Turnout outdoor switching state monitoring and analyzing system based on large model

Abstract

The invention relates to a turnout outdoor switching state monitoring and analyzing system based on a large model, in particular to the field of rail transit, and particularly comprises a data acquisition module, a data processing module, a fault prediction module and a visual interface. The monitoring precision and the fault prediction capability are obviously improved; compared with a traditional method, the large model can identify the turnout state more accurately, predict faults in advance and adapt to different working condition changes; besides, in combination with the Internet of Things, edge computing and digital twinning technologies, the large model realizes efficient and intelligent turnout monitoring and management, and the accuracy and timeliness of fault diagnosis and prediction are optimized.

Description

Technical Field

[0001] The present invention relates to the field of rail transportation, and more specifically, to a large-model-based outdoor switch state monitoring and analysis system. Background Art

[0002] Turnouts are vital infrastructure in rail transit systems. They undertake the key task of guiding trains to switch from one track to another. They are widely used in railway marshaling yards, freight stations, subway stations and stations. They are the core equipment for achieving efficient train operation and rational use of lines. The reliability and stability of their switching state are directly related to driving safety and transportation efficiency, and play a decisive role in the normal operation of the entire rail transit system. Traditional turnout monitoring methods mainly include manual inspections and automated monitoring systems based on simple sensors. Manual inspections can ensure the safety of turnout operation to a certain extent by regularly checking the appearance of turnout equipment, component wear, and whether the switching action is normal. However, their efficiency is low and they are easily affected by human factors, which cannot meet the needs of real-time monitoring. On the other hand, some automated monitoring systems use devices such as travel switches and pressure sensors for data collection, which can provide some information on the status of turnouts. However, the monitoring accuracy and reliability of these sensors are affected by environmental factors, resulting in poor data accuracy and difficulty in fully reflecting the actual operating conditions of turnouts. In addition, traditional monitoring methods also have problems in data processing and analysis. They usually rely on manual collation and analysis of large amounts of monitoring data, which has slow processing speeds and makes it difficult to effectively identify and predict complex failure modes.

[0003] With the rapid development of rail transit technology, the outdoor switch state monitoring and analysis system based on large models has gradually attracted attention. The system uses advanced artificial intelligence algorithms and big data analysis technology to collect turnout operation data in real time and conduct in-depth analysis, thereby improving the efficiency and accuracy of turnout monitoring. However, the existing large model monitoring system still faces some problems. First, although the large model can process a large amount of data, the system often requires higher computing resources and more complex configurations in practical applications, which makes it difficult to achieve efficient operation in some resource-constrained environments. Secondly, the large model training and optimization process in the existing system usually relies on historical data. However, the operating environment of turnouts in rail transit systems is complex and changeable, and historical data may not be sufficient to cover all potential failure modes, resulting in insufficient generalization ability of the large model and difficulty in dealing with various complex and sudden failure situations. Furthermore, although the large model system can provide more accurate monitoring results, it still has room for improvement in terms of real-time and stability. In some high-frequency operation environments, there may be a certain risk of data processing delay or misjudgment, which affects the timely discovery and disposal of faults. In addition, existing monitoring systems rely on a single sensor device or a single data source and lack the fusion and collaborative analysis of multi-source data, which may lead to misjudgment of the system when facing complex faults and fail to provide comprehensive and accurate fault warnings. Summary of the invention

[0004] In view of the technical problems existing in the prior art, the present invention provides a large-model-based outdoor switch status monitoring and analysis system for turnouts. By deeply analyzing multi-source heterogeneous data, the monitoring accuracy and fault prediction capabilities are significantly improved to solve the problems raised in the above-mentioned background technology.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: specifically comprising: a data acquisition module, a data processing module, a fault prediction module and a visual interface;

[0006] Data acquisition module: through the vibration, temperature, displacement and current sensors installed at the key parts of the turnout, various data related to the operation status of the turnout are collected in real time, and the collected data are transmitted to the data processing module;

[0007] Data processing module: pre-processes the data transmitted by the data acquisition module, then integrates the multi-source data from different sensors, and then extracts the key features reflecting the operating status of the turnout from the processed data, and sends the key feature data to the fault prediction module;

[0008] Fault prediction module: Build a large model for turnout monitoring tasks, take the acquired key feature data as input, use the trained large model to judge and predict the operating status of the turnout, and output whether the turnout is operating normally, the fault type and the fault location information. If the prediction result shows an abnormality, the early warning mechanism is automatically triggered and the early warning information is sent to the visualization interface;

[0009] Visual interface: The intuitive interface displays the operating status, fault diagnosis and prediction of turnout equipment, helping users to understand equipment conditions in a timely manner and make maintenance decisions; at the same time, it provides detailed reports and data statistics.

