-1 detection nodes are used as the input of the LMS filter to obtain the high-level traveling wave signal.

[0020] The fault classification model is built on the basis of the CNN model. The convolutional neural network, or CNN, is a deep learning model that is particularly suitable for processing data with a distinct grid-like topological structure, such as signal sequences. The CNN model simulates the working principle of the human visual system, automatically extracts local features of the input data using the convolution layer, and then reduces the spatial dimension of the features through the pooling layer to enhance the invariance of the features, and performs high-level feature synthesis and classification decisions through the fully connected layer. The CNN model has strong generalization capabilities and can process unseen data. With proper training, it can adapt to a variety of complex pattern recognition problems. The fault classification model includes input layer, convolution layer, pooling layer, fully connected layer and output layer; the input layer is the starting point of CNN and is used to input high-level traveling wave signals; the convolution layer is the core of CNN, which extracts key features from the signal by convolving the input signal with a series of learnable convolution kernels (or filters); these features may include the frequency components and amplitude changes of the signal, which are crucial for fault identification; the pooling layer follows the convolution layer and is used to reduce the spatial dimension of the features while increasing the invariance to signal changes; this helps to reduce the amount of computation and make the model more robust to small position changes; the fully connected layer further learns the high-level representation of features through a dense neural network structure to provide more abstract features for classification; the output layer is the last layer of CNN, which usually uses the softmax activation function for multi-class classification tasks and outputs the probability or score of each type of fault to determine the fault type corresponding to the signal.

[0021] For the training of fault classification model, the specific operations are as follows: Acquire several fault classification training samples, each of which contains a high-level traveling wave signal and corresponding fault type information; divide all acquired fault classification training samples into a fault classification training set and a fault classification test set; use the fault classification training set to train the fault classification model, and then input the fault classification test set into the fault classification model to obtain the test result; set the training conditions, and if the test result meets the training conditions, output the trained fault classification model; otherwise, continue to use the fault classification training set to train the fault classification model.

[0022] Specific operations for dividing all line sections into several high-risk sections and low-risk sections: Obtain historical fault occurrence data, and calculate the historical fault occurrence frequency of each line section based on the historical fault occurrence data; set a risk assessment threshold, and classify the line sections with historical fault occurrence frequencies higher than the risk assessment threshold as high-risk sections; and classify the line sections with historical fault occurrence frequencies lower than the risk assessment threshold as low-risk sections.

[0023] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A long-distance collector line fault online positioning system, characterized in that: include: A fault detection module is used to determine whether there is a fault in the current collector line according to the current value and voltage value of the detection node obtained; The traveling wave signal acquisition module is used to acquire the original traveling wave signals of all detection nodes and process the original traveling wave signals using adaptive filtering technology to acquire advanced traveling wave signals; A fault classification module is used to input the advanced traveling wave signal into the fault classification model to obtain fault type information; The fault location module is used to obtain the fault section where the fault point is located according to the time when each node collects the traveling wave signal, and calculate the location of the fault point through the arrival time difference of the original traveling wave signal of the nodes at both ends of the fault section; The invention also includes a long-distance collector line fault online positioning method applied to a long-distance collector line fault online positioning system, and the specific steps are as follows: S1: Set up detection nodes at the main line end of the wind farm collection line, the access point of all wind turbines, and the connection point between the wind turbine and the main line; regard the line between two adjacent monitoring nodes as a section, and obtain several line sections; divide all line sections into several high-risk sections and low-risk sections according to historical fault occurrence data; Obtaining the current value and voltage value of the detection nodes at both ends of all high-risk sections at the current first monitoring time point; Obtaining current values ​​and voltage values ​​of the remaining detection nodes at the current second monitoring time point; For any detection node, determine whether the current value and voltage value obtained by the node at the current time are both within the normal threshold range. If so, no operation is performed; if not, proceed to step S2; S2: For all detection nodes, the original traveling wave signal is obtained; the original traveling wave signal is processed by using adaptive filtering technology to obtain an advanced traveling wave signal; S3: inputting the advanced traveling wave signal into the fault classification model to obtain fault type information; S4: According to the time when each node collects the traveling wave signal, the fault section where the fault point is located is obtained; the location of the fault point is calculated and obtained through the arrival time difference of the original traveling wave signals at the nodes at both ends of the fault section.

