Railway project completion delivery data information extraction method based on digital twinning

Abstract

The invention provides a railway engineering completion delivery data information extraction method based on digital twinning. Relates to the technical field of engineering data processing. The technical key point of the invention lies in how to construct a dynamic digital twinborn model by integrating various technologies, such as BIM, Internet of Things, big data, AI and the like, and in combination with technologies, such as real-time data monitoring, intelligent analysis, data sharing and the like, the quality control, operation optimization, decision support and completion delivery efficiency of a railway construction project is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering data processing, and particularly to a method for extracting railway engineering completion and delivery data information based on digital twin. Background Art

[0002] The completion and delivery of railway projects is a key link in the railway construction process, involving a large amount of complex engineering data, design drawings, construction records, equipment configurations, and other information. Currently, data extraction mainly relies on manual operations, facing many challenges. On the one hand, manual data extraction is inefficient, time-consuming, and prone to human errors, resulting in data errors and affecting the overall progress of completion and delivery. On the other hand, due to the subjectivity of manual operations, data accuracy is difficult to guarantee, which has an adverse impact on the subsequent operation and maintenance of railway projects. In addition, there are obstacles to data sharing and collaboration among participating construction units, and the phenomenon of information islands is serious, further increasing the complexity of information processing and management difficulty.

[0003] In this context, digital twin technology has become an effective way to solve the problem of railway construction data management, especially in the completion and delivery stage. Digital twin can significantly improve the efficiency and accuracy of data processing, ensuring data integrity and consistency. Through real-time data exchange and feedback between digital models and physical entities, the state, behavior, and changes of physical systems can be accurately restored. In the railway field, digital twin technology can accurately model railway equipment, facilities, and environments in the virtual space, and provide high-quality data support for railway management and operation and maintenance through means such as real-time monitoring, data analysis, and simulation prediction.

[0004] Although the application of digital twin technology in the railway industry has made certain progress, there are still technical problems in the process of completion and delivery data extraction and management. For example, problems such as inconsistent data standards, missing information extraction, and unclear data structures have not been effectively solved. Digital twin technology can greatly improve the efficiency and accuracy of data extraction through accurate modeling, real-time data feedback, and intelligent analysis, promote information sharing and collaboration among different participating construction units, and thus optimize the construction, management, and operation and maintenance processes of railway projects. Summary of the Invention

[0005] The purpose of the present invention is to solve the above problems and provide a method for extracting railway engineering completion and delivery data information based on digital twin.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows: According to one aspect of the present invention, a method for extracting railway engineering completion and delivery data information based on digital twin is provided, and the specific steps are as follows: S1: Digital twin model construction (1)Collect data during railway engineering design, construction, acceptance and other stages (2)Use 3D modeling technology to construct a railway engineering entity model After data collection is completed, integrate design and construction data through BIM and CAD technologies, and use Revit software for modeling. The data is converted into precise geometric data for the BIM model through point cloud processing software; (3)Combine technologies such as the Internet of Things and big data to add attribute information to the entity model to form a digital twin model Based on the entity model, form a complete digital twin model by updating the virtual model with real-time data; Install sensors for temperature, vibration, and stress on key facilities. The real-time data is transmitted to the cloud platform through a wireless network for processing. After the data collected by the sensors is preprocessed, it is synchronized with the BIM model in real time. Use big data analysis technology to process the sensor data. Through data mining and pattern recognition, use machine learning models to predict possible fault trends; S2: Real-time collect data information during the completion and delivery process (1)Use devices including sensors and cameras to monitor the railway engineering site in real time; (2)Transmit the monitoring data to the digital twin model through a data transmission network The data collected by sensors and cameras is transmitted to the cloud platform in real time through a wireless network. The cloud platform receives this data using Internet of Things protocols and applies it to update relevant information in the digital twin model. Whenever the on-site monitoring data changes, the model will automatically adjust to reflect the latest on-site status; After the monitoring data changes, the digital twin model automatically updates relevant information mainly through the following key steps: First is the data reception and parsing link. The cloud platform receives the data from sensors and cameras according to Internet of Things protocols and disassembles the original data according to the preset format to obtain data information including data type, source, and value, which serves as the basic data for subsequent update operations; Then, conduct data verification and preprocessing. In terms of data verification, it is necessary to check the data integrity to ensure that all parameters are successfully received; verify the accuracy, judge whether the data is credible according to the set data range and verification rules. For example, temperature data needs to be within a reasonable physical range; it is also necessary to verify the timeliness to ensure that the data is collected and transmitted at appropriate time intervals; for data preprocessing, if there is missing data, interpolate and supplement it with reference to historical trends and surrounding relevant data. For abnormal data, use smoothing algorithms or correct and mark it according to statistical laws.

