CN117171842A - Urban slow-moving bridge health monitoring and digital twin system - Google Patents

Abstract

The invention relates to a health monitoring and digital twin system for an urban slow-moving bridge, which comprises a live-action model module, a BIM modeling module, a full life cycle operation and maintenance management module, an edge equipment module, a digital twin foundation module, a digital twin data conversion and optimization module and a digital twin application module, wherein information transfer interaction can be realized among the modules. The method comprises the following steps: building a BIM three-dimensional information model of the slow bridge structure; establishing a live-action model of the slow-moving bridge and the surrounding geographical hydrologic environment; constructing a digital twin base layer, and converting and optimizing digital twin data; an intelligent operation and maintenance management platform based on a digital twin body is built. The digital twin system is based on bridge construction, operation maintenance stage structure state monitoring information acquisition, structure load working condition real-time acquisition, data processing and analysis, and remote macroscopic control of construction site conditions and construction period progress is realized through regular updating of a live-action model.

Description

An urban slow-traffic bridge health monitoring and digital twin system

Technical field

The present invention relates to the technical fields of intelligent construction of construction projects, intelligent construction and intelligent operation and maintenance, in particular to the field of construction and operation monitoring of bridge projects, and in particular to application tools and implementation methods for intelligent construction and operation monitoring of bridges. Specifically, it is an urban slow-speed Bridge health monitoring and digital twin system.

Background technique

The arrival of the digital age characterized by informatization and intelligence has promoted the development and innovation of bridge engineering technology. It is necessary to integrate strategic emerging industry technologies such as BIM, Internet of Things, digital twins, UAV real-life modeling, and edge equipment with The integration of bridge engineering will promote the industrialization, digitalization, and intelligent upgrading of bridges from multiple dimensions such as intelligent design, intelligent construction, and intelligent operation and maintenance.

BIM technology is a product of the construction industry that adapts to the development of the times. The concept of "Internet of Things" was proposed in 1999, which connects all items to the Internet through information sensing devices such as radio frequency identification to achieve intelligent identification and management. The Internet of Things technology originated in the media field and is the third revolution in the information technology industry. Through information sensing equipment and according to the agreed protocol, any object is connected to the network. The objects exchange information and communicate through the information communication media to achieve Intelligent identification, positioning, tracking, supervision and other functions. The concept of digital twins was first clearly proposed by NASA in the United States. Its meaning is to highly accurately establish corresponding digital information models based on things in the real world, and perform operations such as reality narration, simulation, and prediction evaluation. The proposal of digital twin technology has high hopes as a bridge for interaction and integration between the physical world and the digital world. However, currently digital twin technology is mainly used in industrial-grade high-precision industries, and its application in the field of bridge engineering still needs to be improved. UAV real-life modeling is a three-dimensional model that objectively restores real scenes. It has the characteristics of singleness, materialization, structure, and semantics. It integrates artificial modeling technology, oblique photography technology, and laser radar technology to form a spatially quantifiable model. and comprehensively analyzed three-dimensional models, presenting three-dimensional data scene results that integrate the advantages of multiple models. The important problems that edge computing solves are the problems of high latency, network instability and low bandwidth that exist in the traditional cloud computing (or central computing) model.

Engineering intelligent construction is based on the digitization of engineering construction element resources, using the engineering construction information model as the carrier, and using automation equipment and Internet of Things information technology as means to realize status identification, error analysis, error prediction, and evaluation correction driven by digital chains. , dynamic adjustment and other auxiliary guidance and decision-making for the whole process of construction behavior, and ultimately realize an innovative construction model with high precision, high quality and high efficiency for the delivery of engineering structural products. Digital twin technology combines the data collection of the Internet of Things, makes full use of model and sensor updates, and historical operation data, integrates multi-disciplinary, multi-physical quantities, multi-scale, and multi-probability simulation processes to complete the mapping of bridges in virtual space and realize the mapping of bridges. Understand the historical conditions, evaluate the current operating status, simulate and diagnose the status and behavior of the bridge, and predict the evolution trend of health conditions and possible diseases and risks (Tongji University. A modular building based on a digital twin platform Health Monitoring System: 202011276245.9[P].2021-02-19).

Based on the technological development background mentioned above, it can be realized that digital twin technology has a huge role but is relatively complex. The present invention can promote the development of digital twin technology in the field of engineering construction by combining the Internet of Things technology and BIM technology. For urban slow-moving bridges, the development of intelligent construction technology also needs to be promoted to adapt to the development of the industry. At present, wireless digital transmission is also used for bridge health monitoring (Zhu Jun, Gao Jian, Wang Lei. A wireless digital transmission device for bridge health monitoring [P]. Jiangsu Province: CN208985362U, 2019-06-14.), This can greatly improve the monitoring efficiency of bridge data, which reflects the superiority of IoT technology. Therefore, it is necessary to propose a digital twin system based on BIM+GIS+FEM technology and IoT technology to intelligently intelligentize the construction and operation and maintenance processes of urban slow-moving bridges.

