Online monitoring and early warning method and system for tension of carrier cable of railway contact network

Abstract

The invention belongs to the technical field of railway transportation safety, provides a railway contact net carrier cable tension online monitoring and early warning method and system, and aims at solving the technical problems that in the prior art, capturing is not comprehensive, and different environments are difficult to adapt. Acquiring a tension data stream by using a context awareness technique; analyzing the tension data flow by using a hidden Markov model to obtain a preliminary analysis result, processing by using a neural network, generating an abnormal mode recognition result, performing risk assessment by using a Bayesian network algorithm and a digital twin technology to obtain a preliminary fault estimation probability, and predicting a target fault estimation probability in combination with multi-context sensing data; a self-adaptive control algorithm is used for generating an adjusted alarm threshold value, when the monitored tension value exceeds the adjusted alarm threshold value, early warning information is generated in time, and processing is carried out through the block chain technology to obtain target early warning information; according to the technical scheme provided by the invention, the monitoring and early warning capabilities of the railway overhead line system carrier cable are comprehensively improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of railway transportation safety, and in particular to an online monitoring and early warning method and system for the tension of railway overhead contact cables. Background Art

[0002] The railway overhead line is an important facility for electric locomotives, electric multiple units and other electrified railway vehicles to obtain electricity. The state of the catenary cable, as a key component that bears tension in the overhead line system, is directly related to the stability and safety of the entire power supply system. Under high-speed trains and severe weather conditions, changes in the tension of the catenary cable may affect the height and stability of the contact line, and then affect the good contact between the pantograph and the contact line, resulting in power transmission interruption or accidents. Therefore, real-time online monitoring of the tension of the catenary cable and the establishment of an early warning mechanism are of great significance to ensure the safety of railway transportation.

[0003] At present, the online monitoring and early warning methods of railway contact network cables mainly rely on traditional sensor technology and simple data analysis methods. Sensors installed on the contact network collect tension data in real time, and simple statistical analysis is performed based on historical data to identify obvious abnormal patterns and issue alarms. However, these existing methods have many shortcomings. First, single-factor monitoring cannot fully capture the tension changes under the combined effect of multiple factors; second, fixed alarm thresholds are difficult to adapt to complex and changing environmental conditions, which can easily lead to false alarms or missed alarms; third, there is a lack of intelligent processing. The existing data analysis methods are relatively simple and it is difficult to achieve accurate risk assessment and fault prediction. Summary of the invention

[0004] The embodiments of the present invention provide a method and system for online monitoring and early warning of the tension of the catenary cables of a railway contact network, so as to solve the problems in the prior art that the tension changes under the combined effect of multiple factors cannot be fully captured, the tension changes are difficult to adapt to complex and changeable environmental conditions, false alarms or missed alarms are easily caused, and accurate risk assessment and fault prediction are difficult to achieve.

[0005] In a first aspect, an embodiment of the present invention provides a method for online monitoring and early warning of the tension of a railway overhead contact network catenary cable, comprising: Based on the factors affecting train operation, the dynamic tension data of the railway contact network bearing cables under different environmental conditions are collected using situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; Applying a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and using a convolutional neural network and a graph neural network to process the preliminary analysis result to generate an abnormal pattern recognition result; Based on the abnormal pattern recognition result, a risk assessment is performed using a Bayesian network algorithm and digital twin technology to obtain a preliminary estimated fault probability, and a target estimated fault probability of a fault occurring within a specific time period is predicted by combining multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information; Based on the target fault prediction probability and the multi-scenario perception data, an adaptive control algorithm is used to dynamically adjust the preset alarm threshold to obtain an adjusted alarm threshold. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, a warning message is immediately generated, and the warning message is processed through blockchain technology to obtain target warning information.

[0006] Optionally, based on the abnormal pattern recognition result, a Bayesian network algorithm and digital twin technology are used to perform risk assessment to obtain a preliminary estimated fault probability, and the target estimated fault probability of a fault occurring within a specific time period is predicted by combining the multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information, including: Based on the abnormal pattern recognition results, the Bayesian network is used to analyze the dependencies between the components of the railway contact network, and the conditional probability distribution results between the components are obtained; Using digital twin technology to build a target virtual copy, using the target virtual copy to simulate the behavior of the railway contact network under different assumed working scenarios, and generating physical simulation results; Combining the conditional probability distribution results between the components and the physical simulation results, comprehensively evaluating the state of the railway contact network to obtain a preliminary fault prediction probability; Based on the preliminary fault prediction probability and combined with multi-context perception data, a graph neural network is used to analyze the correlation between various components in the railway contact network structure to obtain a high-dimensional feature vector, and a long short-term memory network is used to process the high-dimensional feature vector to obtain the key time characteristics of the tension change; Based on the critical time characteristics, the target estimated probability of failure occurring within a specific time period is predicted, and Bayesian optimization and Monte Carlo simulation are introduced in the prediction process to quantify the uncertainty in the prediction for risk assessment.

[0007] Optionally, a target virtual replica is constructed using digital twin technology, and the target virtual replica is used to simulate the behavior of the railway contact network under different assumed working scenarios to generate physical simulation results, including: Based on the structure and parameters of the railway contact network, an initial virtual copy corresponding to the railway contact network is constructed using digital twin technology; Based on the real-time status data of the railway contact network, synchronously updating the internal status of the initial virtual copy to obtain an updated virtual copy, so that the internal status of the updated virtual copy reflects the status of the railway contact network catenary cables; Based on real-time operation and monitoring data, the internal parameters of the updated virtual copy are updated to generate a target virtual copy to adapt to changes in factors affecting train operation; Based on the target virtual copy, multi-scenario simulations are performed under different assumed working scenarios to evaluate the different assumed working scenarios and the impact of the assumed working scenarios on the railway contact network load-bearing cables, and obtain physical simulation results.

[0008] Optionally, based on the target virtual copy, multi-scenario simulations are performed under different assumed working scenarios to evaluate the different assumed working scenarios and the impact of the assumed working scenarios on the railway contact network catenary cables, and physical simulation results are obtained, including: Based on the train operation influencing factors, different assumed working scenarios are set to evaluate the influence of the train operation influencing factors on the railway contact network catenary cables; Based on the assumed working scenario, simulation environment parameters are set, and based on the real-time operation and monitoring data and the historical operation and environment data, the simulation environment parameters are optimized to obtain optimized simulation environment parameters; Based on the optimized simulation environment parameters, the target virtual copy is subjected to multi-scenario simulation processing. During the multi-scenario simulation processing, the dynamic environment and operating condition changes reflecting different assumed working scenarios are changed to obtain target state information of the railway contact network catenary under different dynamic environment and operating condition changes. Based on the target state information, the behavior response of the catenary under different assumed working scenarios is obtained; Based on the key indicators of the state of the catenary cable, the influencing factors affecting the behavioral response of the catenary cable are identified, and the abnormal phenomena are recorded at the same time. Based on the influencing factors and the abnormal phenomena, the performance of the catenary cable is evaluated to obtain the performance impact results; A physical simulation result is generated based on the behavior response of the catenary, the influencing factors, the abnormal phenomenon and the performance impact result.

[0009] Optionally, based on the preliminary fault prediction probability and in combination with multi-context perception data, a graph neural network is applied to analyze the correlation between various components in the railway contact network structure to obtain a high-dimensional feature vector, and a long short-term memory network is used to process the high-dimensional feature vector to obtain key time characteristics of tension changes, including: Based on the preliminary fault prediction probability and multi-scenario perception data, the status of each component in the railway contact network is analyzed to obtain component status information; Based on the component status information, nodes and edges are defined to construct a graph structure, wherein the nodes represent the components of the railway contact network, and the edges represent the physical connections between the components; Based on the graph structure, a graph neural network is used to analyze the correlation between various components in the railway contact network, and a high-dimensional feature vector containing the correlation between various components is obtained, and the high-dimensional feature vector reflects the state of the railway contact network load-bearing cables; The high-dimensional feature vector is processed by using an internal memory unit of a long short-term memory network to obtain key time characteristics of the high-dimensional feature vector, wherein the key time characteristics include tension change trend and periodic fluctuation.