[0007] Then comes the model association and positioning step. Based on the data source identifier, the corresponding model components or sub-models are accurately found in the digital twin model; the regional camera image data corresponds to the corresponding spatial model part; Subsequently, it enters the model parameter update stage, directly updating the processed monitoring data to the corresponding device parameters in the model, making the model consistent with the actual situation in this parameter dimension; In terms of model visualization update, first, the geometry and appearance are updated. According to the physical state changes reflected by the monitoring data, the geometry and appearance of the model are adjusted. For example, when the building structure deforms, the corresponding distortion or displacement is shown in the model, and when the temperature of the device shell is too high, the color of the model device is changed to display a high-temperature warning. Second, the animation and dynamic effects are updated. For the monitoring data with dynamic behaviors, such as the operation status of the device, the movement trajectories of personnel or objects, animation effects are added to the model; Finally, it is the model consistency check and optimization. During the consistency check, after the parameter and visualization updates are completed, check whether the logical relationships and physical constraints of each part of the model are correct, and compare whether the changes in the model state before and after the update conform to the physical laws and expected system behaviors. If there are any abnormalities, trace back in time to check for errors in the data processing or model update process. Model optimization adjusts the model parameter calculation algorithm according to multiple monitoring data updates and model operation conditions to improve efficiency and accuracy, simplifies or refines the model structure to better adapt to the characteristics of monitoring data changes and application requirements. At the same time, based on big data analysis, mine the monitoring data and the historical data of model operation, discover new correlation relationships or laws and incorporate them into the model construction or update mechanism to enhance the simulation and prediction capabilities of the digital twin model for the actual system.

[0008] S3: Extract, analyze, and process data information (1) Set data extraction rules according to the completion and delivery standards To ensure that the monitoring data meets the project acceptance standards, first, specific data extraction rules are formulated according to industry specifications. Through the preset rules, the platform can automatically screen the real-time data that meets the requirements; (2) Use technologies such as machine learning and deep learning to extract features from the monitoring data Use a convolutional neural network to extract key features related to structural health from the signals of vibration sensors. Through the training of a large amount of data, the machine learning model can identify the features related to structural damage or faults, and use a supervised learning model to train the monitoring data; (3) Combine with the digital twin model to conduct correlation analysis on the feature data and realize the intelligent extraction of data information; ① Set data extraction rules: Formulate specific rules according to the completion and delivery standards and industry specifications, covering monitoring data including quality, accuracy, and time frequency; ② Feature extraction of the detection data: For the data collected by vibration sensors and the like, a suitable convolutional neural network (CNN) model architecture is constructed. The CNN model includes convolutional layers, pooling layers, and fully connected layer components, and can automatically learn local and global features in the data. Using a supervised learning algorithm, the preprocessed and feature-extracted data is trained to construct an anomaly detection model. During the training process, the data labeled as normal state and abnormal state are respectively input into the model as positive and negative samples, enabling the model to learn the differences in data features between the two states.

[0009] ③ Correlation analysis of feature data and intelligent extraction of information: Combining with the digital twin model, using the relationship between temperature data and track stress, it is found that an increase in temperature may lead to situations affecting railway safety such as track deformation. By performing time-series analysis on the collected sensor data, the data change trends at different time nodes are revealed; using clustering analysis technology, similar data patterns are classified to identify potential structural risks or construction progress issues; S4: Output accurate completion and delivery data information (1) Summarize and organize the extracted data information Classify the data and convert it into a unified format to ensure that the report can intuitively display key data; subsequently, summarize all monitoring data into a comprehensive completion data report; (2) Generate a completion and delivery data report Based on the data summarized in the early stage, generate a detailed completion and delivery report. Through clear data display and chart analysis, ensure that the report can provide strong evidence for project acceptance and at the same time provide reference for subsequent operation and maintenance and management.

[0010] Further, the data materials in step 1 are divided into three stages: design, construction, and acceptance: Design stage: Use BIM technology to digitally model railway line planning, station layout, bridge and tunnel design. Through data docking of design software, collect detailed information including structure, pipelines, and electricity. All design drawings, material specifications, and structure requirement data are digitized and stored in the cloud; Construction stage: Install various Internet of Things sensors at the construction site and record data such as progress, quality control, and construction drawing changes through the construction management system. In addition, use GPS and drone technology to obtain the progress and changes at the construction site in real time; Acceptance stage: Use laser scanning technology and drones to perform precise on-site data collection to verify whether the completed structure meets the design requirements. The on-site detection data will be input into the system for final review.