Contents of the invention

In order to solve the problems of reasonable interaction between finite element analysis and building information model (BIM), the establishment of a digital twin system that combines the construction, operation and maintenance process of urban slow-moving bridges with Internet of Things technology, and the construction process of intelligent urban slow-moving bridges, the present invention A digital twin system for intelligent construction and intelligent operation and maintenance of urban slow-moving bridges is proposed and used in the construction process of intelligent urban slow-moving bridges, including BIM+GIS module, Internet of Things module, data processing module and construction simulation module, with BIM+GIS module, data processing module and construction simulation module as the basis to establish a digital twin of long-span steel bridge construction, and then based on the Internet of Things module to realize data interaction between the digital twin part and the physical entity part, and establish a digital twin system for urban slow-moving bridge construction. Data interaction realized through Internet of Things technology can reflect the internal force conditions and physical characteristics of prefabricated components during transportation and installation in real time, thus realizing the construction process of intelligent long-span steel bridges. For example, managers can compare and analyze whether components are qualified based on digital twins established based on BIM and other technologies. At the same time, the digital twin system can simulate in real time to map the specific construction process. In addition, the digital twin system can simulate the proposed construction process to predict the construction process, thereby conducting analysis and making scientific decisions to guide construction.

The present invention is realized through at least one of the following technical solutions.

An urban slow-traffic bridge health monitoring and digital twin system, including a BIM modeling module, a real-life model module, a full life cycle operation and maintenance management module, a digital twin basic module, a digital twin data conversion and optimization module, and a digital twin application module;

The BIM modeling module includes unified naming and annotation standard units for component information design classification, a modeling module for creating complex curved surface components of bridges, a modeling module for creating regular components of bridges, and a modeling module for creating bridges with regular components. Model conversion module for converting component information between components;

The real-life model module includes a UAV module, an apron module, a live streaming module, a three-dimensional reconstruction module, and a format conversion module; the UAV module integrates RTK positioning and high-resolution cameras; the apron module implements UAV It has a low-battery automatic return-to-home charging function and an operating environment sensing function. The tarmac senses wind speed, rain, and light flight conditions to determine whether flight conditions are met and realize automated operations; the live streaming module uses drones for real-time monitoring and remote viewing at any time. Live work, pictures, and video results are automatically returned and archived, making it convenient to call pictures for three-dimensional reconstruction during real-life modeling;

The three-dimensional reconstruction module finally forms a real-life model by adding photos and coordinates, adding image control points and their coordinates, aligning photos, building dense point clouds, generating grids, and generating textures. The real-life model is a collection of multi-angle pictures of physical entities. Perform three-dimensional reconstruction; the format conversion module converts the real-life model in OSGB format into 3D Tiles format.

The full life cycle operation and maintenance management module includes a 6D BIM initial information unit for comprehensive information application, a comprehensive operation and maintenance information unit for bridge information update and correction, and a full life cycle information operation and maintenance for digital twin application results management. Management module;

The digital twin basic module includes various sensor units used for collecting on-site working conditions and bridge response information during the construction phase, various sensor units used for real-time identification and tracking during the operation and maintenance phase and collecting the load distribution of slow-moving bridges, and various sensor units used for transmitting information. Wireless communication transmission unit, data processing unit for processing information, bridge suspension cable force calculation edge device unit;

The digital twin data conversion and optimization module includes a 6D BIM preliminary information extraction and conversion unit including unit information, a working condition information unit used for numerical analysis of finite element structures, and a slow-moving bridge under real-time load conditions. Structural reaction force calculation and early warning unit, structural reaction comparison unit for simulation analysis under different working conditions, unit parameters and grid optimization unit;

The digital twin application module includes a digital twin application module that combines finite element analysis and intelligent analysis, and an information transfer module for digital twin layer application information interaction.

Further, the initial information unit of 6D BIM includes bridge real-life model and structural model information (3D), overall schedule planning information (4D), project quantity and cost information (5D), and structural analysis related information (6D), among which structural analysis related information Information (6D) includes unit parameters and mesh division information, material property information, working condition load information, boundary condition processing information and structural surface crack damage information;

The full life cycle information operation and maintenance management module builds a comprehensive operation and maintenance information database through the collection and integration of various types of data, data lightweighting and storage backup, and uniformly updates and stores multi-dimensional data at different stages on the server side for customers to use. The web side retrieves and accesses information, and combines the retrieved information with the BIM lightweight model for viewing and output.

The multi-dimensional information of the initial information unit of 6D BIM is added and improved based on the three-dimensional model information. After being lightweight, the multi-dimensional information interacts with the life cycle information operation and maintenance management module.

Further, the digital twin application module includes finite element numerical analysis results calculated based on real-time load conditions, a working condition bridge response database for intelligent analysis based on historical sample analysis results, and an intelligent prediction and safety assessment unit combined with the database. , the structural damage identification unit based on the convolutional neural network, and the remaining effective information units of the digital twin application layer. Among them, the intelligent prediction includes extracting the internal force results of the bridge under similar working conditions at different times, analyzing the internal force changes in key parts of the bridge, and the overall internal force distribution. make predictions;

The information transfer module transfers and transmits the information in the digital twin application module to the life cycle information operation and maintenance management module, and then realizes the integration with the 6D BIM initial information unit through the information viewing and output unit of the life cycle information operation and maintenance management module. information interaction between.

