Flexible overhead line system insulated anchor section shunting method and system based on multi-source data fusion
Through multi-source data fusion and deep learning algorithms, intelligent monitoring and diversion of the insulated anchor segment of the flexible contact network is solved, and the problem of traditional methods relying on a single data source is improved, fault detection accuracy and system reliability are improved, and electrified railway system is ensured. Run safely.
Patent Information
- Application Number
- CN202411573263.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-06
- Publication Date
- 2025-01-03
AI Technical Summary
Traditional contact network monitoring methods rely on a single data source, which is difficult to fully reflect the real-time status of the flexible contact network, and lacks an intelligent diversion strategy, which affects the safety and reliability of train operations.
Using a multi-source data fusion method, the deep learning algorithm model combines sensors, monitoring systems and meteorological data to judge the safety status of the insulated anchor section of the flexible contact network in real time, and perform intelligent shunt operations when unsafe, and realize it through the PLC controller. Shorting and disconnecting of contact lines, construct or unloading the flow bypass, and continuously monitor and alarm.
Achieve comprehensive and accurate monitoring of the insulated anchor section of the flexible contact network, improve the accuracy and speed of fault detection, avoid the risk of fault expansion and damage to the power supply system, improve the reliability and stability of the system, and make the electrified railway system Provides safe and efficient operation guarantee.
Smart Images
Figure CN119239407A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent control of electrified railway systems, and in particular to a flexible contact network insulated anchor section current shunting method and system based on multi-source data fusion. Background Art
[0002] In the electrified railway system, the flexible contact network is an important device for trains to obtain electricity, and the safety status of its insulation anchor section is directly related to the safety and reliability of train operation. However, traditional contact network monitoring methods often rely on a single data source, which makes it difficult to fully reflect the real-time status of the contact network and lack intelligent diversion strategies. Therefore, how to achieve comprehensive monitoring and intelligent diversion of the flexible contact network insulation anchor section has become a technical problem that needs to be solved urgently in the current electrified railway field.
[0003] In recent years, with the rapid development of the Internet of Things, big data and artificial intelligence technologies, multi-source data fusion and deep learning algorithms have been widely used in various fields. Introducing these technologies into the electrified railway system can achieve comprehensive monitoring and intelligent analysis of the insulated anchor section of the flexible contact network, and improve the accuracy and timeliness of fault detection. Summary of the invention
[0004] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a flexible contact network insulating anchor section diversion method and system based on multi-source data fusion, so as to solve the problem of lack of intelligent diversion strategy for flexible contact network insulating anchor sections in the prior art.
[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions: In a first aspect, a flexible contact network insulation anchor section current diversion method based on multi-source data fusion comprises: Step S100: collecting real-time data related to the insulating anchor section of the flexible contact network, and performing data preprocessing on the data related to the insulating anchor section of the flexible contact network; Step S200: when the train stops and is in the insulating anchor section area, using the pre-processed real-time flexible contact network insulating anchor section related data through a deep learning-based algorithm model to determine whether the flexible contact network insulating anchor section is in a safe state; Step S300: If it is in a safe state, the process is terminated; if it is not in a safe state, a shunt operation is performed, a shunt operation instruction is issued through the PLC controller, the main control contactor is closed, the contact wires on both sides of the insulated anchor section of the flexible contact network are short-circuited, and a through-current bypass is constructed; Step S400: After the shunting operation is completed, the flexible contact network insulation anchor section is continuously monitored at preset time intervals to see whether it is in a safe state, the train status information is updated in real time, and an alarm mechanism is triggered when an abnormality is found; Step S500: When the train leaves the insulated anchor section area, the PLC controller sends a termination shunt operation instruction to disconnect the main control contactor and terminate the shunt operation.
[0006] Preferably, in a possible implementation of the first aspect, the data related to the flexible contact network insulation anchor section includes sensor data, monitoring system data and meteorological data; the sensor data includes flexible contact network temperature data, flexible contact network voltage data and flexible contact network current data, the monitoring system data includes train running position data, speed data and power supply voltage data, and the meteorological data includes temperature data and humidity data.