[0010] In a preferred embodiment, in the data acquisition module, the key parts of the turnout include the switch machine, the point rail, and the stock rail, and the vibration sensor is used to monitor the vibration of the turnout during the switching process;

[0011] The temperature sensor is used to monitor the temperature change of the turnout equipment in real time; the displacement sensor is used to accurately measure the displacement of the point rail; the current sensor is used to monitor the working current of the switch machine;

[0012] The data acquisition module adopts a transmission method combining wired and wireless.

[0013] In a preferred embodiment, in the data processing module, the preprocessing operation includes data cleaning, denoising and normalization, wherein the normalization method includes minimum-maximum normalization and Z-score normalization.

[0014] In a preferred embodiment, the data processing module uses time synchronization and data matching technology to ensure that data from different sensors are consistent in time and space.

[0015] In a preferred embodiment, the key features include time domain and frequency domain features of the vibration signal, trend features of temperature change, maximum and minimum values ​​of displacement, and amplitude and phase of current.

[0016] In a preferred embodiment, in the fault prediction module, in the large model training stage, the large model is trained and optimized using a deep learning model of the Transformer architecture in combination with a large amount of historical data and real-time monitoring data; during the training process, the stochastic gradient descent and Adam optimization algorithms are used to adjust the corresponding parameters of the large model; at the same time, regularization and data enhancement techniques are used to improve the generalization ability of the large model; in the large model optimization stage, the structure and parameters of the large model are adjusted and optimized according to the performance of the large model during the training and verification process.

[0017] In a preferred embodiment, in the fault prediction module, turnout faults specifically include loose fit of the point rail, switch machine failure, and component wear; the steps for judging the operating status of the turnout are: when the displacement data of the point rail exceeds the normal range and the duration exceeds the set threshold, the large model judges it as a loose fit of the point rail failure; when the switch machine current fluctuates abnormally and is significantly different from the normal working current curve, the large model judges that there is a fault in the switch machine; in addition, the model can also perform a comprehensive analysis of multiple sensor data, and the steps of the comprehensive analysis are: by calculating the average, maximum, and minimum statistical parameters of various performance indicators of the turnout in different time periods, and comparing and analyzing them with the standard values; judging whether the conversion efficiency of the turnout meets the requirements based on the statistical data of the conversion time; evaluating whether the mechanical properties of the turnout are good by analyzing the changes in the conversion resistance; judging whether the control accuracy and electrical performance of the turnout are normal based on the stability of the point rail displacement and the switch machine current.

[0018] In a preferred embodiment, the large model uses time series analysis and prediction algorithms to model and predict the operating parameters of the turnout, analyze its future change trends, and generate prediction results. If the prediction results show that one of the parameters exceeds the corresponding preset normal range or shows an abnormal change trend, the early warning mechanism is automatically triggered.

[0019] In a preferred embodiment, the warning mechanism adopts multiple warning modes, specifically: first, on the visual interface, the warning information is displayed in a striking red warning box, accompanied by a sound alarm and a vibration prompt to attract the user's attention;

[0020] Second, the warning information is sent to the maintenance personnel's mobile phones and mailboxes by SMS and email to ensure that they receive the notification in the first time; the warning information includes the warning type, turnout number, possible time and location of the fault, and recommended measures, providing comprehensive information support for maintenance personnel so that they can take appropriate measures in time to avoid the occurrence and expansion of faults;

[0021] In addition, in order to further improve the accuracy and reliability of fault warnings, the fault prediction module also combines historical fault data and expert experience to continuously optimize and adjust the warning model. By analyzing historical fault cases, summarizing the characteristics and patterns before the fault occurs, and integrating this knowledge into the large model, the large model can more accurately predict potential faults; cooperate with experts in the railway industry to obtain their professional opinions and suggestions, and optimize the warning rules and thresholds.