Further, the remaining effective information units of the digital twin application layer include updating BIM structural analysis-related information based on the optimized finite element model units and grid parameters, providing initial units and sixth-dimensional structural analysis information for the 6D BIM initial information unit. The value range of grid parameters; based on the comparison and error analysis of various types of bridge responses under different working conditions, boundary type selection and boundary simplification, load simulation effectiveness, and bridge response monitoring layout points are carried out for the finite element structural analysis model. Assessment; based on the results of finite element analysis and intelligent analysis, automatically generate preventive maintenance decisions and assessment documents for bridge structures.

The intelligent analysis optimizes the finite element model boundary conditions, unit parameters and grid information through the working condition response database of historical sample analysis results, so that the finite element results are close to the actual measured values.

Further, the structural damage identification unit based on convolutional neural network includes the following steps:

Data collection and preprocessing: A data set containing structural surface images is collected by drone close-up photography, and the data set is preprocessed, including image resizing, grayscale, and denoising operations;

Divide the data and set up the CNN network architecture model: divide the data set into a training set, a validation set and a test set, set the number and size of the convolutional layers, activation function, pooling layer type and parameters; network training: use training Sets are used to train the CNN. During the training process, the network parameters are updated through the back propagation algorithm to minimize the loss function;

Model evaluation: Use an independent test data set to evaluate the performance of the network. The evaluation index is accuracy. Hyperparameter tuning: Based on the performance of the validation set, adjust the network architecture and hyperparameters, including learning rate, batch size, and regularization parameters;

Model testing: Use the test set to evaluate the performance of the final model.

Furthermore, the bridge suspension cable force calculation edge device unit includes a digital twin operating system (DTOS), an acceleration sensor, a frequency domain data filtering and cable force calculation module, and a suspension cable force status wireless communication transmission unit; multiple edge devices communicate through high-speed The data communication protocol is connected to form digital twin edge equipment, which is managed uniformly through the digital twin edge equipment wireless remote management platform on the mobile client APP, PC client APP or web version;

The frequency domain data filtering processing and cable force calculation module collects the time domain data of the target cable at preset time intervals, and uses the fast Fourier transform method to convert the time domain data into frequency domain data to construct a spectrogram, The second-order derivative method is used to identify all the wave peak data in the spectrum diagram, and then the cable force of the target cable is calculated through the wave peak data.

Further, the calculation and early warning unit for slow-moving bridge structure reaction force under real-time load conditions includes a cross-camera pedestrian recognition and trajectory tracking module, an anonymous pedestrian weight detection module, a finite element numerical analysis module and a result output module. Finite element result error evaluation module and threshold alarm module;

The cross-camera pedestrian recognition and trajectory tracking module uses the timing and spatial relationships of surveillance videos to optimize and infer the multi-camera network topology design and optimize the pedestrian recognition deep learning model, thereby realizing the pedestrian recognition function based on the appearance of pedestrians. In the pedestrian recognition deep learning model Based on this, the dynamic delay topology network model between multiple cameras is implemented to provide a more optimized detection sequence for the search, thereby reducing the blindness of the search.

Furthermore, the update and iterative implementation process of the unit parameters and grid optimization unit includes two levels: on-site operation and maintenance information update through the comprehensive operation and maintenance information unit, and unit parameter and network optimization through comparative analysis of bridge response errors under different working conditions. Grid optimization and iterative update of unit information. Among them, the comprehensive operation and maintenance information unit update content includes the collection and import of on-site working condition information, the collection and comparative analysis of various bridge response information, and the geometric parameters obtained through on-site appearance measurement and component testing. and comprehensive operation and maintenance information on material properties.

The method for implementing the above-mentioned urban slow-traffic bridge health monitoring and digital twin system includes the following steps:

Step 1: Reconstruct the real-life model of the bridge and its surrounding environment to reflect the real physical entities;

Step 2: Establish a BIM three-dimensional information model of the bridge structure;

Step 3: Build the digital twin base layer, which specifically includes collecting the overall status data of the bridge, transmitting the data to the data processing unit and lightweight data processing;

Step 4: Convert and optimize the digital twin data;

Step 5: Build an intelligent operation and maintenance management platform based on the digital twin hierarchical architecture. Specifically, through the information transfer module, the information transfer and comprehensive application between the digital twin application module and the full life cycle information operation and maintenance management module are realized. .