[0010] Optionally, based on the target fault prediction probability and the multi-scenario perception data, an adaptive control algorithm is used to dynamically adjust the preset alarm threshold to obtain an adjusted alarm threshold. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, an early warning message is immediately generated, and the early warning message is processed by blockchain technology to obtain target early warning information, including: Based on the fault prediction probability and the multi-scenario perception data, using an adaptive control algorithm and a reinforcement learning algorithm, dynamically adjusting a preset alarm threshold to obtain an adjusted alarm threshold; Based on the tension data stream, the real-time tension value of the railway contact network load-bearing cable is monitored. When the real-time tension value exceeds the adjusted alarm threshold, an abnormal data point of the tension data stream is identified by using an abnormality detection algorithm, and an early warning message is generated in real time by combining the abnormal data point and the event-driven architecture; The warning information is encrypted using homomorphic encryption technology to obtain encrypted warning information, and blockchain technology is introduced to perform tamper-proof processing on the encrypted warning information through the consensus mechanism and timestamp function of the blockchain to obtain target warning information.

[0011] Optionally, based on the fault prediction probability and the multi-scenario perception data, an adaptive control algorithm and a reinforcement learning algorithm are used to dynamically adjust a preset alarm threshold to obtain an adjusted alarm threshold, including: Based on the target fault prediction probability and the multi-scenario perception data, a Bayesian optimization algorithm is used to construct a comprehensive risk assessment model, and the risk assessment model is used to iteratively update the prior distribution to obtain an updated prior distribution, while maximizing the posterior probability. Based on the updated prior distribution and the maximized posterior probability, the risk level of the railway contact network load-bearing cables in a real-time environment is evaluated to generate a real-time risk assessment result; According to the real-time risk assessment results, the preset alarm threshold is dynamically adjusted by applying an adaptive control algorithm to generate a preliminary adjusted alarm threshold, and a fuzzy logic controller is introduced to handle uncertainty and nonlinear relationships in the process of applying the adaptive control algorithm for dynamic adjustment to optimize the adjustment process; A reinforcement learning algorithm is introduced to evaluate the warning effects under different alarm thresholds by defining a reward function. The preliminary adjusted alarm threshold is optimized based on the warning effect to obtain the adjusted alarm threshold. In the process of introducing the reinforcement learning algorithm for processing, the double Q learning algorithm and the active learning mechanism are combined to adjust the configuration of the preliminary adjusted alarm threshold according to the actual warning effect and user feedback, so as to search for the optimal alarm threshold as the adjusted alarm threshold.

[0012] In a second aspect, an embodiment of the present invention provides an online monitoring and early warning system for the tension of a railway overhead contact network catenary cable, comprising: A collection module, which is used to collect dynamic tension data of railway contact network bearing cables under different environmental conditions based on factors affecting train operation by using situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; An analysis module, used for applying a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and using a convolutional neural network and a graph neural network to process the preliminary analysis result to generate an abnormal pattern recognition result; A prediction module is used to perform risk assessment based on the abnormal pattern recognition result using a Bayesian network algorithm and digital twin technology to obtain a preliminary estimated fault probability, and to predict a target estimated fault probability of a fault occurring within a specific time period by combining multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information; A generation module is used to dynamically adjust the preset alarm threshold value based on the target fault estimation probability and the multi-scenario perception data by using an adaptive control algorithm to obtain an adjusted alarm threshold value. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold value, an early warning message is immediately generated, and the early warning message is processed through blockchain technology to obtain target early warning information.

[0013] In a third aspect, an embodiment of the present invention provides a computing device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute a method for online monitoring and early warning of the tension of a railway contact network load-bearing cable as described in any one of the first aspects.

[0014] In a fourth aspect, an embodiment of the present invention provides a computer storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement an online monitoring and early warning method for the tension of a railway contact network load-bearing cable as described in any one of the first aspects.

[0015] In an embodiment of the present invention, based on the factors affecting train operation, the dynamic tension data of the railway contact network bearing cable under different environmental conditions is collected by using the situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; the hidden Markov model is applied to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and the preliminary analysis result is processed by using a convolutional neural network and a graph neural network to generate an abnormal pattern recognition result; based on the abnormal pattern recognition result, a risk assessment is performed using a Bayesian network algorithm and a digital twin technology to obtain a preliminary fault prediction probability, and the target fault prediction probability of a fault occurring within a specific time period is predicted by combining the multi-situational awareness data and the preliminary fault prediction probability, wherein the multi-situational awareness data includes historical maintenance records and weather forecast information; based on the target fault prediction probability and the multi-situational awareness data, an adaptive control algorithm is used to dynamically adjust the preset alarm threshold to obtain an adjusted alarm threshold, and when the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, an early warning information is immediately generated, and the early warning information is processed by blockchain technology to obtain a target early warning information. The technical solution provided by the present invention introduces situational awareness technology, so that the system can collect dynamic tension data of the contact network bearing cables under different environmental conditions in real time and accurately, ensuring the comprehensiveness and real-time nature of the data. Hidden Markov models, convolutional neural networks and graph neural networks are used to intelligently analyze the tension change pattern and detect anomalies, thereby improving the accuracy and reliability of anomaly identification. Combining the Bayesian network algorithm and digital twin technology, multi-situational awareness data are integrated to perform accurate risk assessment and fault prediction. The adaptive control algorithm dynamically adjusts the alarm threshold according to the estimated probability of the target fault to reduce false alarms and missed alarms, while blockchain technology is used to process early warning information to ensure the security and credibility of the early warning information. In summary, the present invention comprehensively improves the monitoring and early warning capabilities of the railway contact network bearing cables. Among them, by setting a variety of hypothetical working scenarios, the system can comprehensively evaluate the impact of train operation factors on the contact network catenary cables, ensuring the comprehensiveness and accuracy of risk assessment; combining real-time operation and monitoring data with historical data, optimizing simulation environment parameters, improving the accuracy and reliability of the simulation, and ensuring that the simulation results are closer to the actual situation; through multi-scenario simulation processing, the system can evaluate the behavioral responses of the catenary cables under different dynamic environments and changes in operating conditions, capture complex real-life situations, and enhance the adaptability and flexibility of the system; based on the key indicators of the catenary cable status, the system can accurately identify the main factors affecting the behavioral responses of the catenary cables, and record abnormal phenomena, providing a scientific basis for subsequent performance evaluation.

[0016] These and other aspects of the present invention will become more apparent from the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are 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.

[0018] Figure 1 A flow chart of a method for online monitoring and early warning of the tension of a railway overhead contact network catenary provided by an embodiment of the present invention; Figure 2 A schematic structural diagram of a railway overhead contact network catenary tension online monitoring and early warning system provided by an embodiment of the present invention; Figure 3 A schematic diagram of the structure of a computing device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the solutions of 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.

[0020] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0021] 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0022] Figure 1 A flowchart of a method for online monitoring and early warning of the tension of a railway overhead contact network catenary cable is provided for an embodiment of the present invention. Figure 1 As shown, the method includes: With the rapid development of railway transportation, ensuring the safe and stable operation of the catenary cables has become the key to ensuring the safety of railway operations. Existing monitoring and early warning methods have deficiencies in multi-factor comprehensive monitoring, intelligent analysis, and adaptive alarm, and are difficult to meet the increasingly complex operational needs. Based on this, the present invention provides a method for online monitoring and early warning of the tension of railway catenary cables, such as Figure 1 ,include: Step 101: Based on factors affecting train operation, dynamic tension data of railway overhead wire under different environmental conditions are collected using situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; In this step, factors affecting train operation refer to factors that affect the tension of the railway contact network load-bearing cables, mainly including weather conditions (such as temperature, humidity, wind speed) and train operation density (such as train frequency, vehicle model, etc.).