[0011] Compared with the prior art, the present invention has the following beneficial effects: (1)Improve data extraction efficiency: The extraction of completed and delivered data in traditional railway engineering relies on manual operations, resulting in low data extraction efficiency and long time consumption. Based on digital twin technology, this invention constructs a model and combines devices such as sensors and cameras to collect data in real time. Through a series of automated processes, the model can be updated in a timely manner and the required data can be extracted, thus greatly improving the speed of data extraction and effectively solving the problem of low efficiency in manual extraction.

[0012] (2)Ensure data accuracy: Manual operations are highly subjective, making it difficult to guarantee data accuracy. This invention precisely collects data by installing sensors on key facilities and transmits the data to the cloud platform using Internet of Things protocols. Strictly control data quality in key links such as data reception and parsing, verification and preprocessing, including checking data integrity, accuracy, timeliness, and handling missing and abnormal data. In addition, combined with digital twin models for correlation analysis and intelligent extraction, it ensures the accuracy of the data and provides solid and reliable data support for the subsequent operation and maintenance of railway projects.

[0013] (3)Promote data sharing and collaboration: In the past, there were obstacles to data sharing and collaboration among various participating units, resulting in frequent information silos. This invention builds a data sharing platform, integrating data from all parties such as design, construction, supervision, and owner units, covering multi-dimensional information such as technical monitoring, project progress, cost, and quality, enabling each participating unit to access and share data in real time, promoting information exchange and collaborative work. At the same time, it also effectively reduces the complexity of information processing and management difficulty. Specific implementation methods

[0014] To make the technical means, creative features, achieved purposes, and functions of this invention easy to understand, the following further elaborates on this invention in combination with specific implementation methods.

[0015] This invention provides a method for extracting completed and delivered data information based on digital twins, aiming to solve problems such as low efficiency, poor accuracy, and incomplete information in the process of extracting completed and delivered data for railway projects. Its core technical solutions include the following aspects: 1 Digital twin model construction Based on the full-life cycle data of railway construction projects, construct a detailed digital twin model, including not only the geometric shape and layout structure of railway facilities, but also combining the real-time operation data collected by sensors to form an integrated virtual environment model.

[0016] (1)Collect data materials in the design, construction, acceptance and other stages of railway engineering In different stages of railway construction projects, a large amount of data collection and integration are involved. In the design stage, data such as line planning, bridge and tunnel design, and station layout are mainly collected and usually digitally stored through BIM (Building Information Modeling) and CAD technologies. In the construction stage, the focus of data collection includes construction progress, quality control, construction drawings, and on-site real-time data, which are usually supported by Internet of Things sensors and construction management systems. Finally, in the acceptance stage, on-site inspection data are mainly collected through technologies such as drones and laser scanning to evaluate the performance and quality of equipment and ensure that the project meets the design standards.

[0017] (2) Use 3D modeling technology to construct a physical model of the railway project Based on the data collected in the early stage, it is transformed into a virtual entity model of digital twin through 3D modeling technology. Building Information Modeling (BIM) is widely used in the modeling of railway projects and can accurately represent various components including railway tracks, stations, tunnels, and bridges, and integrate design information from different specialties. At the same time, CAD technology can further refine the modeling of specific structural parts. To improve the accuracy and authenticity of the model, point cloud data is obtained through laser scanning or drone technology to enhance the details of the 3D model. Finally, the data in the design and construction stages are integrated together to form a complete virtual entity model of digital twin.

[0018] (3) Combine technologies such as the Internet of Things and big data to add attribute information to the physical model to form a digital twin model The digital twin model is not only a static 3D representation but also needs to be dynamically updated in combination with real-time data. After deploying sensors on key facilities, data such as temperature, stress, and vibration are collected in real time, and these data are uploaded to the cloud platform for processing through Internet of Things technology. Using big data analysis technology, a large amount of sensor data is analyzed to achieve fault prediction and operation optimization. The digital twin platform combines these real-time data with the BIM model to form a continuously updated virtual model, providing real-time monitoring, fault warning, and decision-making support for managers.

[0019] 2 Real-time collection of data information during the completion and delivery process (1) Use devices such as sensors and cameras to monitor the railway project site in real time In the completion and delivery stage of railway projects, real-time monitoring of on-site data is crucial for ensuring project quality and progress. By deploying various types of sensors (such as temperature, stress, vibration, etc.) and devices such as cameras, key parameters of the construction site can be comprehensively monitored. The sensors can perceive changes in infrastructure such as tracks, bridges, and tunnels in real time to ensure the safety and stability of the structure; while the cameras provide real-time video images to help managers remotely monitor the construction progress, discover problems in a timely manner, and intervene.