2. The long-distance collector line fault online positioning system according to claim 1 is characterized in that: Specific operations for obtaining the fault section where the fault point is located: Record the time when the original traveling wave signal arrives at each detection node, and obtain the time difference when the original traveling wave signal arrives at each detection node; calculate and obtain the theoretical time difference between each detection node according to the time difference of each detection node and the speed of the original traveling wave signal; For any line section, the time difference of the detection nodes at both ends of the line section is compared with the theoretical time difference. If the difference exceeds the normal threshold, the line section is judged to be a fault section.

3. The long-distance collector line fault online positioning system according to claim 2 is characterized in that: Specific operations for calculating the fault point location: Obtain the distance between the detection nodes at both ends of the fault section; calculate the sum of the time it takes for the traveling wave signal to reach the detection nodes at both ends of the fault section to obtain the total time; take the detection node that first collects the traveling wave signal as the nearest node, and divide the time it takes for the traveling wave signal to reach the nearest node by the total time to obtain the distance ratio; multiply the distance between the detection nodes at both ends by the distance ratio to obtain the distance between the fault point and the nearest node; obtain the location of the fault point based on the distance between the fault point and the nearest node.

4. The long-distance collector line fault online positioning system according to claim 3 is characterized in that: The adaptive filtering technology adopts LMS algorithm.

5. The long-distance collector line fault online positioning system according to claim 4 is characterized in that: The specific operations for processing the original traveling wave signal to obtain the advanced traveling wave signal are as follows: A1: Set the order M of the LMS filter, initialize the filter coefficient to zero, and set the learning rate; A2: For any detection node, the original traveling wave signal obtained by the detection node and the original traveling wave signals obtained by the first M-1 detection nodes are used as the input of the LMS filter, and the output value of the LMS filter is compared with the preset expected value to obtain the error value; the filter coefficient of the LMS filter is updated based on the error value; A3: Repeat the above operation until the preset maximum number of iterations is reached; for any detection node, the original traveling wave signal obtained by the detection node and the original traveling wave signals obtained by the first M-1 detection nodes are again used as the input of the LMS filter to obtain the high-level traveling wave signal.

6. A long-distance collector line fault online location system according to claim 5, characterized in that: The fault classification model is established based on the CNN model, including input layer, convolution layer, pooling layer, fully connected layer and output layer; the input layer is used to input high-level traveling wave signals; the convolution layer is used to extract the features of high-level traveling wave signals; the pooling layer is used to reduce the spatial dimension of the features; the fully connected layer is used to further learn the feature representation; the output layer is used to output the fault type information.

7. The long-distance collector line fault online positioning system according to claim 6 is characterized in that: For the training of fault classification model, the specific operations are as follows: Acquire a number of fault classification training samples, each of which contains a high-level traveling wave signal and corresponding fault type information; divide all acquired fault classification training samples into a fault classification training set and a fault classification test set; Use the fault classification training set to train the fault classification model, then input the fault classification test set into the fault classification model to obtain the test results; Set the training conditions. If the test results meet the training conditions, output the trained fault classification model; otherwise, continue to use the fault classification training set to train the fault classification model.

8. The long-distance collector line fault online positioning system according to claim 7 is characterized in that: Specific operations for dividing all line sections into several high-risk sections and low-risk sections: Obtain historical fault occurrence data, and calculate the historical fault occurrence frequency of each line section based on the historical fault occurrence data; Set a risk assessment threshold and classify the line sections where the historical fault frequency is higher than the risk assessment threshold as high-risk sections; Line sections where the historical fault frequency is lower than the risk assessment threshold are classified as low-risk sections.

Images