[0023] This step is based on the factors affecting train operation and uses situational awareness technology to collect dynamic tension data of the contact network under different environmental conditions in real time to form a tension data stream. For example, data is collected containing a temperature of 25°C, a humidity of 60%, and a wind speed of 10m / s on a certain day. At the same time, there are 3 trains running in this section. The system records the tension value once a second and generates a continuous tension data stream, such as [80N, 79N, 81N,...].

[0024] Step 102: applying a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and using a convolutional neural network and a graph neural network to process the preliminary analysis result to generate an abnormal pattern recognition result; In this step, the tension variation pattern refers to the law or trend of the change of the tension of the catenary cable with factors such as time, environmental conditions and train operation status.

[0025] In this step, the hidden Markov model is used to analyze the tension change pattern in the tension data stream to obtain preliminary analysis results. The hidden Markov model trains multiple hidden states based on historical data, and each state corresponds to a different tension change pattern. By analyzing the current tension data stream, the most likely hidden state sequence is output. The preliminary analysis results are processed using convolutional neural networks and graph neural networks to generate abnormal pattern recognition results. Convolutional neural networks are used to extract local features and identify potential abnormal patterns; graph neural networks capture the correlation between the various parts of the load-bearing cable and further confirm the existence of abnormal patterns. For example, abnormal tension fluctuations in certain time periods are identified.

[0026] Step 103: Based on the abnormal pattern recognition result, a Bayesian network algorithm and digital twin technology are used to perform risk assessment to obtain a preliminary estimated fault probability, and a target estimated fault probability of a fault occurring within a specific time period is predicted by combining the multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information; In this step, multi-context perception data including historical maintenance records, weather forecast information and other data are used to comprehensively evaluate the operating status and risk level of the system.

[0027] This step uses the Bayesian network algorithm and digital twin technology to perform risk assessment based on the abnormal pattern recognition results and obtain a preliminary estimated probability of failure. For example, the Bayesian network combined with digital twin technology calculates the probability of failure of the catenary in the next 24 hours as 0.05 based on the current abnormal pattern and historical data. Combined with multi-context perception data (such as historical maintenance records and weather forecast information), the target estimated probability of failure within a specific time period is predicted. If the weather forecast shows that there will be strong winds in the next few days, and historical maintenance records show that there have been minor problems with the catenary recently, the system will increase the target estimated probability of failure to 0.1.

[0028] Step 104: Based on the target fault prediction probability and the multi-scenario perception data, the preset alarm threshold is dynamically adjusted by using an adaptive control algorithm to obtain an adjusted alarm threshold. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, an early warning message is immediately generated, and the early warning message is processed by using blockchain technology to obtain target early warning information. This step uses an adaptive control algorithm to dynamically adjust the preset alarm threshold based on the target failure prediction probability and multi-scenario perception data to obtain the adjusted alarm threshold. For example, if the target failure prediction probability is high, the system adjusts the alarm threshold from the default 80N to 70N to detect potential problems more sensitively. If the tension value reaches 75N at a certain moment, exceeding the adjusted alarm threshold of 70N, the system immediately generates an early warning message. The early warning information is processed through blockchain technology to obtain the target early warning information to ensure its security and credibility. For example, the early warning information is recorded on the blockchain, including timestamps, locations, tension values ​​and other information to ensure that the information cannot be tampered with and is transparent and traceable.

[0029] The embodiments of the present invention use situational awareness technology and a variety of sensors to collect dynamic tension data of the contact network load-bearing cables in real time and accurately, ensuring the comprehensiveness and real-time nature of the data; using hidden Markov models, convolutional neural networks and graph neural networks, the system can efficiently identify complex tension change patterns and improve the accuracy and reliability of anomaly detection; combined with the Bayesian network algorithm and digital twin technology, it can comprehensively consider the influence of multiple factors, provide more scientific risk assessment and fault prediction, and discover potential problems in advance; through adaptive control algorithms and blockchain technology, the system can dynamically adjust the alarm threshold to reduce false alarms and missed alarms, while ensuring the security and credibility of warning information; through the above-mentioned technical means, the present invention can not only detect potential problems in a timely manner and issue warnings, but also provide scientific decision-making support for maintenance personnel, helping them to take preventive measures in advance to ensure the safety and efficiency of railway operations.

[0030] On the basis of abnormal pattern recognition, in order to further improve the accuracy and reliability of risk assessment, the present invention provides a specific embodiment, based on this, step 103, based on the abnormal pattern recognition result, using the Bayesian network algorithm and the digital twin technology to perform risk assessment, obtain a preliminary fault prediction probability, combine the multi-context perception data and the preliminary fault prediction probability, and predict the target fault prediction probability of a fault occurring within a specific time period, wherein the multi-context perception data includes historical maintenance records and weather forecast information, specifically including the following steps: Step 301: Based on the abnormal pattern recognition result, the dependency relationship between the components of the railway contact network is analyzed using the Bayesian network to obtain the conditional probability distribution results between the components; In this step, the conditional probability distribution result refers to the specific quantification of the dependencies between components in the Bayesian network, which is expressed as the probability distribution of the states of other components under certain given conditions.

[0031] For example, after identifying abnormal fluctuations in the tension of the catenary, the Bayesian network can analyze the dependencies between the catenary and components such as support rods and insulators, and quantify the state probability distribution of these components under different conditions. In this way, the system can more comprehensively understand the mutual influence between the components and provide a basis for subsequent risk assessment.

[0032] Step 302: construct a target virtual copy using digital twin technology, and use the target virtual copy to simulate the behavior of the railway contact network under different assumed working scenarios to generate physical simulation results; In this step, the target virtual replica refers to a virtual model of the railway contact network created using digital twin technology, which can simulate its behavior under different assumed working scenarios and generate physical simulation results. The physical simulation results refer to the data on the behavior of each component of the contact network obtained through the simulation of the target virtual replica under different assumed working scenarios for further analysis and evaluation. The assumed working scenario refers to a series of hypothetical environments and operating conditions set to evaluate and simulate the behavior of the railway contact network under different conditions. These scenarios are usually based on historical data, forecast information, and possible extreme situations, including normal operation scenarios, high-density train operation scenarios, and extreme weather conditions scenarios.

[0033] This step uses digital twin technology to build a target virtual copy of the railway contact network. This virtual copy can simulate the behavior of the contact network under different assumed working scenarios and generate physical simulation results. For example, assuming that the temperature is 30°C, the wind speed is 15m / s, and there are 5 trains passing frequently on a certain day, the system inputs these conditions into the virtual copy to simulate the dynamic response of the catenary under these environmental conditions. Through this simulation, the system can observe the changes in the behavior of the catenary and other components under different working conditions, providing an important basis for evaluating the state of the contact network.

[0034] Step 303: combining the conditional probability distribution results between the components and the physical simulation results, comprehensively evaluating the state of the railway contact network to obtain a preliminary fault prediction probability; For example, if the Bayesian network analysis shows a strong dependency between the catenary cables and the support poles, and the physical simulation results show an abnormal increase in the tension of the catenary cables under high wind conditions, the system will increase the initial failure estimate probability, indicating a higher risk of failure.

[0035] Step 304: Based on the preliminary fault prediction probability and in combination with multi-context perception data, a graph neural network is applied to analyze the correlation between various components in the railway contact network structure to obtain a high-dimensional feature vector, and the high-dimensional feature vector is processed using a long short-term memory network to obtain key time characteristics of tension changes; In this step, the key time characteristics refer to the characteristics of the evolution of tension changes over time obtained after processing the high-dimensional feature vector through the long short-term memory network, which are used for fault prediction.

[0036] For example, the high-dimensional feature vectors extracted by graph neural networks describe the complex correlations between the load-bearing cables and other components. The graph neural network further analyzes how these features change over time, thereby accurately capturing the key time characteristics of tension changes and providing support for subsequent fault prediction.