[0020] (2) Transmit the monitoring data to the digital twin model through the data transmission network To centrally manage and analyze real-time monitoring data, it is necessary to upload the information collected on-site to the digital twin model through the data transmission network. Through Internet of Things (IoT) technology, the data collected by sensors and cameras will be transmitted to the cloud platform or local data center via wireless networks (such as 5G, Wi-Fi, etc.). On this basis, the digital twin platform can update the virtual model in real time to ensure its synchronization with the on-site situation. The timely transmitted data provides accurate on-site feedback for project managers and supports scientific decision-making.

[0021] When the monitoring data changes, the digital twin model automatically updates the relevant information mainly through the following key steps: First is the data reception and parsing link. The cloud platform receives the data transmitted by sensors and cameras according to the IoT protocol, disassembles the original data according to the preset format, and obtains key information such as data type (temperature, pressure, etc.), source (specific sensor or camera identifier), and value, which serves as the basic data for subsequent update operations.

[0022] Then comes data verification and preprocessing. In terms of data verification, it is necessary to check the data integrity to ensure that all parameters are successfully received; verify the accuracy, judge whether the data is credible according to the set data range and verification rules, for example, the temperature data needs to be within a reasonable physical range; it is also necessary to verify the timeliness to ensure that the data is collected and transmitted at an appropriate time interval. For data preprocessing, if there is missing data, interpolate and supplement it with reference to historical trends and surrounding relevant data. For example, when temperature data is missing, it can be estimated by combining the data of the previous and subsequent moments and other relevant sensors in the same area; for abnormal data, such as jumps caused by sensor failures, use smoothing algorithms or correct and mark them according to statistical laws.

[0023] Then is the model association and positioning step. According to the data source identifier, accurately find the corresponding model component or sub-model in the digital twin model. For example, the sensor data from a specific device can locate the representation part of the device in the model; the regional camera image data corresponds to the corresponding spatial model part.

[0024] Subsequently, it enters the model parameter update stage. For device temperature, pressure, speed, etc., directly update the processed monitoring data to the corresponding device parameters in the model, so that the model is consistent with the actual situation in this parameter dimension. For example, when the motor speed changes from 1000 revolutions per minute to 1200 revolutions per minute, the motor speed parameter in the model is also updated to 1200 revolutions per minute.

[0025] In terms of model visualization updates, first, there are geometric shape and appearance updates. The geometric shape and appearance of the model are adjusted according to the physical state changes reflected by the monitoring data. For example, when the building structure deforms, the corresponding distortion or displacement is shown in the model; when the temperature of the equipment housing is too high, the color of the model equipment is changed to display a high-temperature warning. Second, there are animation and dynamic effect updates. For monitoring data with dynamic behaviors, such as equipment operation status, movement trajectories of personnel or objects, etc., animation effects are added to the model. For example, when a camera monitors the movement path of a person, an animation track is generated in the corresponding area of the model; when the equipment starts, a startup animation (such as a motor rotating, a conveyor belt running, etc.) is shown.

[0026] Finally, there is model consistency checking and optimization. During consistency checking, after parameter and visualization updates are completed, check whether the logical relationships and physical constraints of each part of the model are correct, and compare whether the changes in the model state before and after the update conform to physical laws and the expected system behavior. If there are any abnormalities, trace back in time to check for errors in the data processing or model update process. Model optimization adjusts the model parameter calculation algorithm based on multiple monitoring data updates and model operation conditions to improve efficiency and accuracy, simplifies or refines the model structure to better adapt to the characteristics of monitoring data changes and application requirements. At the same time, based on big data analysis of the monitoring data and the historical data of model operation, new correlation relationships or laws are discovered and incorporated into the model construction or update mechanism to enhance the simulation and prediction capabilities of the digital twin model for the actual system.

[0027] 3 Extract, analyze, and process data information (1)Set data extraction rules according to the completion and delivery standards In the completion and delivery stage of railway projects, setting data extraction rules is a crucial step to ensure that the monitoring data meets the delivery standards. These rules are based on the specific requirements of the project and industry specifications, covering aspects such as the quality, accuracy, and time frequency of the monitoring data. By clarifying the extraction rules, it is ensured that the data obtained from the real-time monitoring system meets the requirements of quality control and acceptance evaluation. In addition, the rules also need to consider the types and functional requirements of different monitoring points to ensure that each data point can accurately reflect the on-site status and provide reliable data support for subsequent analysis and decision-making.