[0007] Preferably, in a possible implementation of the first aspect, the data preprocessing includes data cleaning, data conversion and data normalization; the data cleaning is used to remove outliers and missing values, the data conversion is used to unify data in different formats into a format recognizable by an algorithm model; the data normalization is used to convert data of different dimensions to the same order of magnitude.
[0008] Preferably, in a possible implementation manner of the first aspect, the algorithm model structure based on deep learning includes an input layer, a feature extraction layer, a latent space mapping layer, a feature fusion layer, and an output layer; The input layer is responsible for inputting the pre-processed data related to the insulated anchor section of the flexible contact network into the model; The feature extraction layer uses a graph neural network to capture the spatial relationship between the flexible contact network sensor and the train, extracts the feature representation of nodes and graphs, uses a long short-term memory network to capture the temporal feature representation of the monitoring system data, and uses a convolutional neural network to extract the feature representation of meteorological data; The latent space mapping layer uses the encoder part of the variational autoencoder to map the features output by the feature extraction layer to a low-dimensional latent space to capture the potential distribution of the data; The feature fusion layer fuses the feature vectors from different data sources to form a comprehensive feature vector; The output layer uses a fully connected neural network to judge the safety status of the train based on the comprehensive feature vector output by the feature fusion layer.
[0009] Preferably, in a possible implementation manner of the first aspect, the shunt operation is performed by issuing an instruction through a PLC controller to close the main control contactor to short-circuit the contact wires on both sides of the insulating anchor section of the flexible contact network, electrically connecting the working branch on one side of the segment with the non-working branch on the other side of the segment, forming a current bypass, and changing the electrical state of the insulating anchor section.
[0010] Preferably, in a possible implementation of the first aspect, after the flow bypass is established, the flexible contact network current data is used to determine whether the current is transferred to the short-circuit bypass. If the current cannot be effectively transferred to the short-circuit bypass, a secondary instruction is immediately sent to the PLC controller to try to close the main control contactor again, and at the same time, the early warning mechanism is activated.
[0011] Preferably, in a possible implementation manner of the first aspect, the continuously monitoring whether the flexible contact network insulating anchor section is in a safe state at a preset time interval specifically includes: Acquire real-time flexible contact network insulation anchor section related data at preset time intervals; Input the real-time flexible contact network insulation anchor section related data into the deep learning-based algorithm model; Whether the insulated anchor section of the flexible contact network is in a safe state is determined based on the model output results.
[0012] Preferably, in a possible implementation of the first aspect, the alarm mechanism includes starting an audible and visual alarm device, and simultaneously sending a detailed alarm signal to a remote monitoring center through a wireless communication module, including the fault type, fault location, and current flexible contact network insulation anchor section related data.
[0013] In a second aspect, the present invention provides a flexible contact network insulation anchor section shunt system based on multi-source data fusion, comprising: Data acquisition module, used to collect sensor data, monitoring system data and meteorological data related to the flexible contact network insulation anchor section; Data preprocessing module, used to clean, convert and normalize the collected data; A deep learning algorithm module is used to receive preprocessed data and determine whether the flexible contact network insulation anchor section is in a safe state; A programmable logic controller module is used to control the closing and opening of the main control contactor when the flexible contact network is in an unsafe state, to achieve the short-circuiting and disconnection of the contact wires on both sides of the insulating anchor section of the flexible contact network, and to build or release the through-current bypass; The monitoring and alarm module is used to continuously monitor the safety status of the flexible contact network at preset time intervals after the diversion operation is completed, and trigger the alarm mechanism when an abnormality is found. The alarm mechanism includes starting the sound and light alarm device and sending a detailed alarm signal to the remote monitoring center through the wireless communication module.
[0014] The beneficial effects of the present invention are as follows: by collecting and integrating multiple data sources of the insulated anchor section of the flexible contact network, the real-time status of the contact network can be fully and accurately reflected, and the accuracy and speed of fault detection can be improved. At the same time, by using the deep learning algorithm model for intelligent analysis, the system can judge the safety status of the flexible contact network, and implement intelligent shunting operations through the PLC controller when necessary, effectively avoiding the risk of fault expansion and damage to the power supply system. In addition, the system can also continuously monitor the safety status of the flexible contact network, and trigger the alarm mechanism when an abnormality is found, and send a detailed alarm signal to the remote monitoring center, providing timely and accurate information support for operation and maintenance personnel, and further improving the reliability and stability of the system. In summary, the present invention provides a strong technical guarantee for the safe and efficient operation of the electrified railway system. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] 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, other drawings can be obtained based on these drawings without paying creative work.