[0037] Step 305: Based on the critical time characteristics, predict the target failure probability of failure within a specific time period, and introduce Bayesian optimization and Monte Carlo simulation in the prediction process to quantify the uncertainty in the prediction for risk assessment; This step predicts the target estimated probability of failure within a specific time period based on the above key time characteristics. Bayesian optimization and Monte Carlo simulation are introduced in the prediction process to quantify the uncertainty in the prediction for risk assessment. For example, the model parameters are adjusted through Bayesian optimization to ensure the accuracy of the prediction results; at the same time, Monte Carlo simulation is used to perform multiple random sampling to evaluate the uncertainty in the prediction. Finally, the system outputs the target estimated probability of failure of the catenary within the next 24 hours, providing maintenance personnel with a scientific basis for decision-making.

[0038] The embodiment of the present invention analyzes the dependencies between components through Bayesian networks, so that the system can more comprehensively understand the complex interactions within the contact network and improve the accuracy of risk assessment. The target virtual copy constructed using digital twin technology can simulate the behavior of the contact network under different assumed working scenarios, provide detailed physical simulation results, and enhance the reliability of the prediction. Combining the conditional probability distribution results and the physical simulation results, the system can comprehensively evaluate the overall state of the contact network and obtain a more accurate preliminary fault prediction probability. Through the application of graph neural networks and long short-term memory networks, the system can efficiently capture the key time characteristics of tension changes and provide strong support for fault prediction. By introducing Bayesian optimization and Monte Carlo simulation, the system can quantify the uncertainty in the prediction and ensure the scientificity and reliability of risk assessment.

[0039] In order to more accurately predict and evaluate the behavior of the railway contact network catenary under various complex environments, based on this, the present invention provides a specific embodiment, step 302, using digital twin technology to build a target virtual copy, using the target virtual copy to simulate the behavior of the railway contact network under different assumed working scenarios, and generating physical simulation results, specifically including the following steps: Step 311: Based on the structure and parameters of the railway contact network, construct an initial virtual copy corresponding to the railway contact network using digital twin technology; In this step, the structure and parameters of the railway contact network refer to the specific physical structure of the railway contact network (such as load-bearing cables, support rods, insulators, etc.) and its related parameters (such as length, diameter, material properties, etc.), which are used to build the basis of the virtual model. The initial virtual copy refers to the preliminary virtual model of the railway contact network created using digital twin technology, which can simulate the behavior of the actual contact network in the computer based on its structure and parameters.

[0040] For example, suppose a section of railway overhead contact network contains load-bearing cables, support rods and insulators of specific lengths. These structures and parameters are input into the virtual modeling software to generate a detailed virtual copy that can simulate the basic physical behavior of the overhead contact network and provide the basis for subsequent real-time updates and simulations.

[0041] Step 312: Based on the real-time status data of the railway contact network, synchronously update the internal status of the initial virtual copy to obtain an updated virtual copy, so that the internal status of the updated virtual copy reflects the status of the railway contact network catenary cables; In this step, real-time status data refers to data collected in real time from sensors and other monitoring equipment on the railway contact network, including environmental conditions such as tension, temperature, humidity, wind speed, and train operation status information. Synchronous update processing refers to dynamically adjusting the internal state of the virtual copy based on the real-time status data to ensure that the virtual copy can accurately reflect the status of the actual contact network. The updated virtual copy refers to the virtual copy that has been synchronized and updated, and its internal state is consistent with the actual contact network, which can truly reflect the current working situation.

[0042] For example, sensors installed on the railway contact network collect data such as tension, temperature, and humidity in real time, and transmit these data to the virtual copy, dynamically adjusting the internal state of the virtual copy to keep it consistent with the state of the actual contact network. In this way, the virtual copy can reflect the dynamic response of the catenary under different environmental conditions in real time.

[0043] Step 313: Based on the real-time operation and monitoring data, the internal parameters of the updated virtual copy are updated to generate a target virtual copy to adapt to the changes in the factors affecting the train operation; In this step, real-time operation and monitoring data refers to data continuously generated during the actual operation of the railway contact network, including but not limited to train operation density, weather changes, etc., which are used to further adjust the parameters of the virtual copy. The target virtual copy refers to the virtual copy updated and processed by the real-time operation and monitoring data, which can adapt to the changes in the factors affecting the operation of the train and more accurately simulate the behavior of the actual contact network.

[0044] For example, by combining real-time operational data (such as train operation density, vehicle type, etc.) and monitoring data (such as weather forecast information), the parameters in the virtual copy can be further adjusted to adapt it to different train operation conditions and environmental changes. In this way, the target virtual copy not only reflects the current state of the overhead line, but also predicts possible changes in the future, providing a more accurate basis for subsequent simulations.

[0045] Step 314: Based on the target virtual copy, multi-scenario simulation is performed under different assumed working scenarios to evaluate the different assumed working scenarios and the influence of the assumed working scenarios on the railway contact network catenary cables, and obtain physical simulation results; For example, under assumed extreme weather conditions (such as strong winds, heavy rain, heavy snow, etc.) and high-density train operation scenarios, simulations are performed using target virtual copies to observe the behavioral changes of catenary cables and other components. Through these simulations, the system can evaluate the impact of different working conditions on the contact network, identify potential risk points, and provide a scientific basis for maintenance decisions.

[0046] The initial virtual copy constructed by the digital twin technology in the embodiment of the present invention can accurately simulate the structure and behavior of the railway contact network, providing a solid foundation for subsequent real-time updates and simulations; based on the synchronous update processing of real-time status data, it ensures that the virtual copy can reflect the status of the actual contact network in real time, thereby improving the accuracy and reliability of the simulation; through the update processing of real-time operation and monitoring data, the generated target virtual copy can adapt to the changes in factors affecting train operation, thereby enhancing the adaptability and flexibility of the system; multi-scenario simulation under different assumed working scenarios can comprehensively evaluate the impact of various working conditions on the contact network, identify potential risk points, and provide a scientific basis for maintenance decisions.

[0047] In order to systematically evaluate the impact of different working scenarios on the railway contact network catenary cables, based on this, the present invention provides a specific embodiment, step 314, based on the target virtual copy, multi-scenario simulation is performed under different assumed working scenarios to evaluate the impact of different assumed working scenarios and the assumed working scenarios on the railway contact network catenary cables, and obtain physical simulation results, which specifically includes the following steps: Step 321: Based on the train operation influencing factors, different assumed working scenarios are set to evaluate the influence of the train operation influencing factors on the railway contact network catenary cables; For example, the following hypothetical working scenarios are set: high-density train operation scenario is that 10 trains pass frequently every hour; strong wind scenario is that the wind speed reaches 20m / s; heavy rain scenario is that the rainfall exceeds 50mm / h and the humidity is close to 100%; heavy snow scenario is that the snow thickness exceeds 10cm and the temperature is lower than -10°C; high temperature scenario is that the temperature reaches 40°C; low temperature scenario is that the temperature drops to -20°C. These scenarios are designed to comprehensively evaluate the different effects of train operation influencing factors on the overhead contact network load-bearing cables.

[0048] Step 322: Based on the assumed working scenario, setting simulation environment parameters, optimizing the simulation environment parameters based on the real-time operation and monitoring data and the historical operation and environment data, and obtaining optimized simulation environment parameters; In this step, the simulated environmental parameters refer to the parameter settings used to simulate specific working scenarios, including but not limited to temperature, humidity, wind speed, train running density, etc., which are used to build a virtual environment. Real-time operation and monitoring data refer to data collected in real time from sensors and other monitoring equipment on the railway contact network, including environmental conditions such as tension, temperature, humidity, wind speed, and train running status information. Historical operation and environmental data refer to the actual operation data and environmental condition records of the railway contact network over a period of time in the past, which are used to optimize the simulation environment parameters.