Further, the intelligent operation and maintenance management platform layers the digital twin application module, the information transfer module, and the full life cycle information operation and maintenance management module, specifically into the data perception layer, the digital twin data processing layer, and the digital twin Application layer and intelligent construction and operation and maintenance management layer;

The data sensing layer includes various types of IoT sensors, bridge suspension cable force calculation edge devices, RTK positioning technology drones, environment and bridge structure monitoring modules, human-machine, material, law and environmental intelligent identification modules, and pedestrian identification and trajectory tracking modules;

The digital twin data processing layer includes bridge digital twin modeling, precise real-life three-dimensional modeling modules, on-site load condition response data, mechanical performance testing data and other effective operation and maintenance data, and collects, organizes, integrates and exchanges the data. , lightweight preprocessing, and finally establish an effective information database according to different module requirements;

The digital twin application layer includes the bridge structure digital twin, BIM system, GIS system, finite element simulation system, database-based comprehensive operation and maintenance management system, preventive assessment system based on intelligent early warning analysis, knowledge map-based And machine learning maintenance plan formulation system and other digital twin expansion applications, damage identification of structural surface cracks based on convolutional neural network;

The intelligent construction and operation and maintenance management layer includes multi-dimensional information such as three-dimensional models, construction progress information, project quantity and cost information, and structural analysis divided into different information dimensions, and multi-stage information such as planning and design, construction management, and operation and maintenance divided into different construction stages. , as well as functional modules such as information entry and management, collaborative communication of various sub-items, and preventive maintenance decision-making;

The output end of the data sensing layer is connected to the input end of the digital twin data processing layer, the output end of the digital twin data processing layer is connected to the input end of the digital twin application layer, and the output end of the digital twin application layer is connected to the intelligent construction and The input terminal of the operation and maintenance management layer is connected.

Compared with existing technology, the beneficial effects of the present invention are:

1. This invention has certain reference value for the in-depth application of urban slow-traffic bridge BIM+GIS+FEM models combined with Internet of Things, digital twins and other technologies, and provides a reference for promoting BIM applications in the industry.

2. The establishment of digital twins of urban slow-travel bridges reflects the construction conditions of urban slow-travel bridges in real time, which is of great significance to promoting the digitalization of urban slow-travel bridge construction.

3. This invention uses digital twin technology to intelligentize the urban slow-moving bridge construction process, which has great guiding significance for the optimization of bridge construction technology. At the same time, it conforms to the future development direction of the industry, promotes the rapid development of the industry, and has huge practical value and economic benefits.

Description of drawings

Figure 1 is a schematic diagram of the module level of the embodiment;

Figure 2 is a structural flow chart of an urban slow-traffic bridge health monitoring and digital twin system according to the embodiment;

Figure 3 is a schematic diagram of the hierarchical architecture of the digital twin according to the embodiment;

Figure 4 is a schematic diagram of the hierarchical architecture of an edge device according to an embodiment;

Figure 5 is a schematic flowchart of an urban slow-traffic bridge health monitoring and digital twin method according to the embodiment.

Detailed ways

The present invention will be explained in more detail below with reference to specific implementation examples and the accompanying drawings. It should be understood that the examples are only for illustrating the implementation process of the present invention more specifically to facilitate understanding by those in the industry, but the present invention is not limited thereto. Based on the main concept of the present invention, some simple deformations or improvements may be made within the protection scope of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

In the examples described in the present invention, unless otherwise explicitly stated and limited, the terms "conversion", "connection", "transfer", "update", "interaction", "implementation" and other terms should be understood in a broad sense. For example, the transfer can be a simple transfer between the two, or it can be an indirect transfer through an intermediary; the interaction can be a direct interaction between the two units, or it can be an indirect interaction including certain technical means. For different actual situations, the specific meanings of the above terms can be understood as needed.

As shown in Figure 1, an urban slow-traffic bridge health monitoring and digital twin system includes a BIM comprehensive modeling and information management subsystem and a digital twin technology subsystem based on BIM-GIS-FEM. The BIM comprehensive modeling and information management subsystem The management subsystem includes BIM (Building Information Model) modeling module, reality model module, and full life cycle operation and maintenance management module; the digital twin technology subsystem based on BIM-GIS-FEM includes digital twin basic module, digital twin data conversion and optimization modules and digital twin application modules.

The BIM modeling module shown in Figure 1 is used to achieve accurate modeling between different types of components of complex bridge structures. Specifically, the BIM modeling module includes a variety of comprehensive modeling and conversion software (such as Revit , Rhino, 3dxMAX), various three-dimensional digital modeling software can be used for modeling according to the needs of complex bridge structures, and then the created models can be integrated into unified BIM software through model conversion software for subsequent information improvement and comprehensive application;

By formulating unified naming and annotation standards, it facilitates unified information management of different components of various modeling software. Different modeling software can directly convert and position the component models through the model modeling software, and finally realize the splicing integration and information management of the overall model in the BIM software platform with better comprehensive information management capabilities and more reasonable engineering application.

The unified naming and annotation standards mentioned above include: unit project name + location + section number + component category + detailed unit division information, etc., establishing an information classification and coding method for the full bridge components, and finally forming a similar "XX-XX" .XX.XX.XX” unified standard for hierarchical information management.