[0016] Figure 1 A flow chart of a method for current diversion of insulated anchor sections of flexible contact networks based on multi-source data fusion is provided for this application.
[0017] Figure 2 A schematic diagram of the system structure of flexible contact network insulated anchor section current diversion based on multi-source data fusion is provided for this application.
[0018] Explanation of the accompanying drawings: 1-data acquisition module, 2-data preprocessing module, 3-deep learning algorithm module, 4-programmable logic controller module, 5-monitoring and alarm module. DETAILED DESCRIPTION
[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0020] Example 1: Figure 1 As shown, the present invention provides a flexible contact network insulation anchor section current diversion method based on multi-source data fusion, comprising: Step S100: collecting real-time data related to the insulating anchor section of the flexible contact network, and performing data preprocessing on the data related to the insulating anchor section of the flexible contact network.
[0021] Specifically, the data related to the insulated anchor section of the flexible contact network include sensor data, monitoring system data and meteorological data; the sensor data include flexible contact network temperature data, flexible contact network voltage data and flexible contact network current data, the monitoring system data include train running position data, speed data and power supply voltage data, and the meteorological data include temperature data and humidity data.
[0022] Temperature sensors are installed in the flexible contact network insulation anchor section area to monitor temperature changes in the flexible contact network insulation anchor section area; voltage sensors are installed on the power supply lines in the flexible contact network insulation anchor section area to monitor voltage fluctuations in the flexible contact network insulation anchor section area; current sensors are installed in the power supply circuit of the flexible contact network insulation anchor section area to monitor the current size and direction in the flexible contact network insulation anchor section area. Temperature sensors and humidity sensors are installed near the flexible contact network insulation anchor section area to obtain meteorological data.
[0023] The sensor data and meteorological data are sent to the train control system through wireless transmission, and the monitoring system data, namely the train running position data, speed data and power supply voltage data, are sent to the train control system through the train line.
[0024] After that, the relevant data of the insulated anchor section of the flexible contact network are preprocessed, and the sensor data, monitoring system data and meteorological data are stored as csv files, named sensor_data.csv, monitor_data.csv and weather_data.csv respectively.
[0025] Then the data was cleaned. First, the missing values in the data were removed by using the dropna() function to ensure the integrity of the data. Then, outliers were removed by conditional screening. In terms of data conversion, the timestamp column was converted to datetime format, and the numerical data was standardized. The StandardScaler() function was used to standardize the numerical columns of flexible contact network temperature, voltage, current, train speed, power supply voltage, ambient temperature and humidity, eliminating the impact of different dimensions on the data.
[0026] Step S200: When the train stops and is in the insulating anchor section area, the pre-processed real-time flexible contact network insulating anchor section related data is used to determine whether the flexible contact network insulating anchor section is in a safe state through an algorithm model based on deep learning.
[0027] Specifically, when the train speed data is zero and the train is judged to be in the insulating anchor section area based on the train running position data, the pre-processed flexible contact network insulating anchor section related data is input into the algorithm model based on deep learning.
[0028] The algorithm model based on deep learning includes input layer, feature extraction layer, latent space mapping layer, feature fusion layer and output layer.
[0029] The input layer is responsible for receiving and processing input data. In this embodiment, the input data includes a graph representation of flexible contact network sensor data (x_graph), a time series representation of monitoring system data (x_time_series), and a local feature representation of meteorological data (x_weather).
[0030] The feature extraction layer consists of three parts: graph neural network (GNN), long short-term memory network (LSTM) and convolutional neural network (CNN).
[0031] Graph neural network, uses the GCNConv layer to perform convolution operations on graph data and extract feature representations of nodes and graphs. These features represent the spatial relationship between flexible contact network sensors.
[0032] The long short-term memory network processes the time series data and captures the time series dependency and long-term memory of the monitoring system data. The LSTM layer receives the time series data as input and outputs the hidden state of each time step. In this embodiment, the last hidden state of the LSTM is taken as the time series feature.