[0049] For example, when setting a high wind scenario, the initial simulation environment parameter is a wind speed of 20m / s, but combined with real-time operation and monitoring data and historical operation and environmental data, it is found that under similar conditions, the wind speed usually fluctuates by ±5m / s, so the optimized simulation environment parameter is a wind speed of 15-25m / s. In this way, the simulation environment parameters are ensured to be closer to the actual situation, improving the accuracy and reliability of the simulation.

[0050] Step 323: Based on the optimized simulation environment parameters, a multi-scenario simulation process is performed on the target virtual copy. During the multi-scenario simulation process, the dynamic environment and operating condition changes reflecting different assumed working scenarios are changed to obtain target state information of the railway contact network catenary under different dynamic environment and operating condition changes. Based on the target state information, the behavior response of the catenary under different assumed working scenarios is obtained. In this step, dynamic environmental and operating condition changes refer to the environmental conditions (such as weather changes) and operating conditions (such as changes in train running density) that are constantly changing during the simulation process to reflect the complex situations in actual operations.

[0051] For example, in a high-density train operation scenario, 10 trains passing per hour are simulated to observe the tension changes, vibration amplitude and other behavioral responses of the catenary cables; in a high-wind scenario, the wind speed change of 15-25m / s is simulated to evaluate its impact on the catenary cables. Through these simulations, the system can obtain detailed target state information of the catenary cables under various dynamic environments and operating conditions, and finally obtain the behavioral responses of the catenary cables under different assumed working scenarios.

[0052] The change in tension of the railway contact network catenary cables is affected by many factors, including meteorological conditions (such as temperature, humidity, wind speed), train operation density, and maintenance activities. In order to more accurately predict and evaluate these influences, a model that can comprehensively consider multiple factors is needed. The traditional single-factor model cannot fully capture the complex dynamic behavior of the system, so a multi-factor dynamic state evolution model is introduced to integrate multiple environmental parameters and operating conditions. More specifically, the present invention provides a calculation formula for a multi-factor dynamic state evolution model, through which the changes in the state of the railway contact network catenary cables at different time points can be accurately described. The specific calculation formula is as follows: ; Indicates at time When , the state information of the kth assumed working scenario is used to describe the state evolution from the current time t to the next time t+1; represents the status information of the kth assumed working scenario at the current time t, is a vector containing multiple state variables, such as the tension, temperature, and displacement of the load-bearing cable, which can be expressed as , represents the value of the nth state variable at time t; Represents the transpose operation, converting the row vector into a column vector, ensuring that the vector The correct representation of , to facilitate subsequent mathematical operations; Represents the dynamic system model function, which is used to calculate the specific mathematical model of the state evolution of the railway contact network load-bearing cables; Represents the activation function, which is used to introduce nonlinear transformation; Represents the weight matrix, used for linear transformation vector ; Represents the bias vector, used for linear transformation vector ; Represents the optimized simulation environment parameters, which are used to adjust the assumed working scene conditions to ensure that the physical simulation results are as close to the actual situation as possible; It represents the environmental parameters under the kth assumed working scenario, describes the parameters of specific environmental conditions, and reflects the impact of external factors on system behavior; Represents the dynamic environment and operating condition change vector, which contains multiple components; Parameters representing the dynamic system model function; represents the random noise term, that is, unforeseen interference or error.

[0053] The multi-factor dynamic state evolution model comprehensively considers the influence of multiple factors and provides a more accurate prediction of the behavior of the contact cables. The model can adjust parameters in real time according to the latest operating data and environmental conditions to ensure that the simulation results are always close to the actual situation. The multi-factor dynamic state evolution model can be used to perform multi-scenario simulations under different assumed working scenarios, evaluate the impact of various factors on the contact network contact cables, and provide a more comprehensive analysis perspective.

[0054] Among them, the specific expressions of the dynamic environment and operating condition change vector are as follows: ; in, Represents the dynamic environment and operating condition change vector, which contains multiple components; Indicates combining historical temperature data and the parameters of the time series forecasting model , use the time series prediction model to predict the temperature at the current time t; Indicates that combined with historical humidity data and the parameters of the time series forecasting model , use the time series prediction model to predict the humidity at the current time t; Indicates the combination of historical wind speed and wind direction data and the parameters of the time series forecasting model , use the time series prediction model to predict the wind speed and wind direction at the current time t; Indicates that combined with weather forecast data and the parameters of the probability distribution model , use the probability distribution model to predict the precipitation at the current time t; Indicates that the historical train operation data is combined , Scheduling plan and the parameters of the time series forecasting model , use the time series forecasting model to predict the current time The train operation density; Indicates that real-time monitoring data and historical speed data and the parameters of the mixed model , use the hybrid model to predict the vehicle speed at the current time t; Indicates the historical data related to the vehicle model and the parameters of the classification model , use the classification model to predict the model of the current time t; Indicates combined maintenance log data and maintenance plan data and the parameters of the time series forecasting model , use the time series forecasting model to predict the maintenance activities at the current time t; Represents geographic information data combined with tracks and the parameters of the space-time model , using the space-time model to predict the orbital conditions at the current time t; The above-mentioned dynamic environment and operating condition change vector combines multiple data sources such as historical data, real-time monitoring data, weather forecasts, maintenance logs, etc., and introduces advanced algorithms such as time series prediction models, classification models, and space-time models for prediction, which enhances the modeling ability of dynamic environment and operating condition changes, so that the final multi-factor dynamic state evolution model can flexibly respond to various unforeseen operating conditions and environmental changes, and provide more reliable early warning information.

[0055] Step 324: Based on the key indicators of the state of the catenary cable, identifying the influencing factors that affect the behavioral response of the catenary cable, and recording the abnormal phenomena, and evaluating the performance of the catenary cable based on the influencing factors and the abnormal phenomena to obtain the performance impact results; In this step, the behavior response of the catenary refers to the response characteristics of the catenary under different assumed working scenarios, such as tension change, vibration amplitude, etc. The key indicators of the catenary state refer to the key parameters used to evaluate the performance of the catenary, such as maximum tension, minimum tension, average tension change rate, etc. Abnormal phenomena refer to behaviors or states that are beyond the normal range identified during the simulation process, which may indicate potential problems. Performance impact results refer to the results of a comprehensive evaluation of the performance of the catenary based on factors such as the behavior response of the catenary, influencing factors, and abnormal phenomena.

[0056] For example, in a high wind scenario, if the maximum tension of the catenary exceeds the safety threshold, the system will record this abnormal phenomenon and further analyze whether it is caused by excessive wind speed or other factors (such as insufficient support rod strength). Through this evaluation, the system can identify the main influencing factors and conduct a comprehensive evaluation of the performance of the catenary to obtain performance impact results.

[0057] Step 325: generating a physical simulation result based on the behavior response of the catenary, the influencing factors, the abnormal phenomenon and the performance impact result; For example, by integrating all simulation data, the system generates a detailed physical simulation report, including the behavior response of the catenary under different assumed working scenarios, the main influencing factors, abnormal phenomena and their specific impact on the performance of the catenary. This report not only provides a scientific basis for risk assessment, but also helps maintenance personnel take preventive measures in advance to ensure the safety and efficiency of railway operations.

[0058] By setting a variety of hypothetical working scenarios, the embodiment of the present invention can enable the system to comprehensively evaluate the impact of train operation influencing factors on the contact network catenary cables, thereby ensuring the comprehensiveness and accuracy of risk assessment; combining real-time operation and monitoring data with historical data, optimizing simulation environment parameters, improving the accuracy and reliability of simulation, and ensuring that the simulation results are closer to the actual situation; through multi-scenario simulation processing, the system can evaluate the behavioral response of the catenary cables under different dynamic environments and changes in operating conditions, capture complex real-life situations, and enhance the adaptability and flexibility of the system; based on the key indicators of the catenary cable status, the system can accurately identify the main factors affecting the behavioral response of the catenary cables, and record abnormal phenomena, providing a scientific basis for subsequent performance evaluation.