[0022] In a preferred embodiment, the visual interface displays the operating status of the turnout equipment, including real-time curves of data such as vibration, temperature, displacement, current, etc., as well as the current position and working status of the turnout. Through this function, the user can understand the operating status of the turnout in real time and find abnormalities in time;

[0023] The reports include daily, monthly and annual reports on turnout operation status and fault statistics.

[0024] In a preferred embodiment, the method specifically includes the following steps:

[0025] The beneficial effects of the present invention are: the large model of the present application significantly improves the monitoring accuracy and fault prediction ability in rail transit turnout monitoring by deeply analyzing multi-source heterogeneous data; compared with traditional methods, the large model can more accurately identify the turnout status, predict faults in advance and adapt to changes in different working conditions; in addition, combined with the Internet of Things, edge computing and digital twin technology, the large model realizes efficient and intelligent turnout monitoring and management, and optimizes the accuracy and timeliness of fault diagnosis and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of the method of the present invention;

[0027] Figure 2 This is a system structure block diagram of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0029] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0030] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0031] Example 1

[0032] This embodiment provides Figure 1-2 A large-model-based outdoor switch state monitoring and analysis system is shown, which specifically includes: a data acquisition module, a data processing module, a fault prediction module and a visualization interface;

[0033] Data acquisition module: Vibration, temperature, displacement and current sensors installed at key parts of the turnout are used to collect various data related to the turnout operation status in real time, and the collected data is transmitted to the data processing module to provide a basis for subsequent analysis;

[0034] Data processing module: pre-processes the data transmitted by the data acquisition module, then integrates the multi-source data from different sensors, fully explores the correlation information between the data, improves the comprehensiveness and reliability of the data, and then extracts the key features that can reflect the running status of the turnout from the processed data to improve the quality and availability of the data, provide high-quality data for the model layer, and send the key feature data to the fault prediction module;

[0035] Fault prediction module: Build a large model for turnout monitoring tasks, take the acquired key feature data as input, use the trained large model to judge and predict the operating status of the turnout, and output whether the turnout is operating normally, the fault type and the fault location information. If the prediction result shows an abnormality, the early warning mechanism is automatically triggered and the early warning information is sent to the visualization interface;

[0036] Visual interface: The intuitive interface displays the operating status, fault diagnosis and prediction of turnout equipment, helping users to understand the equipment status in a timely manner and make maintenance decisions; at the same time, it provides detailed reports and data statistics to ensure efficient and safe equipment management.

[0037] In this embodiment, the data acquisition module is specifically required to be explained. The key parts of the turnout include the switch machine, the point rail, and the stock rail. The vibration sensor is used to monitor the vibration of the turnout during the switching process. By analyzing the frequency, amplitude, phase and other characteristics of the vibration signal, it can be determined whether the turnout is loose, worn or stuck. Abnormal conditions;

[0038] The temperature sensor is used to monitor the temperature change of the turnout equipment in real time, because too high a temperature may indicate problems such as equipment overload, excessive friction or poor heat dissipation; the displacement sensor is used to accurately measure the displacement of the point rail to ensure that the point rail can be accurately in place during the switching process to avoid faults such as the point rail not being closely attached; the current sensor is used to monitor the working current of the switch machine, and the change in current can reflect the load condition and working status of the switch machine;

[0039] In terms of sensor layout, the structural characteristics of the turnout and the key parameters during operation are fully considered, and sensors are installed in the following locations:

[0040] 1) Switch machine: As the power source for turnout switching, the working state of the switch machine directly affects the switching effect of the turnout; a current sensor is installed on the switch machine to monitor the current change when the switch machine is working; the current signal can reflect the load condition of the switch machine, the running state of the motor, and whether there is a problem such as jamming; when the switch machine encounters a large resistance, the current will increase significantly. By monitoring the abnormal change of the current, the hidden dangers of the switch machine can be discovered in time; a temperature sensor is installed inside the switch machine to monitor the temperature of the switch machine in real time; excessive temperature may mean that the switch machine is overloaded, the heat dissipation is poor, or the wear of the internal components is aggravated. By monitoring the temperature, the potential failure of the switch machine can be warned in advance;

[0041] 2) Point rail: The displacement and close fitting state of the point rail are key indicators for the normal operation of the turnout. A displacement sensor is installed on the point rail to accurately measure the displacement change of the point rail during the conversion process to ensure that the point rail can be accurately positioned to avoid the situation where the point rail is not close fitting. The close fitting of the point rail will affect the driving safety of the train. By real-time monitoring of the displacement of the point rail, the position deviation of the point rail can be discovered and corrected in time. A pressure sensor is installed at the joint of the point rail and the base rail to detect the close fitting pressure between the point rail and the base rail. The appropriate close fitting pressure can ensure the close fit between the point rail and the base rail to prevent dangerous situations such as shaking or derailment when the train passes. By monitoring the changes in the close fitting pressure, it can be determined whether the close fitting state between the point rail and the base rail is normal.