[0033] Convolutional neural network,convolution operation on meteorological data to extract local features.,CNN layer receives meteorological data and outputs convolutional feature maps.,These feature maps are then flattened for splicing with other features.
[0034] The latent space mapping layer uses the encoder part of the variational autoencoder to map the features output by the feature extraction layer to a low-dimensional latent space. This is achieved through two fully connected layers: one for calculating the mean (mu) and the other for calculating the logarithmic variance (logvar). Then, the latent variable (z) is sampled from the latent distribution using the reparameterization technique, which captures the latent distribution of the data.
[0035] Feature fusion layer, the feature fusion layer fuses feature vectors from different data sources. In this model, the features extracted by the graph neural network, long short-term memory network and convolutional neural network are spliced together to form a comprehensive feature vector. Then, this comprehensive feature vector is transformed through a fully connected layer to obtain the fused feature representation.
[0036] The output layer uses a fully connected neural network to judge the safety status of the train based on the comprehensive feature vector output by the feature fusion layer. It receives the fused features as input and outputs a binary classification result through a fully connected layer.
[0037] The preprocessed data related to the insulated anchor section of the flexible contact network is passed through an algorithm model based on deep learning to finally output whether the current insulated anchor section of the flexible contact network is in a safe state.
[0038] The current flexible contact network insulation anchor section is in a safe state, which refers to an overall evaluation of the working state of the current flexible contact network insulation anchor section. When the model outputs the conclusion that the current flexible contact network insulation anchor section is in a safe state, it means that under the current conditions, the working state of the flexible contact network insulation anchor section area meets the safety standards, and no abnormalities or faults that may cause danger are detected.
[0039] Step S300: If it is in a safe state, the process is terminated. If it is not in a safe state, a shunt operation is performed, and a shunt operation instruction is issued through the PLC controller to close the main control contactor to achieve short-circuiting of the contact wires on both sides of the insulated anchor section of the flexible contact network and construct a current bypass.
[0040] Specifically, the main function of the main control contactor is to short-circuit the contact wires on both sides of the insulated anchor section of the flexible contact network when the train stays in the designated insulated anchor section area and the flexible contact network is in poor condition, turning the originally isolated anchor section into a non-insulated state to form a through-current bypass. The purpose of this design is to prevent excessive current generated by the carbon slide contacting the working branch and the non-working branch at the same time, and to avoid the current passing through the "working branch-carbon slide-non-working branch" path, causing the wire to be partially overheated and melted.
[0041] When the train control system confirms that the train has stopped in the insulating anchor section area and the algorithm model based on deep learning determines that the flexible contact network insulating anchor section is in an unsafe state, the PLC issues a shunt operation instruction and controls the main control contactor to close to short-circuit the contact wires on both sides of the flexible contact network insulating anchor section. The electrical connection between the working branch and the non-working branch is completed, forming a through-current bypass, thereby changing the electrical characteristics of the insulating anchor section.
[0042] After the bypass is successfully established, the system immediately monitors the current data in real time through the current sensors deployed on the flexible contact network. These data are transmitted to the train control system in real time, and the system conducts a comprehensive evaluation based on the preset current threshold and change pattern to determine whether the current has been effectively transferred to the newly built short-circuit bypass.
[0043] If the current data does not show an obvious trend of transferring to the short-circuit bypass within the specified time window, that is, the current still mainly flows through the original working branch, the system determines that the current transfer has failed. At this time, the train control system responds and sends a secondary instruction to the PLC to try to reclose the main control contactor. At the same time, the early warning mechanism is activated, including triggering an audible and visual alarm in the train control room, and sending a detailed alarm signal to the remote monitoring center to report the current transfer failure and the countermeasures that have been taken. After receiving the alarm signal, the remote monitoring center will quickly initiate the emergency response procedure to ensure the operation of the train and the safety of passengers.
[0044] Step S400: After the shunting operation is completed, the flexible contact network insulation anchor section is continuously monitored at preset time intervals to see whether it is in a safe state, the train status information is updated in real time, and an alarm mechanism is triggered when an abnormality is found.