[0059] In order to improve the accuracy and timeliness of the fault prediction of the catenary of the railway contact network, based on this, the present invention provides a specific embodiment, step 304, based on the preliminary fault prediction probability, combined with multi-context perception data, applying a graph neural network to analyze the correlation between various components in the railway contact network structure, obtaining a high-dimensional feature vector, and using a long short-term memory network to process the high-dimensional feature vector to obtain the key time characteristics of the tension change, specifically including the following steps: Step 331: Analyze the status of each component in the railway contact network based on the preliminary fault prediction probability and multi-scenario perception data to obtain component status information; In this step, the component status information refers to the status information of each component in the railway contact network (such as load-bearing cables, support rods, insulators, etc.) under current working conditions, including parameters such as tension value, temperature, and humidity.

[0060] For example, assuming that the initial fault prediction probability is 0.05, combined with the current weather forecast information (such as temperature 25°C, humidity 60%, wind speed 10m / s) and real-time monitoring data (such as the tension of the catenary cable 80N), the system can analyze the specific status information of components such as the catenary cable, support rod and insulator, including their tension value, temperature, humidity and other parameters. This status information provides the basis for the subsequent construction of the graph structure.

[0061] Step 332: Based on the component status information, define nodes and edges to construct a graph structure, wherein the nodes represent the components of the railway contact network, and the edges represent the physical connections between the components; In this step, the graph structure refers to a mathematical structure composed of nodes and edges, which is used to represent the components of the railway contact network and the associations between them. Nodes refer to the components of the railway contact network in the graph structure, such as load-bearing cables, support rods, insulators, etc. Edges refer to the physical connection relationship between the components in the graph structure, reflecting the interaction and dependency between the components.

[0062] For example, the catenary cables, support rods and insulators are defined as nodes in the graph, and the edges are defined according to the physical connection between them (such as the catenary cables are connected to the support rods, and the support rods are connected to the insulators). In this way, the system builds a complete graph structure that intuitively shows the association and interaction between the components.

[0063] Step 333: Based on the graph structure, a graph neural network is used to analyze the correlation between various components in the railway contact network to obtain a high-dimensional feature vector containing the correlation between various components, and the high-dimensional feature vector reflects the state of the railway contact network catenary cables; In this step, the high-dimensional feature vector refers to a set of multidimensional features extracted by the graph neural network that describes the correlation between the components of the railway contact network, which is used for further key time characteristics analysis.

[0064] For example, through the graph neural network, the system can capture the dependency between the catenary and the support rod, as well as the mutual influence between the support rod and the insulator. These relationships are encoded as high-dimensional feature vectors, which contain rich association information and provide important input for subsequent time characteristic analysis.

[0065] Step 334: Processing the high-dimensional feature vector using the internal memory unit of the long short-term memory network to obtain key time characteristics of the high-dimensional feature vector, wherein the key time characteristics include tension change trend and periodic fluctuation; For example, the LSTM network can identify the time-varying trend (such as gradual increase or decrease) and periodic fluctuations (such as periodic peaks in tension when a fixed number of trains pass through every hour) of the catenary through its internal memory units. These key time characteristics provide an important basis for fault prediction and risk assessment.

[0066] Through preliminary fault prediction probability and multi-scenario perception data, the embodiment of the present invention enables the system to accurately analyze the status of each component in the railway contact network to ensure the accuracy and reliability of the status information; by defining nodes and edges to construct a graph structure, the system can intuitively display the correlation and interaction between components, thereby enhancing the interpretability of the system; by using graph neural networks to analyze component correlations, the system can capture complex interaction relationships and extract high-dimensional feature vectors, providing rich input for subsequent time characteristic analysis; by processing high-dimensional feature vectors through long short-term memory networks, the system can efficiently capture the trend and periodic fluctuations of tension changes, thereby providing a scientific basis for fault prediction and risk assessment.

[0067] In order to solve the problems of inadaptability, false alarm and missed alarm of traditional fixed alarm threshold in complex and changeable environment, based on this, the present invention provides a specific embodiment, step 104, based on the target fault prediction probability and the multi-context perception data, the preset alarm threshold is dynamically adjusted by using an adaptive control algorithm to obtain the adjusted alarm threshold, when the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, early warning information is immediately generated, and the early warning information is processed by blockchain technology to obtain the target early warning information, which specifically includes the following steps: Step 401: Based on the fault prediction probability and the multi-scenario perception data, the preset alarm threshold is dynamically adjusted by using an adaptive control algorithm and a reinforcement learning algorithm to obtain an adjusted alarm threshold; This step uses adaptive control algorithms and reinforcement learning algorithms to dynamically adjust preset alarm thresholds based on the estimated probability of failure (for example, 0.03 means there is a 3% chance of failure within the next 24 hours) and multi-context perception data (such as real-time weather conditions and train flow). For example, in a busy railway section, due to the additional stress caused by frequent trains passing by, coupled with adverse weather conditions (such as strong winds or low temperatures), adaptive control algorithms and reinforcement learning algorithms may increase the alarm threshold to reduce false alarms while ensuring that any potential problems can be discovered in a timely manner. The final adjusted alarm threshold reflects the most appropriate safety standard at the moment.

[0068] Step 402: Based on the tension data stream, monitor the real-time tension value of the railway contact network catenary cable, and when the real-time tension value exceeds the adjusted alarm threshold, use an anomaly detection algorithm to identify abnormal data points of the tension data stream, and combine the abnormal data points with the event-driven architecture to generate early warning information in real time; This step monitors the real-time tension value of the catenary cables of the railway contact network based on the tension data stream (i.e., the continuously collected catenary cable tension readings). When the real-time tension value exceeds the above-adjusted alarm threshold, the anomaly detection algorithm is immediately activated to identify abnormal data points in the tension data stream. For example, if the tension at a certain moment suddenly increases significantly and exceeds the normal fluctuation range, it is considered an abnormal point. Combining these abnormal data points with the event-driven architecture, the system will instantly generate early warning information to notify relevant personnel to pay attention and take necessary preventive measures.

[0069] Step 403: Encrypt the warning information using homomorphic encryption technology to obtain encrypted warning information, introduce blockchain technology, and perform tamper-proof processing on the encrypted warning information through the consensus mechanism and timestamp function of blockchain to obtain target warning information; In this step, anti-tampering processing refers to the use of blockchain characteristics to ensure that the warning information cannot be changed once it is recorded, thereby enhancing the reliability and security of the information.

[0070] This step uses homomorphic encryption technology to encrypt the warning information to obtain encrypted warning information, ensuring that even if it is intercepted during network transmission, the specific content cannot be leaked. Then, blockchain technology is introduced to perform tamper-proof processing on the encrypted warning information through its consensus mechanism and timestamp function, thereby obtaining the target warning information. This not only protects the privacy of the information, but also ensures the authenticity and integrity of the warning information due to the tamper-proof nature of the blockchain, preventing malicious tampering or deletion.

[0071] The embodiments of the present invention reduce the problem of false alarms or missed alarms caused by fixed thresholds by dynamically adjusting the alarm thresholds, thereby improving the accuracy of the early warning system; adopt an event-driven architecture so that the system can respond quickly when abnormal data points are detected and generate early warning information in real time, thereby improving the speed of emergency response; use homomorphic encryption technology and blockchain technology to ensure the security and tamper-proofness of early warning information throughout the entire processing process, thereby improving the level of information security; provide more accurate and timely early warning information, which helps the maintenance team to plan and implement effective maintenance measures in advance, thereby reducing the risk of accidents.