[0028] (2)Adopt technologies such as machine learning and deep learning to extract features from the monitoring data Machine learning and deep learning technologies can automatically extract key features from a large amount of monitoring data, helping to identify potential patterns and abnormal changes. These technologies can discover important laws in the data through training on historical data. For example, algorithms such as deep neural networks (DNN) and convolutional neural networks (CNN) can identify key parameters affecting the safety of railway structures, such as track deformation and bridge stress. When faced with large-scale and high-dimensional data, machine learning can automatically extract the most representative features, avoiding the bias of manual intervention in traditional methods, thereby improving the efficiency and accuracy of data analysis.

[0029] (3) Combine the digital twin model to conduct correlation analysis on the feature data and achieve intelligent extraction of data information By combining the digital twin model for correlation analysis of feature data, the real-time collected monitoring data is combined with the dynamic state in the virtual model to achieve intelligent extraction. The digital twin model provides a visual and interactive virtual environment, enabling on-site data to be correlated and compared with the physical properties and behaviors in the model. By applying data correlation technologies such as multivariate analysis, time series analysis, and clustering analysis, the relationships between different monitoring points and potential risks can be revealed. For example, the correlation between temperature changes and material stress, and between vibration and structural displacement, can help detect potential faults in advance and assist project managers in making more accurate decisions. Intelligent analysis not only improves the data processing efficiency but also enhances the early warning and prediction capabilities of the digital twin model.

[0030] 4 Output accurate completion delivery data information (1) Summarize and organize the extracted data information During the completion delivery process, the extracted data needs to be carefully summarized and organized. This process involves classifying, standardizing, and formatting various types of monitoring data (such as structural health monitoring, construction progress, quality inspection, etc.) to ensure the integrity and consistency of the data. By uniformly organizing the monitoring data from different sources, redundant information can be effectively removed, highlighting key information and providing an accurate data basis for subsequent analysis and report generation. In addition, the organized data needs to meet industry standards and project requirements to ensure smooth passage through the acceptance process.

[0031] (2) Generate a completion delivery data report The organized data will be used to generate a completion delivery data report to ensure the smooth delivery of the project. The key indicators, abnormal situations, quality control measures, and rectification records of the monitoring data are detailed to ensure clarity, accuracy, and traceability. In addition, not only the overall project data is summarized, but also specific analyses are conducted for sub-projects (such as bridges, tracks, tunnels, etc.), providing information such as technical data, performance evaluation, and safety status. These reports will provide important bases for project acceptance, later operation, and maintenance decisions.

[0032] (3) Realize data sharing and collaboration among participating units through the data sharing platform In order to improve the efficiency and transparency of the completion and delivery data, the data sharing platform has become a key tool. By building a unified platform, all participating units (such as design, construction, supervision and owner units) can access and share relevant data in real time, promoting information exchange and collaborative work. The platform not only integrates technical monitoring information, but also covers data in multiple dimensions such as project progress, cost, and quality. Through this integrated data sharing, all parties can cooperate more efficiently, promote project delivery, and provide complete and transparent information support for subsequent operations and maintenance.

[0033] The specific steps are as follows: S1: Digital twin model construction (1) Collecting data and information from the design, construction, and acceptance stages of railway projects Data collection is divided into three stages: design, construction and acceptance: Design stage: Use BIM (Building Information Modeling) technology to digitally model railway line planning, station layout, bridge and tunnel design. Through data docking with design software such as Revit and AutoCAD, detailed information such as structure, pipeline, and electrical are collected. All design drawings, material specifications, structural requirements and other data are digitized and stored in the cloud.

[0034] Construction phase: Various IoT sensors (such as temperature, stress, vibration sensors, etc.) are installed at the construction site, and data such as progress, quality control, and changes in construction drawings are recorded through the construction management system. In addition, GPS and drone technology are used to obtain real-time progress and changes at the construction site.

[0035] Acceptance stage: Accurate on-site data collection through laser scanning technology and drones to verify whether the completed structure meets the design requirements. On-site inspection data (such as structural health monitoring reports, equipment performance test data, etc.) will be entered into the system for final review.

[0036] (2) Use 3D modeling technology to build a solid model of the railway project After completing data collection, integrate the design and construction data through BIM and CAD technologies, and use software such as Revit for modeling. Create a three-dimensional solid model using BIM technology that includes elements such as railway tracks, stations, bridges, and tunnels. The geometric information and related parameters (such as design strength, material properties, etc.) of each component (such as bridges, tracks, tunnels) are input into the model. Obtain high-precision point cloud data from the construction site through laser scanning technology. These data are converted into precise geometric data that can be used in the BIM model through point cloud processing software (such as CloudCompare), thus ensuring that the actual construction is consistent with the design. The design drawings, construction data, and on-site feedback are integrated through dedicated interfaces such as IFC to generate a comprehensive virtual entity model covering all construction details.