[0042] 3) Basic rail: Basic rails are subjected to huge pressure and impact force when trains pass through them, and their vibration can reflect the overall stability of the turnout. Vibration sensors are installed on basic rails to monitor the vibration signals of basic rails during train passing and turnout switching. The frequency, amplitude and phase characteristics of vibration signals can reflect whether the turnout is loose, worn or in other abnormal conditions. By analyzing the vibration signals, potential faults of the turnout can be discovered in time to ensure the safe operation of the train.

[0043] The data acquisition module adopts a transmission method that combines wired and wireless. For sensors that are closer to the data processing center, such as switch machine sensors installed in the machine room near the turnout, wired transmission methods such as RS485 bus and Ethernet are adopted. Wired transmission has the advantages of stable transmission, strong anti-interference ability, and high data transmission rate, which can ensure the real-time and accuracy of data. For some sensors with more remote installation locations or difficult wiring, such as sensors installed on point rails and base rails, wireless transmission methods such as ZigBee and LoRa are adopted. Wireless transmission methods have the advantages of easy installation, high flexibility, and low cost, and can adapt to complex on-site environments. In order to ensure the reliability of data transmission, data verification, retransmission mechanism and other technologies are adopted in the wireless transmission process to ensure the integrity and accuracy of the data. At the same time, the data collected by the sensor is marked in real time, including information such as collection time, sensor location, turnout number, etc., for subsequent data processing and analysis.

[0044] Since the raw data collected by the data acquisition module have problems such as noise, outliers and inconsistent data formats, which will affect the subsequent data analysis and model training effects, it is necessary to preprocess the collected data to improve the quality and availability of the data. In the data processing module, the preprocessing operations include data cleaning, denoising and normalization, where the normalization methods include minimum-maximum normalization and Z-score normalization;

[0045] Among them, data cleaning is mainly used to remove noise and outliers in the data; during the data collection process, due to the accuracy limitations of the sensor, environmental interference, equipment failure and other reasons, the collected data may contain noise and outliers; these noise and outliers will interfere with the data analysis results, so cleaning is required; this application adopts a method based on statistical analysis, such as the 3σ criterion, to clean the data; the 3σ criterion means that the data should obey the normal distribution. In the normal distribution, the probability that the data falls within the range of the mean plus or minus 3 times the standard deviation is 99.7%, so the data outside this range is regarded as an outlier and is eliminated; for some obviously erroneous data, such as negative values ​​output by the sensor (data that cannot be negative in the actual physical sense), corresponding corrections or elimination are also performed;

[0046] Denoising is an important part of data preprocessing, and its purpose is to remove noise from data and improve the signal-to-noise ratio of data. In this application, filtering algorithms are mainly used to denoise data; for data containing high-frequency noise such as vibration signals, low-pass filters are used to remove high-frequency noise components and retain low-frequency valid signals; low-pass filters can set cutoff frequencies and only allow signals below the cutoff frequency to pass, thereby achieving the purpose of removing high-frequency noise; for some data interfered by random noise, median filtering algorithms are used; median filtering is a nonlinear filtering method that replaces the value of each data point with the median value of the data points in its neighborhood, which can effectively remove random noise while retaining the edge and detail information of the data;

[0047] Normalization is to unify data with different features into the same scale range to eliminate the dimensional differences between data features and improve the training effect and convergence speed of the model. In this application, the minimum-maximum normalization method is mainly used to map the data to the [0,1] interval. The formula for minimum-maximum normalization is:

[0048]

[0049] Among them, x represents the original data, x min and x max Respectively represent the minimum and maximum values ​​of the data, x norm represents the normalized data; in addition, for some data with special distribution, this application adopts the Z-score normalization method, and the specific formula of the Z-score normalization is:

[0050]

[0051] Among them, μ represents the mean of the data, σ represents the standard deviation of the data, and z represents the normalized data. The Z-score normalization method converts the data into a standard normal distribution with a mean of 0 and a standard deviation of 1, which can make the data more stable and comparable.