[0072] In order to further improve the reliability of early warning information and solve the limitations of traditional fixed alarm thresholds in dynamic environments, the present invention provides a specific embodiment. Step 401 dynamically adjusts the preset alarm threshold based on the fault prediction probability and the multi-context perception data using an adaptive control algorithm and a reinforcement learning algorithm to obtain an adjusted alarm threshold, which specifically includes the following steps: Step 411: Based on the target fault prediction probability and the multi-scenario perception data, a comprehensive risk assessment model is constructed using a Bayesian optimization algorithm, and the risk assessment model is used to iteratively update the prior distribution to obtain an updated prior distribution, while maximizing the posterior probability. Based on the updated prior distribution and the maximized posterior probability, the risk level of the railway overhead contact network load-bearing cable in a real-time environment is evaluated to generate a real-time risk assessment result; In this step, the comprehensive risk assessment model refers to a risk assessment model that combines multiple factors (such as historical data, real-time monitoring data, environmental conditions, etc.) to predict and assess the risk level of railway contact network load-bearing cables under different working conditions.

[0073] For example, the initial prior distribution assumes that the catenary cables are working properly, but as new observation data (such as tension change patterns) are continuously introduced, the prior distribution is gradually updated. Ultimately, based on the updated prior distribution and the maximized posterior probability, the system can accurately assess the risk level of the catenary cables of the railway contact network in a real-time environment, and generate real-time risk assessment results, providing a scientific basis for risk warning. For example, the system outputs a risk score of 75 points (out of 100 points) for the catenary cables at the current moment, indicating that there is a certain risk but it has not yet reached the warning level.

[0074] Step 412: According to the real-time risk assessment result, the preset alarm threshold is dynamically adjusted by applying the adaptive control algorithm to generate a preliminary adjusted alarm threshold, and a fuzzy logic controller is introduced to handle uncertainty and nonlinear relationship in the process of applying the adaptive control algorithm to perform dynamic adjustment, so as to optimize the adjustment process; For example, if the real-time risk score exceeds 60 points, the system adjusts the alarm threshold from the default 80 N to 70 N. At the same time, the fuzzy logic controller will fine-tune the alarm threshold based on the uncertainty factor of the sudden increase in ambient temperature to ensure that it adapts to the new environmental conditions, thereby improving the flexibility and accuracy of the alarm threshold setting.

[0075] Step 413: introducing a reinforcement learning algorithm, defining a reward function, evaluating the warning effect under different alarm thresholds, optimizing the alarm threshold after the preliminary adjustment based on the warning effect, and obtaining the adjusted alarm threshold. In the process of introducing the reinforcement learning algorithm for processing, the configuration of the alarm threshold after the preliminary adjustment is adjusted in combination with the double Q learning algorithm and the active learning mechanism, according to the actual warning effect and user feedback, so as to search for the optimal alarm threshold as the adjusted alarm threshold; For example, the system tests different alarm thresholds multiple times and gives positive or negative rewards based on the actual warning effect (such as whether potential problems are discovered in a timely manner). It uses the dual Q learning algorithm to reduce over-estimation problems and combines the active learning mechanism to select the most valuable data points for training. Ultimately, the optimal alarm threshold is determined to be 65N, ensuring that the alarm threshold configuration is always optimal, thereby improving the reliability and accuracy of the early warning system.

[0076] The embodiment of the present invention uses a comprehensive risk assessment model constructed by a Bayesian optimization algorithm. The system can more accurately assess the risk level of railway contact network load-bearing cables in a real-time environment and provide a scientific basis for risk warning. Combined with an adaptive control algorithm and a fuzzy logic controller, the system can dynamically adjust the alarm threshold to adapt to different working conditions and environmental changes, reduce false alarms and missed alarms, and improve the flexibility and accuracy of the early warning system. Through a reinforcement learning algorithm and its improved technology (such as a dual-Q learning algorithm and an active learning mechanism), the system can continuously optimize the alarm threshold configuration according to actual warning effects and user feedback to ensure that it is always in the optimal state, thereby improving the reliability and response speed of the early warning system.

[0077] Figure 2 A schematic diagram of a railway overhead line tension online monitoring and early warning system is provided for an embodiment of the present invention. Figure 2 As shown, the system includes: The acquisition module 21 is used to acquire dynamic tension data of the railway contact network catenary under different environmental conditions by using the situational awareness technology based on the factors affecting train operation, and obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; An analysis module 22 is used to apply a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and use a convolutional neural network and a graph neural network to process the preliminary analysis result to generate an abnormal pattern recognition result; A prediction module 23 is used to perform risk assessment based on the abnormal pattern recognition result using a Bayesian network algorithm and digital twin technology to obtain a preliminary estimated fault probability, and to predict a target estimated fault probability of a fault occurring within a specific time period by combining multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information; The generation module 24 is used to dynamically adjust the preset alarm threshold value based on the target fault estimation probability and the multi-scenario perception data by using an adaptive control algorithm to obtain an adjusted alarm threshold value. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold value, an early warning message is immediately generated, and the early warning message is processed through blockchain technology to obtain target early warning information.

[0078] Figure 2 The railway overhead contact network catenary cable tension online monitoring and early warning system can be implemented Figure 1 The implementation principle and technical effect of the method for online monitoring and early warning of the tension of catenary cables of a railway contact network described in the embodiment shown are not described in detail. The specific manner in which each module and unit performs operations in the online monitoring and early warning system for the tension of catenary cables of a railway contact network in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0079] In one possible design, Figure 2 The railway overhead contact network catenary cable tension online monitoring and early warning system of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0080] The processing component 32 is used to: based on the factors affecting train operation, collect dynamic tension data of the railway contact network bearing cable under different environmental conditions using situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; apply a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, use a convolutional neural network and a graph neural network to process the preliminary analysis result, and generate an abnormal pattern recognition result; based on the abnormal pattern recognition result, use a Bayesian network algorithm and a digital twin technology to perform risk assessment to obtain a preliminary fault prediction probability, combine the multi-situational awareness data and the preliminary fault prediction probability, and predict a target fault prediction probability of a fault occurring within a specific time period, wherein the multi-situational awareness data includes historical maintenance records and weather forecast information; based on the target fault prediction probability and the multi-situational awareness data, use an adaptive control algorithm to dynamically adjust a preset alarm threshold to obtain an adjusted alarm threshold, and when the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, generate an early warning message in real time, and process the early warning message through blockchain technology to obtain a target early warning message.

[0081] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0082] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0083] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0084] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0085] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0086] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0087] The embodiment of the present invention further provides a computer storage medium storing a computer program, which can achieve the above-mentioned Figure 1 The illustrated embodiment is a method for online monitoring and early warning of the tension of the catenary cables of a railway contact network.

[0088] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0089] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0090] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for online monitoring and early warning of the tension of a railway overhead contact network, characterized in that: include: Based on the factors affecting train operation, the dynamic tension data of the railway contact network bearing cables under different environmental conditions are collected using situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; Applying a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and using a convolutional neural network and a graph neural network to process the preliminary analysis result to generate an abnormal pattern recognition result; Based on the abnormal pattern recognition result, a risk assessment is performed using a Bayesian network algorithm and digital twin technology to obtain a preliminary estimated fault probability, and a target estimated fault probability of a fault occurring within a specific time period is predicted by combining multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information; Based on the target fault prediction probability and the multi-scenario perception data, an adaptive control algorithm is used to dynamically adjust the preset alarm threshold to obtain an adjusted alarm threshold. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, a warning message is immediately generated, and the warning message is processed through blockchain technology to obtain target warning information.

2. The method according to claim 1, characterized in that Based on the abnormal pattern recognition results, Bayesian network algorithm and digital twin technology are used to perform risk assessment to obtain a preliminary estimated fault probability. Combined with multi-context perception data and the preliminary estimated fault probability, the target estimated fault probability of a fault occurring within a specific time period is predicted. The multi-context perception data includes historical maintenance records and weather forecast information, including: Based on the abnormal pattern recognition results, the Bayesian network is used to analyze the dependencies between the components of the railway contact network, and the conditional probability distribution results between the components are obtained; Using digital twin technology to build a target virtual copy, using the target virtual copy to simulate the behavior of the railway contact network under different assumed working scenarios, and generating physical simulation results; Combining the conditional probability distribution results between the components and the physical simulation results, comprehensively evaluating the state of the railway contact network to obtain a preliminary fault prediction probability; Based on the preliminary fault prediction probability and combined with multi-context perception data, a graph neural network is used to analyze the correlation between various components in the railway contact network structure to obtain a high-dimensional feature vector, and a long short-term memory network is used to process the high-dimensional feature vector to obtain the key time characteristics of the tension change; Based on the critical time characteristics, the target estimated probability of failure occurring within a specific time period is predicted, and Bayesian optimization and Monte Carlo simulation are introduced in the prediction process to quantify the uncertainty in the prediction for risk assessment.