[0037] (3) Combine technologies such as the Internet of Things and big data to add attribute information to the physical model to form a digital twin model Based on the physical model, update the virtual model by combining real-time data to form a complete digital twin model; Install sensors such as temperature, vibration, and stress on key facilities (such as the connection parts of bridges, the entrances and exits of tunnels, and the walls of stations). The real-time data is transmitted to the cloud platform for processing through wireless networks such as 5G and LoRa. After the data collected by the sensors is preprocessed, it is synchronized with the BIM model in real time. For example, the temperature sensor data of the station wall will be automatically associated with the BIM model of the wall to provide feedback for real-time monitoring. Use big data analysis technology to process the sensor data, and through data mining and pattern recognition, use machine learning models to predict possible fault trends (such as bridge stress overload, etc.).

[0038] S2: Real-time collect data information during the completion and delivery process (1) Use devices such as sensors and cameras to monitor the railway engineering site in real time During the completion and delivery stage, comprehensive monitoring of the construction site is crucial. For this purpose, various technical means can be used to ensure the safety and quality of the facilities. Deploy sensors on key facilities (such as tracks, bridges, tunnels, stations, etc.). For example, install stress sensors at the load-bearing support points of bridges and displacement sensors at the top of tunnels to monitor the state changes of these facilities in real time. At the same time, set up multiple cameras at the construction site to transmit video images to the background in real time, facilitating managers to remotely view the construction progress and conduct real-time tracking of quality control.

[0039] (2) Transmit the monitoring data to the digital twin model through the data transmission network Data collected by sensors and cameras is transmitted to the cloud platform in real time via a wireless network (such as 5G or Wi-Fi). The cloud platform uses Internet of Things protocols (such as MQTT) to receive this data and applies it to update relevant information in the digital twin model. Whenever the on-site monitoring data changes, the model automatically adjusts to reflect the latest on-site status. For example, if the stress value of a bridge exceeds the set safety range, the bridge part in the digital twin model will automatically display a red alarm sign to remind the management to take timely countermeasures.

[0040] (3)Preprocess the monitoring data to ensure data quality When processing data collected by sensors, data cleaning is required to remove outliers and invalid data. For example, when a temperature sensor fails, it may generate abnormally high or low temperature data, and the system will automatically identify and eliminate these outliers. For missing data, data interpolation algorithms (such as linear interpolation or polynomial interpolation) are used to fill in the missing values to ensure data continuity. Finally, for high-frequency data such as vibration sensors, the system uses signal processing techniques (such as low-pass filtering) to reduce noise and extract effective monitoring signals to ensure data accuracy.

[0041] S3: Extract, analyze, and process data information (1)Set data extraction rules according to the completion and delivery standards To ensure that the monitoring data meets the project acceptance standards, specific data extraction rules are first formulated according to industry specifications. For example, the acquisition frequency of temperature data is set to once per hour, and stress data is collected every 10 minutes. At the same time, clarify which data points are key parameters (such as track verticality, bridge stress value) and pay special attention to them. To improve data extraction efficiency, an automated extraction is achieved using a data management platform. Through preset rules, the platform can automatically screen real-time data that meets the requirements to ensure that no key data is missed.

[0042] (2)Adopt technologies such as machine learning and deep learning to extract features from the monitoring data Use a convolutional neural network to extract key features related to structural health from the signals of vibration sensors, such as vibration frequency and amplitude. Through training on a large amount of data, the machine learning model can identify features related to structural damage or failure, providing a basis for subsequent maintenance. A supervised learning model is used to train the monitoring data to automatically detect abnormal states. For example, the model can learn the patterns of temperature and stress changes, timely identify sudden changes, and issue warnings, thus achieving early fault diagnosis and risk prevention.

[0043] (3)Combine with the digital twin model to conduct correlation analysis on the feature data and achieve intelligent extraction of data information Using the relationship between temperature data and track stress, it can be found that an increase in temperature may lead to track deformation, which in turn affects the safety of the railway. In addition, by performing time series analysis on the collected sensor data, the data change trends at different time nodes can be revealed. Using clustering analysis techniques, similar data patterns are classified to identify potential structural risks or construction progress problems. These analysis methods can not only provide real-time risk assessment but also provide data support for decision-making.

[0044] ① Set data extraction rules: Specific rules are formulated according to the completion and delivery standards and industry norms, covering aspects such as the quality, accuracy, and time frequency of monitoring data. For example, it is stipulated that the temperature data is collected once per hour, and the stress data is collected once every 10 minutes. At the same time, key parameters such as track verticality and bridge stress values are clearly defined and particularly concerned about. Use the data management platform to screen out the real-time data that meets the requirements without missing key data.