[0045] Specifically, after the shunting operation is completed, the system automatically collects and processes the relevant data of the flexible contact network insulation anchor section every minute. These data include the temperature, voltage, and current data of the flexible contact network, as well as the monitoring system data such as the train running position, speed, and power supply voltage, and meteorological data such as ambient temperature and humidity. After receiving these data, the train control system first performs preprocessing, including data cleaning, format conversion, and standardization steps to ensure the accuracy and consistency of the data.
[0046] Subsequently, the system uses a deep learning-based algorithm model to analyze these pre-processed data in real time to determine whether the flexible contact network insulation anchor section is still in a safe state. During the continuous monitoring process, the system updates the train status information in real time, including the train's location, speed, power supply status, etc., so that the train driver and the remote monitoring center can understand the train's operating status at any time.
[0047] If the system finds that the flexible contact network insulation anchor section is abnormal or in an unsafe state during the analysis process, the alarm mechanism will be triggered immediately. This mechanism includes triggering the sound and light alarm device in the train control room to attract the driver's attention, and sending a detailed alarm signal to the remote monitoring center at the same time. The alarm signal will contain detailed information on the abnormal data, the time and location of the abnormality, and the response measures taken, so that the remote monitoring center can respond quickly and take appropriate measures.
[0048] Step S500: When the train leaves the insulated anchor section area, the PLC controller sends a termination shunt operation instruction to disconnect the main control contactor and terminate the shunt operation.
[0049] Specifically, when the train has completely left the insulated anchor section area according to the position data provided by the train monitoring system, the PLC controller sends a termination shunt operation instruction to disconnect the main control contactor. After receiving this instruction, the PLC controller will immediately execute the corresponding control program to release the short-circuit state of the contact wires on both sides of the insulated anchor section of the flexible contact network, that is, terminate the shunt operation.
[0050] After disconnecting the main control contactor, the system will perform subsequent operations to ensure safety. This includes reconfirming whether the current has been completely transferred from the short-circuit bypass and whether the flexible contact network insulation anchor section has returned to its normal electrical isolation state. In addition, the system will also record and store relevant information about this operation, including the time when the main control contactor was disconnected and the exact time when the train left the insulation anchor section area.
[0051] Example 2: Figure 2 As shown, the present invention provides a flexible contact network insulation anchor section shunt system based on multi-source data fusion, comprising: Data acquisition module 1, used to collect sensor data, monitoring system data and meteorological data related to the flexible contact network insulation anchor section; Data preprocessing module 2, used for cleaning, converting and normalizing the collected data; The deep learning algorithm module 3 is used to receive the preprocessed data and determine whether the flexible contact network insulation anchor section is in a safe state; The programmable logic controller module 4 is used to control the closing and opening of the main control contactor when the flexible contact network is in an unsafe state, to achieve the short-circuiting and disconnection of the contact wires on both sides of the insulating anchor section of the flexible contact network, and to build or release the through-current bypass; The monitoring and alarm module 5 is used to continuously monitor the safety status of the flexible contact network at preset time intervals after the diversion operation is completed, and trigger the alarm mechanism when an abnormality is found. The alarm mechanism includes starting the sound and light alarm device and sending a detailed alarm signal to the remote monitoring center through the wireless communication module.
[0052] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A flexible contact network insulation anchor section current diversion method based on multi-source data fusion, characterized in that: include: Step S100: collecting real-time data related to the insulating anchor section of the flexible contact network, and performing data preprocessing on the data related to the insulating anchor section of the flexible contact network; Step S200: when the train stops and is in the insulating anchor section area, using the pre-processed real-time flexible contact network insulating anchor section related data through a deep learning-based algorithm model to determine whether the flexible contact network insulating anchor section is in a safe state; Step S300: If it is in a safe state, the process is terminated; if it is not in a safe state, a shunt operation is performed, a shunt operation instruction is issued through the PLC controller, the main control contactor is closed, the contact wires on both sides of the insulated anchor section of the flexible contact network are short-circuited, and a through-current bypass is constructed; Step S400: After the shunting operation is completed, the flexible contact network insulation anchor section is continuously monitored at preset time intervals to see whether it is in a safe state, the train status information is updated in real time, and an alarm mechanism is triggered when an abnormality is found; Step S500: When the train leaves the insulated anchor section area, the PLC controller sends a termination shunt operation instruction to disconnect the main control contactor and terminate the shunt operation.