The real-life modeling module includes a drone module, an apron module, a live streaming module, a three-dimensional reconstruction module, and a format conversion module. The drone module integrates RTK positioning and high-resolution cameras. The apron module realizes the low-power automatic return-to-home charging function of the drone and the operating environment sensing function. The apron senses wind speed, rain, light and other flight environments to determine whether the flight conditions are met and realize automated operations. The live streaming module allows real-time monitoring by drones, which facilitates online monitoring of the job site and enables remote viewing of the actual work status at any time. Pictures and video results are automatically returned and archived, which facilitates the use of pictures for three-dimensional reconstruction during real-life modeling, eliminating the need for personnel to intervene in the process of inserting and unplugging memory cards to transfer pictures. The three-dimensional reconstruction process usually includes the following steps: adding photos and coordinates, adding image control points and their coordinates, aligning photos, building dense point clouds, generating meshes, generating textures, and finally forming a real-life model. Reality models are the result of three-dimensional reconstruction using multi-angle pictures of physical entities. The commonly used format for real-life models is OSGB format, which needs to be converted into lightweight 3D Tiles format that can be browsed and operated smoothly on the web. The format conversion module converts the real-life model in OSGB format into 3D Tiles format.

The full life cycle information operation and maintenance management module is used to classify and store the operation and maintenance management information during the whole life cycle of the bridge, and process the relevant information into a formatted expression to facilitate integration with 6D BIM (six-dimensional building information model). ) for information interaction. The presentation method of this module can be based on the comprehensive information management platform based on lightweight web pages for real-time display;

As shown in Figure 1, the full life cycle operation and maintenance management module includes the 6D BIM initial information unit for the initial design modeling process, the comprehensive operation and maintenance information unit for bridge information update and correction, and the digital twin application result management Full life cycle information operation and maintenance management;

The 6D BIM initial information unit includes bridge three-dimensional model information 3D, overall schedule planning information 4D, project quantity and cost information 5D, structural analysis related information 6D; on-site monitoring information and unit parameter optimization of the model revision process that are regularly updated during the construction process information; later comprehensive operation and maintenance information and application information from the digital twin functional layer.

The structural analysis related information (6D) includes unit parameters and mesh division information, material property information, working condition load information, boundary condition processing information and structural surface crack damage information;

The multi-dimensional information of the initial information unit of 6D BIM is added and improved based on the three-dimensional model information. After being lightweight, the multi-dimensional information interacts with the life cycle information operation and maintenance management module.

The comprehensive operation and maintenance information unit includes structural geometry information, mechanical performance information, inspection and test information, assessment decision-making information and maintenance recommendation information involving geometric parameters and material properties. Among them, structural geometry information or mechanical performance information involving geometric parameters and material properties are used to update the 6D BIM initial information unit; the remaining effective operation and maintenance information is used to update the full life cycle information operation and maintenance management module.

The full life cycle information operation and maintenance management module includes a comprehensive database unit, a configuration application server unit, a data call and feedback unit, and an information viewing and output unit. The comprehensive database unit builds an information database for comprehensive operation and maintenance through processes such as collection and integration of data at various stages, data lightweighting, and storage and backup. Unifiedly connect the information database to the server so that customers can retrieve and access information from the backend server on the Web or other customer service terminals, and combine the retrieved information with the BIM lightweight model for viewing and output functions.

The edge device unit is based on a high-performance micro development board and peripheral equipment as the core machine, adopts a distributed edge equipment system architecture, is driven by the collected monitoring big data and edge-side artificial intelligence deep learning algorithms, and is based on the edge-side inference framework. Real-time monitoring and grid-connected operation of the operation and maintenance process of civil engineering projects, facing the full life cycle of civil engineering design, construction and operation and maintenance, anchoring the digital twin goals and strategies of baseline modeling, comprehensive perception, and real-time prediction, based on E-BIM /E-GIS/E-HIM/E-AIM and physical/data/knowledge-driven digital twin baseline status, dynamic evolution management and predictive operation and maintenance decision-making enabling platform to achieve edge IoT operation monitoring - digital twin Iterative optimization—integration and efficient collaboration of structural health evolution trend prediction.

The digital twin basic module is used for on-site working conditions and various types of bridge response information collection, communication transmission and preliminary data processing, etc., through the collection and fusion, lightweight processing, data transfer and other processes of data collected by different sensors, for Establish a comprehensive operation and maintenance database to provide basic conditions, and input lightweight effective on-site working condition information into finite element analysis software to conduct bridge response analysis under specific loads. This is the basis for building a digital twin of a complex bridge structure and provides an on-site information source for the digital twin data conversion and optimization module.

Specifically, the digital twin basic module includes various sensor units used for collecting on-site working conditions and bridge response information during the construction phase, various sensor units used for real-time identification, tracking and inspection during the operation and maintenance phase, and collecting slow-moving bridge load distribution. Wireless communication transmission unit for collecting information, data processing unit for collecting information, bridge suspension cable force calculation edge device unit; various sensor units include various types of sensors.

The bridge suspension cable force calculation edge device unit includes a digital twin operating system (DTOS), a high-sensitivity acceleration sensor, a frequency domain data filtering processing and cable force calculation module, a suspension cable stress status wireless communication transmission unit; edge device integrated wireless communication transmission The unit supports wireless broadband (WIFI) and 5G/4G commercial networks, and can decide which wireless communication method to use for data transmission based on the on-site operating environment. If there is WIFI coverage where the edge device is placed, connect the edge device to the on-site LAN through the on-site WIFI network, and use File Transfer Protocol (FTP) to transfer the data to the database in the LAN, so that it can be called into the intelligent operation and maintenance management platform. show. If there is no local area network coverage where the edge device is placed, a 5G or 4G commercial network can be used. In this case, a valid SIM card needs to be inserted into the edge device and the TCP protocol or FTP protocol is used for data transmission. Multiple edge devices are connected through high-speed data communication protocols to form digital twin edge equipment, which is managed uniformly through the digital twin edge equipment wireless remote management platform on the mobile client APP, PC client APP or web version.