[0052] In this embodiment, it is specifically necessary to explain that the data processing module uses time synchronization and data matching technology to make the data of different sensors consistent in time and space;

[0053] The key features include the time domain and frequency domain characteristics of the vibration signal, the trend characteristics of temperature changes, the maximum and minimum displacements, and the amplitude and phase of the current. The key features are used as inputs to the model in the fault prediction module for training and reasoning. The data processing module is also responsible for storing the processed data in the database for subsequent query and analysis. The database uses distributed storage technology to ensure the security and reliability of the data, while supporting efficient data reading and writing operations.

[0054] In this embodiment, it is specifically necessary to explain the fault prediction module. In the large model training stage, the deep learning model of the Transformer architecture is used, and a large amount of historical data and real-time monitoring data are combined to train and optimize the large model; in the training process, random gradient descent and Adam optimization algorithms are used to adjust the corresponding parameters of the large model, so that the large model can accurately learn the mapping relationship between the turnout operation status and the monitoring data. Among them, the Adam optimization algorithm used in this application can automatically adjust the learning rate according to the size of the gradient, use a larger learning rate in the early stage of training to speed up the convergence speed, and gradually reduce the learning rate in the later stage of training to improve the accuracy of the large model. The hyperparameters of the large model, such as the number of layers, the number of heads, and the hidden layer dimension, are also adjusted and optimized; at the same time, in order to prevent the large model from overfitting, the regularization is used. , data enhancement technology, to improve the generalization ability of the large model. Among them, due to the relative scarcity of turnout fault data, the large model may be overfitted if it is trained solely on the actual collected data, and it cannot show good performance on the new data. The data enhancement technology of this application is to generate new training data by performing a series of transformations on the original data, such as translation, rotation, scaling, and noise addition, so as to expand the scale and diversity of the training data set. When processing the vibration signal of the turnout, the vibration signal is randomly time-shifted and amplitude-scaled to generate new vibration signal data. When processing the image data of the turnout, the image is rotated, cropped, and noise is added to increase the diversity of the image data, so that the large model can learn more data features and patterns, and improve the large model's ability to recognize the turnout status under different working conditions.

[0055] Regularization techniques mainly include L2 regularization (also known as weight decay) and Dropout regularization. L2 regularization refers to adding a penalty term proportional to the sum of squares of model parameters in the loss function to force the model to learn smaller parameters, thereby reducing the complexity of the model and preventing overfitting. Dropout regularization refers to randomly setting the output of a part of neurons to 0 during model training, which is equivalent to using a different sub-model each time training, thereby reducing the co-adaptation phenomenon between neurons and improving the generalization ability of the model. In the Transformer model, Dropout regularization is applied to the fully connected layer and the attention layer. By setting the Dropout probability, a certain proportion of neuron connections are randomly discarded during the training process to make the large model more robust. In the large model optimization stage, the structure and parameters of the large model are adjusted and optimized according to the performance of the large model during training and verification. The optimal large model configuration can be selected by comparing the performance indicators under different model structures and parameter settings. In addition, transfer learning and model fusion technology can be used to further improve the performance of the large model. Transfer learning refers to transferring the model parameters trained on one task to another related task for fine-tuning, thereby speeding up the training speed of the new task and improving the model performance. In the turnout monitoring task, since collecting and annotating a large amount of turnout fault data is a time-consuming and costly task, the Transformer model that has been trained in other related fields (such as mechanical fault diagnosis, industrial equipment monitoring, etc.) can be used to transfer its parameters to the turnout monitoring model. By fine-tuning on a small amount of turnout monitoring data, the model can quickly adapt to the turnout monitoring task, and the general features and patterns that have been learned by the pre-trained model can be used to reduce the training time and computing resources of the large model on the turnout monitoring data, while improving the generalization ability and accuracy of the model.