[0045] ② Extract features from the detection data: For the data collected by vibration sensors and other devices, a suitable convolutional neural network (CNN) model architecture is constructed. The CNN model can automatically learn local and global features in the data through components such as convolutional layers, pooling layers, and fully connected layers. For example, when analyzing bridge vibration data, the convolutional layer can extract the characteristic patterns of vibration signals within different time windows, such as the change trend of vibration amplitude in a specific frequency band; the pooling layer reduces the dimensionality of these features, compressing the data volume while retaining key feature information; the fully connected layer integrates and maps the extracted features to output a feature vector related to the structural health status. By training on a large amount of labeled vibration data and continuously adjusting the weight parameters of the CNN model, it can accurately extract key features related to structural health from the original vibration signal, such as the change in vibration characteristics caused by the appearance of cracks and the deviation of vibration frequency caused by structural loosening. Supervised learning algorithms such as support vector machines (SVM) and decision trees are used to train the preprocessed and feature-extracted data to construct an anomaly detection model. During the training process, the data labeled as normal and abnormal states are respectively used as positive and negative samples and input into the model to let the model learn the differences in data features between the two states.

[0046] ③ Correlation analysis of feature data and intelligent extraction of information: Combining with the digital twin model, using the relationship between temperature data and track stress, etc., it can be found that an increase in temperature may lead to situations such as track deformation that affect railway safety. By performing time series analysis on the collected sensor data, the data change trends at different time nodes are revealed; using clustering analysis techniques, similar data patterns are classified to identify potential structural risks or construction progress problems, realizing the intelligent extraction of data information and providing support for decision-making.

[0047] S4: Output accurate completion and delivery data information (1) Summarize and organize the extracted data information To ensure clear and orderly presentation of the data, first classify the data according to different monitoring items (such as structural health, progress control, quality inspection, etc.) and convert it into a unified format. For example, data such as bridge stress, temperature, vibration, etc. will be output in a predetermined standardized format to ensure that the report can intuitively display key data. Subsequently, summarize all monitoring data into a comprehensive completion data report to ensure complete and accurate information for subsequent acceptance.

[0048] (2) Generate a completion and delivery data report Generate a detailed completion and delivery report based on the previously summarized data. This report should not only include the comprehensive data of all monitoring items, but also cover key information such as the overall quality of the project, construction progress, and structural health status. Through clear data presentation and chart analysis, ensure that the report can provide strong evidence for project acceptance and also serve as a reference for subsequent operation and maintenance and management.

[0049] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, in any regard, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included within the present invention.