The frequency domain data filtering processing and cable force calculation module collects the time domain data of the target cable at preset time intervals, and uses the fast Fourier transform method to convert the time domain data into frequency domain data to construct a spectrogram, The second-order derivative method is used to identify all the wave peak data in the spectrum diagram, and then the cable force of the target cable is calculated through the wave peak data.

The various sensor units described mainly include temperature sensors and wind speed and direction meters arranged at the main span and tower of the bridge, static levels and fiber Bragg grating strain gauges arranged at the bottom of the bridge deck at the control points of the main structure of the bridge. Displacement sensors and acceleration sensors at the end supports and pier supports, cable force calculation edge equipment arranged on the stay cables or main cables, infrared induction surveillance cameras arranged on the bridge deck for pedestrian identification, arranged on the bridge The entrance and exit gates are used to obtain pressure sensors of load size, drones and their cameras, and GNSS sensors arranged in the middle of the span and at the top of the tower.

Specifically, the information collected by various sensors is connected to the data processing unit through 5G wireless communication through different regional forwarding nodes. The data processing unit further standardizes the data format by integrating, classifying and lightweighting the collected data. In order to improve the applicability of information, it is conducive to the rapid input of on-site working condition loads for finite element analysis and the error comparison of different types of bridge responses. In particular, Soli computing edge devices have data processing capabilities and can directly transmit calculation results and conclusions to the full life cycle operation and maintenance management module through 5G wireless communication.

The digital twin data conversion and optimization module described in Figure 1 is used to build a digital twin corresponding to the real structure. Through the secondary development plug-in, the extraction and conversion between the initial BIM information and the finite element model information are realized. On this basis, the error comparison analysis of the finite element analysis results of the bridge response and the bridge response collected on site is carried out based on the real-time load. After a continuous iterative process The unit parameters and grid optimization information in the model conversion are optimized and updated until the comparison of bridge responses under different working conditions and loads can meet the error conditions.

Specifically, the digital twin data conversion and optimization module includes a 6D BIM initial information extraction and conversion unit including unit information, a model and working condition information unit for numerical analysis of finite element structures, and a real-time load condition unit. Calculation and early warning unit for structural reaction force of slow-moving bridges, various types of bridge response comparison units for simulating and analyzing structures under specific working conditions, unit parameters and grid optimization units based on intelligent optimization algorithms.

The working condition information unit of the numerical analysis of the finite element structure inputs the real-time load conditions of the slow-moving bridge into the finite element model. According to the structural response comparison unit used for simulation analysis under specific working conditions, the bridge structure expert group determines a series of specific adverse load conditions, calculates the bridge response under these adverse working conditions, and determines the danger threshold of the bridge reaction. Use this threshold to compare the bridge response under real-time load conditions.

The unit parameters and grid optimization units based on the intelligent optimization algorithm optimize the finite element model boundary conditions, unit parameters and grid information through the historical sample analysis results working condition response database, so that the finite element results are close to the measured values. It solves the problem that the building information model (BIM) does not consider the unit parameters and grid information in the structural simulation analysis, lacks the necessary unit parameter optimization and building information model correction process, and cannot simulate the real bridge completely equivalently.

The preliminary information extraction and conversion unit mainly includes component information and unit information. The component information can be updated through manual direct input or adjustment of operation and maintenance information such as geometric parameters and material properties to update the 6D BIM initial information unit; the unit information is based on intelligent The unit parameters and grid optimization units of the optimization algorithm are updated through a cyclic iterative process, and the model is finally modified into a digital twin that is consistent with the real bridge structure.

The iterative update implementation process of the unit parameters and grid optimization units includes two levels. At the macro level, on-site operation and maintenance information can be modified through the comprehensive operation and maintenance information unit, and then the 6D BIM initial information unit and preliminary information extraction and conversion unit to realize information update and optimization of the finite element analysis model; among them, the update content of the comprehensive operation and maintenance information unit mainly includes the collection and import of on-site working condition information, the collection and comparative analysis of various bridge response information, through on-site appearance measurement and Comprehensive operation and maintenance information on geometric parameters and material properties obtained through component testing. At the micro level, the finite element results can be made close to the actual measured values through comparison of response errors of various bridges under different working conditions and intelligent optimization analysis. Intelligent optimization analysis continuously and iteratively updates unit parameters, grid optimization and related unit information through the working condition response database of historical sample analysis results.

The unit parameters and grid optimization unit include iterative optimization of preliminary information on unit types, grid types, and unit division fineness. On the basis of meeting error requirements under different working conditions, the optimized and improved information is transmitted to the digital twin. The application module performs classified storage and can use the information transfer module to transmit information to the full life cycle information operation and maintenance management module, thereby guiding the update of structural analysis unit parameters and grid information contained in the initial information unit of 6D BIM.