2. The flexible contact network insulation anchor section current diversion method according to claim 1, characterized in that: The data related to the flexible contact network insulation anchor section include sensor data, monitoring system data and meteorological data; the sensor data include flexible contact network temperature data, flexible contact network voltage data and flexible contact network current data, the monitoring system data include train running position data, speed data and power supply voltage data, and the meteorological data include temperature data and humidity data.
3. The flexible contact network insulation anchor section current diversion method according to claim 1, characterized in that: The data preprocessing includes data cleaning, data conversion and data normalization; the data cleaning is used to remove outliers and missing values, the data conversion is used to unify data in different formats into a format recognizable by the algorithm model; the data normalization is used to convert data of different dimensions to the same order of magnitude.
4. The flexible contact network insulation anchor section current diversion method according to claim 1, characterized in that: The algorithm model structure based on deep learning includes an input layer, a feature extraction layer, a latent space mapping layer, a feature fusion layer, and an output layer; The input layer is responsible for inputting the pre-processed data related to the insulated anchor section of the flexible contact network into the model; The feature extraction layer uses a graph neural network to capture the spatial relationship between the flexible contact network sensor and the train, extracts the feature representation of nodes and graphs, uses a long short-term memory network to capture the temporal feature representation of the monitoring system data, and uses a convolutional neural network to extract the feature representation of meteorological data; The latent space mapping layer uses the encoder part of the variational autoencoder to map the features output by the feature extraction layer to a low-dimensional latent space to capture the potential distribution of the data; The feature fusion layer fuses the feature vectors from different data sources to form a comprehensive feature vector; The output layer uses a fully connected neural network to judge the safety status of the train based on the comprehensive feature vector output by the feature fusion layer.
5. The flexible contact network insulation anchor section current diversion method according to claim 1, characterized in that: The shunt operation is carried out by issuing instructions through the PLC controller to close the main control contactor to short-circuit the contact wires on both sides of the insulating anchor section of the flexible contact network, electrically connect the working branch on one side of the section with the non-working branch on the other side of the section, form a current bypass, and change the electrical state of the insulating anchor section.
6. The flexible contact network insulation anchor section current diversion method according to claim 5, characterized in that: After the flow bypass is established, the flexible contact network current data is used to determine whether the current is transferred to the short-circuit bypass. If the current cannot be effectively transferred to the short-circuit bypass, a secondary instruction is immediately sent to the PLC controller to try to close the main control contactor again and start the early warning mechanism at the same time.
7. The flexible contact network insulation anchor section current diversion method according to claim 1, characterized in that: The method of continuously monitoring whether the flexible contact network insulating anchor section is in a safe state according to a preset time interval specifically includes: Acquire real-time flexible contact network insulation anchor section related data at preset time intervals; Input the real-time flexible contact network insulation anchor section related data into the deep learning-based algorithm model; Whether the insulated anchor section of the flexible contact network is in a safe state is determined based on the model output results.
8. The flexible contact network insulation anchor section current diversion method according to claim 1, characterized in that: The alarm mechanism includes starting the sound and light alarm device, and sending a detailed alarm signal to the remote monitoring center through the wireless communication module, including the fault type, fault location, and current flexible contact network insulation anchor section related data.
9. The flexible contact network insulation anchor section shunt system based on multi-source data fusion is characterized by: include: Data acquisition module, used to collect sensor data, monitoring system data and meteorological data related to the flexible contact network insulation anchor section; Data preprocessing module, used to clean, convert and normalize the collected data; A deep learning algorithm module is used to receive preprocessed data and determine whether the flexible contact network insulation anchor section is in a safe state; A programmable logic controller module is used to control the closing and opening of the main control contactor when the flexible contact network is in an unsafe state, to achieve the short-circuiting and disconnection of the contact wires on both sides of the insulating anchor section of the flexible contact network, and to build or release the through-current bypass; The monitoring and alarm module is used to continuously monitor the safety status of the flexible contact network at preset time intervals after the diversion operation is completed, and trigger the alarm mechanism when an abnormality is found. The alarm mechanism includes starting the sound and light alarm device and sending a detailed alarm signal to the remote monitoring center through the wireless communication module.