3. The method according to claim 2, characterized in that The digital twin technology is used to construct a target virtual copy, and the target virtual copy is used to simulate the behavior of the railway contact network under different assumed working scenarios to generate physical simulation results, including: Based on the structure and parameters of the railway contact network, an initial virtual copy corresponding to the railway contact network is constructed using digital twin technology; Based on the real-time status data of the railway contact network, synchronously updating the internal status of the initial virtual copy to obtain an updated virtual copy, so that the internal status of the updated virtual copy reflects the status of the railway contact network catenary cables; Based on real-time operation and monitoring data, the internal parameters of the updated virtual copy are updated to generate a target virtual copy to adapt to changes in factors affecting train operation; Based on the target virtual copy, multi-scenario simulations are performed under different assumed working scenarios to evaluate the different assumed working scenarios and the impact of the assumed working scenarios on the railway contact network load-bearing cables, and obtain physical simulation results.

4. The method according to claim 3, characterized in that Based on the target virtual copy, multiple scenario simulations are performed under different assumed working scenarios to evaluate the different assumed working scenarios and the impact of the assumed working scenarios on the railway contact network catenary cables, and physical simulation results are obtained, including: Based on the train operation influencing factors, different assumed working scenarios are set to evaluate the influence of the train operation influencing factors on the railway contact network catenary cables; Based on the assumed working scenario, simulation environment parameters are set, and based on the real-time operation and monitoring data and the historical operation and environment data, the simulation environment parameters are optimized to obtain optimized simulation environment parameters; Based on the optimized simulation environment parameters, the target virtual copy is subjected to multi-scenario simulation processing. During the multi-scenario simulation processing, the dynamic environment and operating condition changes reflecting different assumed working scenarios are changed to obtain target state information of the railway contact network catenary under different dynamic environment and operating condition changes. Based on the target state information, the behavior response of the catenary under different assumed working scenarios is obtained; Based on the key indicators of the state of the catenary cable, the influencing factors affecting the behavioral response of the catenary cable are identified, and the abnormal phenomena are recorded at the same time. Based on the influencing factors and the abnormal phenomena, the performance of the catenary cable is evaluated to obtain the performance impact results; A physical simulation result is generated based on the behavior response of the catenary, the influencing factors, the abnormal phenomenon and the performance impact result.

5. The method according to claim 2, characterized in that: Based on the preliminary fault prediction probability and combined with multi-context perception data, a graph neural network is used to analyze the correlation between various components in the railway contact network structure to obtain a high-dimensional feature vector. The high-dimensional feature vector is processed using a long short-term memory network to obtain the key time characteristics of tension changes, including: Based on the preliminary fault prediction probability and multi-scenario perception data, the status of each component in the railway contact network is analyzed to obtain component status information; Based on the component status information, nodes and edges are defined to construct a graph structure, wherein the nodes represent the components of the railway contact network, and the edges represent the physical connections between the components; Based on the graph structure, a graph neural network is used to analyze the correlation between various components in the railway contact network, and a high-dimensional feature vector containing the correlation between various components is obtained, and the high-dimensional feature vector reflects the state of the railway contact network load-bearing cables; The high-dimensional feature vector is processed by using an internal memory unit of a long short-term memory network to obtain key time characteristics of the high-dimensional feature vector, wherein the key time characteristics include tension change trend and periodic fluctuation.

6. The method according to claim 1, characterized in that Based on the target fault prediction probability and the multi-scenario perception data, the preset alarm threshold is dynamically adjusted using an adaptive control algorithm to obtain an adjusted alarm threshold. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold, an early warning message is immediately generated, and the early warning message is processed through blockchain technology to obtain target early warning information, including: Based on the fault prediction probability and the multi-scenario perception data, using an adaptive control algorithm and a reinforcement learning algorithm, dynamically adjusting a preset alarm threshold to obtain an adjusted alarm threshold; Based on the tension data stream, the real-time tension value of the railway contact network load-bearing cable is monitored. When the real-time tension value exceeds the adjusted alarm threshold, an abnormal data point of the tension data stream is identified by using an abnormality detection algorithm, and an early warning message is generated in real time by combining the abnormal data point and the event-driven architecture; The warning information is encrypted using homomorphic encryption technology to obtain encrypted warning information, and blockchain technology is introduced to perform tamper-proof processing on the encrypted warning information through the consensus mechanism and timestamp function of the blockchain to obtain target warning information.

7. The method according to claim 6, characterized in that Based on the fault prediction probability and the multi-scenario perception data, the preset alarm threshold is dynamically adjusted by using an adaptive control algorithm and a reinforcement learning algorithm to obtain an adjusted alarm threshold, including: Based on the target fault prediction probability and the multi-scenario perception data, a Bayesian optimization algorithm is used to construct a comprehensive risk assessment model, and the risk assessment model is used to iteratively update the prior distribution to obtain an updated prior distribution, while maximizing the posterior probability. Based on the updated prior distribution and the maximized posterior probability, the risk level of the railway contact network load-bearing cables in a real-time environment is evaluated to generate a real-time risk assessment result; According to the real-time risk assessment results, the preset alarm threshold is dynamically adjusted by applying an adaptive control algorithm to generate a preliminary adjusted alarm threshold, and a fuzzy logic controller is introduced to handle uncertainty and nonlinear relationships in the process of applying the adaptive control algorithm for dynamic adjustment to optimize the adjustment process; A reinforcement learning algorithm is introduced to evaluate the warning effects under different alarm thresholds by defining a reward function. The preliminary adjusted alarm threshold is optimized based on the warning effect to obtain the adjusted alarm threshold. In the process of introducing the reinforcement learning algorithm for processing, the double Q learning algorithm and the active learning mechanism are combined to adjust the configuration of the preliminary adjusted alarm threshold according to the actual warning effect and user feedback, so as to search for the optimal alarm threshold as the adjusted alarm threshold.

8. An online monitoring and early warning system for the tension of railway contact network catenary cables, characterized in that: include: A collection module, which is used to collect dynamic tension data of railway contact network bearing cables under different environmental conditions based on factors affecting train operation by using situational awareness technology to obtain a tension data stream, wherein the factors affecting train operation include weather conditions and train operation density; An analysis module, used for applying a hidden Markov model to analyze the tension change pattern in the tension data stream to obtain a preliminary analysis result, and using a convolutional neural network and a graph neural network to process the preliminary analysis result to generate an abnormal pattern recognition result; A prediction module is used to perform risk assessment based on the abnormal pattern recognition result using a Bayesian network algorithm and digital twin technology to obtain a preliminary estimated fault probability, and to predict a target estimated fault probability of a fault occurring within a specific time period by combining multi-context perception data and the preliminary estimated fault probability, wherein the multi-context perception data includes historical maintenance records and weather forecast information; A generation module is used to dynamically adjust the preset alarm threshold value based on the target fault estimation probability and the multi-scenario perception data by using an adaptive control algorithm to obtain an adjusted alarm threshold value. When the tension value monitored from the tension data stream exceeds the adjusted alarm threshold value, an early warning message is immediately generated, and the early warning message is processed through blockchain technology to obtain target early warning information.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an online monitoring and early warning method for the tension of a railway contact network load-bearing cable as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, an online monitoring and early warning method for the tension of a railway contact network load-bearing cable as described in any one of claims 1 to 7 is implemented.

Images