The calculation and early warning unit for the reaction force of slow-moving bridge structures under real-time load conditions includes a cross-camera pedestrian recognition and trajectory tracking module, an anonymous pedestrian weight detection module, a finite element numerical analysis module and a result output module, and a finite element result module. Error evaluation module, threshold alarm module.

The anonymous pedestrian weight detection module is installed at the entrance and exit gates of slow-moving bridges to anonymously weigh and record pedestrians passing through the gates and track their locations in real time. Through this method, the real-time load condition of the slow-moving bridge can be accurately determined at any time period. This load condition is also the load condition imposed on the corresponding finite element numerical analysis model of the bridge, so that the real-time bridge response can be obtained through finite element calculation. . The bridge strain calculated by finite element and the strain value of embedded strain gauge were compared and analyzed to evaluate the error of finite element result and verify the validity and accuracy of the finite element model.

The cross-camera pedestrian recognition and trajectory tracking module uses the timing and spatial relationships of surveillance videos to optimize and infer the multi-camera network topology design and optimize the pedestrian recognition deep learning model, thereby realizing the pedestrian recognition function based on the appearance of pedestrians. In the pedestrian recognition deep learning model Based on this, the dynamic delay topology network model between multiple cameras is implemented to provide a more optimized detection sequence for the search, thereby reducing the blindness of the search.

The finite element result error evaluation module obtains the reference solution under specific working conditions from the database, that is, the values monitored by each sensor on the slow-moving bridge. The finite element calculation results under this working condition are compared with the reference solution, and the error indicators used are H1 error (positive semi-definite error) and L2 error (square root error). Only when the error is less than a certain user-set value, the finite element model can be considered valid, and the results can be displayed on the intelligent operation and maintenance management platform. Threshold alarm modules monitor specific finite element calculation results. The collected real-time load conditions of the slow-moving bridge are applied to the finite element model for calculation. When specific finite element calculation results, such as bridge deflection, etc., exceed the allowable value of the specification, the entrance and exit gate of the slow-moving bridge will be set to " At the same time, the intelligent operation and maintenance management platform will send alarm messages to users with specific permissions; when the specific finite element calculation results significantly exceed the allowable value of the specification, the slow-moving bridge alarm system will be triggered and a broadcast will be issued to urge pedestrians on the bridge to get off. Bridge to avoid danger.

The secondary development plug-in between Building Information Model (BIM) and finite element structural analysis model mainly involves any language compatible with .NET (such as C#, C++, F#, Visual Basic.NET, etc.) and integrates a variety of Computer language development tools. In addition, there are SDK files provided for developers and some officially provided development plug-ins, including many help files and source code examples and methods.

The data information conversion type of the secondary development plug-in mainly includes geometric information such as component unit size and coordinate position; material properties such as elastic modulus, Poisson's ratio, density and strength; unit type and size, network used in finite element analysis The type and fineness of grid division. Among them, the geometric information and material properties can be updated periodically through on-site data collection of the comprehensive operation and maintenance information unit or the simultaneous component testing method; the unit parameters and grid optimization unit can be compared with the errors of the bridge response from the finite element analysis results and the on-site measured bridge response. , unit information adjustment and cycle iteration process to achieve optimization.

Among them, the initial information unit conversion of 6D BIM adopts the method of secondary development plug-in to generate command flow. In the process of iterative correction of the model, the relevant parameters are directly added or modified in the finite element software, and then the correction can be made in the finite element software. The processing step information and solution analysis result information are exported to a file in a specific format in the form of a command stream to facilitate data transfer and interaction in later applications.

As shown in Figure 1, the digital twin application module includes a digital twin application module that combines finite element analysis and intelligent analysis, and an information transfer module for digital twin layer application information interaction. The information transfer module transfers and transmits the information in the digital twin application module to the life cycle information operation and maintenance management module, and then realizes the integration with the 6D BIM initial information unit through the information viewing and output unit of the life cycle information operation and maintenance management module. information exchange between.

The digital twin application module mainly includes the finite element numerical analysis results calculated in real time according to the actual load, the working condition bridge response database for intelligent analysis created by combining the historical sample analysis results, the intelligent prediction and safety assessment unit combined with the database, and the volume-based The structural damage identification unit of the cumulative neural network and the remaining effective information units of the digital twin application layer are combined with historical samples to establish a bridge response database under various working conditions. Intelligent optimization algorithms are used to intelligently apply and analyze the database, which can The twin structure performs bridge response prediction and safety assessment and effective information management in other application layers. The digital twin application module pioneers the integration application between the digital twin hierarchical architecture and finite element numerical analysis, and further connects the information transfer module with the full life cycle information operation and maintenance management module to realize the full life cycle intelligence of the bridge. Operations management provides the architectural foundation.