[0056] Turnout faults specifically include loose fitting of the point rail, switch machine failure, and component wear; the steps for judging the operating status of the turnout are: when the displacement data of the point rail exceeds the normal range and the duration exceeds the set threshold, the large model judges it as a loose fitting of the point rail fault; when the switch machine current fluctuates abnormally, such as the current is too large or too small, and is significantly different from the normal working current curve, the large model judges that the switch machine has a fault; in addition, the model can also conduct a comprehensive analysis of multiple sensor data, and the steps of the comprehensive analysis are: by calculating the average, maximum, and minimum statistical parameters of various performance indicators of the turnout in different time periods, and comparing and analyzing them with the standard values; judging whether the conversion efficiency of the turnout meets the requirements based on the statistical data of the conversion time; evaluating whether the mechanical properties of the turnout are good by analyzing the changes in the conversion resistance; judging whether the control accuracy and electrical performance of the turnout are normal based on the stability of the point rail displacement and the switch machine current; in addition, a comprehensive score is given to the performance of the turnout to intuitively reflect the overall performance of the turnout;

[0057] The big model uses time series analysis and prediction algorithms to model and predict the operating parameters of the turnout, analyze its future change trend, and generate prediction results. If the prediction results show that one of the parameters exceeds the corresponding preset normal range or shows an abnormal change trend, the early warning mechanism is automatically triggered;

[0058] The early warning mechanism adopts a variety of early warning methods, specifically: ensuring that the early warning information can be conveyed to relevant personnel in a timely manner; first, displaying the early warning information in a striking red warning box on the visual interface, accompanied by sound alarms and vibration prompts to attract the user's attention;

[0059] Second, the warning information is sent to the maintenance personnel's mobile phones and mailboxes by SMS and email to ensure that they receive the notification in the first time; the warning information includes the warning type, turnout number, possible time and location of the fault, and recommended measures, providing comprehensive information support for maintenance personnel so that they can take appropriate measures in time to avoid the occurrence and expansion of faults;

[0060] In addition, in order to further improve the accuracy and reliability of fault warnings, the fault prediction module also combines historical fault data and expert experience to continuously optimize and adjust the warning model. By analyzing historical fault cases, summarizing the characteristics and patterns before the fault occurs, and integrating this knowledge into the large model, the large model can more accurately predict potential faults; cooperate with experts in the railway industry to obtain their professional opinions and suggestions, optimize the warning rules and thresholds, and ensure that the large model can play the greatest role in practical applications.

[0061] In this embodiment, it is specifically necessary to explain the visual interface, which displays the operating status of the turnout equipment including real-time curves of data such as vibration, temperature, displacement, current, etc., as well as the current position and working status of the turnout. Through this function, the user can understand the operating status of the turnout in real time and find abnormalities in time;

[0062] Reports include daily, monthly and annual reports on turnout operation status and fault statistics. These reports provide users with detailed data statistics and analysis results, which help users understand the operation status and maintenance requirements of turnout equipment and provide a basis for decision-making; the user management function is used to manage the user information and permissions of the system to ensure the security of the system and the confidentiality of data; the system supports multi-user login and assigns different operating permissions according to the user's role and responsibilities, such as viewing data, modifying parameters, and performing maintenance tasks.

[0063] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0064] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0065] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0066] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0067] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0068] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0069] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A large model-based outdoor switch state monitoring and analysis system, characterized in that: Specifically include: Data acquisition module, data processing module, fault prediction module and visualization interface; Data acquisition module: through the vibration, temperature, displacement and current sensors installed at the key parts of the turnout, various data related to the operation status of the turnout are collected in real time, and the collected data are transmitted to the data processing module; Data processing module: pre-processes the data transmitted by the data acquisition module, then integrates the multi-source data from different sensors, and then extracts the key features reflecting the operating status of the turnout from the processed data, and sends the key feature data to the fault prediction module; Fault prediction module: Build a large model for turnout monitoring tasks, take the acquired key feature data as input, use the trained large model to judge and predict the operating status of the turnout, and output whether the turnout is operating normally, the fault type and the fault location information. If the prediction result shows an abnormality, the early warning mechanism is automatically triggered and the early warning information is sent to the visualization interface; Visual interface: The intuitive interface displays the operating status, fault diagnosis and prediction of turnout equipment, helping users to understand equipment conditions in a timely manner and make maintenance decisions; at the same time, it provides detailed reports and data statistics.

2. The large model-based outdoor switch state monitoring and analysis system according to claim 1 is characterized in that: In the data acquisition module, the key parts of the turnout include the switch machine, the point rail, and the stock rail, and the vibration sensor is used to monitor the vibration of the turnout during the switching process; The temperature sensor is used to monitor the temperature change of the turnout equipment in real time; the displacement sensor is used to accurately measure the displacement of the point rail; the current sensor is used to monitor the working current of the switch machine; The data acquisition module adopts a transmission method combining wired and wireless.