[0050] In addition, it should be understood that although this specification is described in terms of embodiments, not every embodiment only contains an independent technical solution. This narrative way of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for extracting railway engineering completion delivery data information based on digital twin, characterized in that: The specific steps are as follows: S1: Digital Twin Model Construction (1) Collect data during railway engineering design, construction, acceptance and other stages (2) Use 3D modeling technology to construct a railway engineering entity model After data collection, integrate design and construction data through BIM and CAD technologies, and use Revit software for modeling. The data is converted into precise geometric data for the BIM model through point cloud processing software; (3) Combine technologies such as the Internet of Things and big data to add attribute information to the entity model to form a digital twin model Based on the entity model, form a complete digital twin model by combining real-time data to update the virtual model; Install sensors including temperature, vibration, and stress sensors at key facilities. The real-time data is transmitted to the cloud platform through a wireless network for processing. After the data collected by the sensors is preprocessed, it is synchronized with the BIM model in real time. Use big data analysis technology to process the sensor data. Through data mining and pattern recognition, use machine learning models to predict possible fault trends; S2: Real-time collection of data information during the completion and delivery process (1) Use sensors and camera devices to monitor the railway engineering site in real time (2) Transmit the monitoring data to the digital twin model through a data transmission network The data collected by sensors and cameras is transmitted to the cloud platform in real time through a wireless network. The cloud platform uses Internet of Things protocols to receive this data and applies it to update relevant information in the digital twin model. Whenever the on-site monitoring data changes, the model will automatically adjust to reflect the latest on-site status; When the monitoring data changes, the digital twin model automatically updates relevant information mainly through the following key steps: First is the data reception and parsing link. The cloud platform receives the data from sensors and cameras according to the Internet of Things protocol, and disassembles the original data according to the preset format to obtain data information including data type, source, and value, which is used as the basic data for subsequent update operations; Then perform data verification and preprocessing. In terms of data verification, it is necessary to check data integrity to ensure that all parameters are successfully received; verify accuracy, judge whether the data is credible according to the set data range and verification rules, such as temperature data needs to be within a reasonable physical range; it is also necessary to verify timeliness to ensure that the data is collected and transmitted at appropriate time intervals; for data preprocessing, if there is missing data, interpolate and supplement it with reference to historical trends and surrounding relevant data. For abnormal data, use smoothing algorithms or correct and mark it according to statistical laws; Then is the model association and positioning step. According to the data source identifier, accurately find the corresponding model component or sub-model in the digital twin model; the regional camera image data corresponds to the corresponding spatial model part; Subsequently, enter the model parameter update stage, directly update the processed monitoring data to the corresponding device parameters in the model, so that the model is consistent with the actual situation in this parameter dimension; In terms of model visualization updates, first, there are geometric shape and appearance updates. The geometric shape and appearance of the model are adjusted according to the physical state changes reflected in the monitoring data. For example, when the building structure deforms, the corresponding distortion or displacement is shown in the model, and when the temperature of the equipment housing is too high, the color of the model equipment is changed to display a high-temperature warning. Second, there are animation and dynamic effect updates. For monitoring data with dynamic behaviors, such as equipment operation status, personnel or object movement trajectories, animation effects are added to the model. Finally, there is model consistency checking and optimization. During consistency checking, after parameter and visualization updates are completed, check whether the logical relationships and physical constraints of each part of the model are correct, and compare whether the changes in the model state before and after the update conform to physical laws and expected system behaviors. If there are any abnormalities, trace back in time to check for errors in the data processing or model update process. Model optimization is based on multiple monitoring data updates and model operation conditions. Adjust the model parameter calculation algorithm to improve efficiency and accuracy, simplify or refine the model structure to better adapt to the characteristics of monitoring data changes and application requirements. At the same time, based on big data analysis of monitoring data and model operation historical data, discover new correlation relationships or laws and incorporate them into the model construction or update mechanism to enhance the simulation and prediction capabilities of the digital twin model for the actual system. S3: Extract, analyze, and process data information (1)Set data extraction rules according to the completion and delivery standards To ensure that the monitoring data meets the project acceptance standards, first, specific data extraction rules are formulated according to industry norms. Through the preset rules, the platform can automatically screen the real-time data that meets the requirements. (2)Adopt technologies such as machine learning and deep learning to extract features from the monitoring data Use a convolutional neural network to extract key features related to structural health from the signals of vibration sensors. Through training on a large amount of data, the machine learning model can identify features related to structural damage or faults, and a supervised learning model is used to train the monitoring data. Combine with the digital twin model to conduct correlation analysis on the feature data and achieve intelligent extraction of data information. ① Set data extraction rules: Formulate specific rules according to the completion and delivery standards and industry norms, covering monitoring data including quality, accuracy, and time frequency. ② Extract features from the detection data: For the data collected by vibration sensors and other devices, construct a suitable convolutional neural network (CNN) model architecture. The CNN model includes convolutional layers, pooling layers, and fully connected layer components, which can automatically learn local and global features in the data. Adopt a supervised learning algorithm to train the preprocessed and feature-extracted data to construct an anomaly detection model. During the training process, the data labeled as normal state and abnormal state are used as positive and negative samples and input into the model respectively, allowing the model to learn the differences in data features between the two states. ③ Correlation analysis of feature data and intelligent extraction of information: Combine with the digital twin model, use the relationship between temperature data and track stress to discover that an increase in temperature may lead to situations such as track deformation that affect railway safety. Through time series analysis of the collected sensor data, reveal the data change trends at different time nodes. Using clustering analysis techniques, similar data patterns are classified to identify potential structural risks or construction schedule problems; S4: Output accurate completion and delivery data information (1) Summarize and organize the extracted data information Classify the data and convert it into a unified format to ensure that the report can intuitively display key data; Subsequently, summarize all monitoring data into a comprehensive completion data report; (2) Generate a completion and delivery data report Based on the previously summarized data, generate a detailed completion and delivery report. Through clear data display and chart analysis, ensure that the report can provide strong evidence for project acceptance and also serve as a reference for subsequent operation and maintenance and management.

2. The method for extracting railway engineering completion delivery data information based on digital twin according to claim 1, wherein: The data materials in Step 1 are divided into three stages: design, construction, and acceptance: Design stage: Use BIM technology to digitally model railway line planning, station layout, bridge and tunnel design. Through data docking of design software, collect detailed information including structure, pipelines, and electricity. All design drawings, material specifications, and structural requirement data are digitized and stored in the cloud; Construction stage: Install various Internet of Things sensors at the construction site and record data such as progress, quality control, and construction drawing changes through the construction management system; in addition, use GPS and drone technology to obtain the progress and changes at the construction site in real time; Acceptance stage: Use laser scanning technology and drones to accurately collect on-site data, verify whether the completed structure meets the design requirements, and input the on-site inspection data into the system for final review.