3. The large model-based outdoor switch state monitoring and analysis system according to claim 2 is characterized in that: In the data processing module, the preprocessing operations include data cleaning, denoising and normalization, wherein the normalization methods include minimum-maximum normalization and Z-score normalization.

4. The large model-based outdoor switch state monitoring and analysis system according to claim 3 is characterized by: The data processing module uses time synchronization and data matching technology to make the data of different sensors consistent in time and space.

5. The large model-based outdoor switch state monitoring and analysis system according to claim 4 is characterized in that: The key features include the time domain and frequency domain features of the vibration signal, the trend features of the temperature change, the maximum and minimum values ​​of the displacement, and the amplitude and phase of the current.

6. The large model-based outdoor switch state monitoring and analysis system according to claim 5 is characterized in that: In the fault prediction module, during the large model training stage, the deep learning model of the Transformer architecture is used, and a large amount of historical data and real-time monitoring data are combined to train and optimize the large model; during the training process, the stochastic gradient descent and Adam optimization algorithms are used to adjust the corresponding parameters of the large model; at the same time, regularization and data enhancement techniques are used to improve the generalization ability of the large model; in the large model optimization stage, the structure and parameters of the large model are adjusted and optimized according to the performance of the large model in the training and verification process.

7. The large model-based outdoor switch state monitoring and analysis system according to claim 6 is characterized by: In the fault prediction module, turnout faults specifically include loose fit of the point rail, switch machine failure, and component wear; the steps for judging the operating status of the turnout are: when the displacement data of the point rail exceeds the normal range and the duration exceeds the set threshold, the large model judges it as a loose fit of the point rail; when the switch machine current fluctuates abnormally and is significantly different from the normal working current curve, the large model judges that there is a fault in the switch machine; in addition, the model can also perform a comprehensive analysis of multiple sensor data, and the steps of the comprehensive analysis are: by calculating the average, maximum, and minimum statistical parameters of various performance indicators of the turnout in different time periods, and comparing and analyzing them with the standard values; judging whether the conversion efficiency of the turnout meets the requirements based on the statistical data of the conversion time; evaluating whether the mechanical properties of the turnout are good by analyzing the changes in the conversion resistance; judging whether the control accuracy and electrical performance of the turnout are normal based on the stability of the point rail displacement and the switch machine current.

8. The large model-based outdoor switch state monitoring and analysis system according to claim 7 is characterized in that: The large model uses time series analysis and prediction algorithms to model and predict the operating parameters of the turnout, analyze its future change trends, and generate prediction results. If the prediction results show that one of the parameters exceeds the corresponding preset normal range or shows an abnormal change trend, the early warning mechanism is automatically triggered.

9. The large model-based outdoor switch state monitoring and analysis system according to claim 8 is characterized in that: The warning mechanism adopts a variety of warning methods, specifically: first, on the visual interface, the warning information is displayed in a striking red warning box, accompanied by a sound alarm and a vibration prompt to attract the user's attention; Second, the warning information is sent to the maintenance personnel's mobile phones and mailboxes by SMS and email to ensure that they receive the notification in the first time; the warning information includes the warning type, turnout number, possible time and location of the fault, and recommended measures, providing comprehensive information support for maintenance personnel so that they can take appropriate measures in time to avoid the occurrence and expansion of faults; In addition, in order to further improve the accuracy and reliability of fault warnings, the fault prediction module also combines historical fault data and expert experience to continuously optimize and adjust the warning model. By analyzing historical fault cases, summarizing the characteristics and patterns before the fault occurs, and integrating this knowledge into the large model, the large model can more accurately predict potential faults; cooperate with experts in the railway industry to obtain their professional opinions and suggestions, and optimize the warning rules and thresholds.

10. The large model-based outdoor switch state monitoring and analysis system according to claim 9 is characterized in that: The visual interface displays the operating status of the turnout equipment, including real-time curves of data such as vibration, temperature, displacement, current, etc., as well as the current position and working status of the turnout. Through this function, users can understand the operating status of the turnout in real time and discover abnormalities in time; The reports include daily, monthly and annual reports on turnout operation status and fault statistics.

Images