The structural damage identification unit based on convolutional neural network includes the following steps:

Data collection and preprocessing: A data set containing structural surface images is collected by drone close-up photography, and the data set is preprocessed, including image resizing, grayscale, and denoising operations;

Divide the data and set up the CNN network architecture model: divide the data set into a training set, a validation set and a test set, set the number and size of the convolutional layers, activation function, pooling layer type and parameters; network training: use training Sets are used to train the CNN. During the training process, the network parameters are updated through the back propagation algorithm to minimize the loss function;

Model evaluation: Use an independent test data set to evaluate the performance of the network. The evaluation index is accuracy. Hyperparameter tuning: Based on the performance of the validation set, adjust the network architecture and hyperparameters, including learning rate, batch size, and regularization parameters;

Model testing: Use the test set to evaluate the performance of the final model.

The remaining effective information units of the digital twin application layer include updating BIM structural analysis-related information based on the optimized finite element model units and grid parameters, and providing initial unit and grid parameters for the sixth-dimensional structural analysis information of the 6D BIM initial information unit. The value range of The results of finite element analysis and intelligent analysis automatically generate preventive maintenance decision-making and assessment documents for the bridge structure.

Intelligent analysis optimizes the finite element model boundary conditions, unit parameters and grid information through the working condition response database of historical sample analysis results, so that the finite element results are close to the actual measured values. It solves the problem that the building information model (BIM) does not consider the unit parameters and grid information in the structural simulation analysis, lacks the necessary unit parameter optimization and building information model correction process, and cannot simulate the real bridge completely equivalently.

The transfer and interaction between the digital twin application module, the full life cycle information operation and maintenance management module and the 6D BIM initial information need to be carried out with the help of information conversion files in multiple formats, mainly including secondary development plug-in extraction and generation that meet the parameterization of finite element analysis software. Command flow files with specific requirements for the design language, inp files used in the model correction preprocessing stage, res files and animation files for storing solution result data, etc. The implementation of relevant functions of the digital twin application module requires the combination of various technical principles, which mainly involves the establishment of sample databases, bridge response prediction combined with intelligent optimization algorithms, intelligent expanded applications, and the development of comprehensive information display platforms.

Specifically, the intelligent operation and maintenance management platform based on the digital twin application module integrates the hardware and software foundation of the digital twin, and mainly includes the data perception layer, the digital twin data processing layer, the digital twin application layer, and intelligent construction and operation and maintenance. management. The data perception layer includes multi-dimensional data collection from various types of sensors and classified storage and transmission of different data sources; the digital twin data processing layer is based on the bridge digital twin modeling and performs related processing on the collected data, including data collection. , interactive fusion, lightweight and export analysis, etc., to build a comprehensive information database; the digital twin application layer includes the bridge structure digital twin, BIM system, GIS system, finite element simulation system, and database-based comprehensive operation and maintenance management system , intelligent early warning assessment system and other expanded applications of digital twins; intelligent construction and operation and maintenance management are based on the expanded applications of the digital twin integration platform, which can display multi-dimensional data in modules and achieve full life cycle management.

This invention puts forward higher requirements for modeling accuracy, performs iterative optimization of unit parameters and grid information on the model, and continuously updates the bridge information model in combination with comprehensive operation and maintenance information, mainly to build a more realistic digital twin for Conduct real-time finite element analysis, establish a more reliable historical information database, and provide realistic guarantees for bridge safety assessment and bridge response prediction under specific natural disasters or structural disease conditions. Through the application of drone oblique photography and close-up photography, precise real-life modeling of on-site bridges and damage identification of on-site bridge structures can be used to help operation and maintenance personnel identify easily damaged parts in a timely manner and provide corresponding maintenance measures or Construction plan.

The following is combined with Figure 2, taking a slow-moving bridge in a certain city as an example to illustrate the method of digital twin of a bridge structure based on BIM-GIS-FEM, which includes the following steps.

Step 1. Regularly reconstruct the real-life model of the bridge and its surrounding environment, including:

a. Five-way tilt photography and close-up photography. Five-way oblique photography is used to build rough models, and close-up photography is used to build fine models. During the construction phase, long-distance five-way tilt photography is regularly performed to visually monitor the project progress; during the operation and maintenance phase, close-up photography is performed based on the rough mold obtained from the five-way tilt photography to monitor the health of each component of the bridge.

Step 2: Establish the BIM three-dimensional information model of the bridge, including:

b. Bridge component information design classification: preliminary classification of bridge components based on complexity, and unified management based on component information classification and coding standards. The component information includes geometric dimensions and coordinate positions, elastic modulus, Poisson's ratio, density and strength material properties, etc.; the bridge component model is established using BIM modeling software that conforms to the drawing style of the traditional construction industry and the operating habits of engineering personnel. .

Step 3. The implementation process of the digital twin basic module, including:

c. Data collection: In order to collect the overall status data of the bridge, determine the main control points of slow-moving bridges in a certain city, and deploy relevant sensors according to actual requirements. These include temperature sensors and wind speed and direction meters arranged at the main span and tower of the bridge, static levels and fiber Bragg grating strain gauges arranged at the bottom of the bridge deck at the control points of the main structure of the bridge, and at the end supports and pier supports of the bridge. Displacement sensors and acceleration sensors at the base, cable force calculation edge equipment arranged on the main cables and tie rods, GNSS sensors arranged at the mid-span and at the top of the tower, infrared induction surveillance cameras arranged on the bridge deck for pedestrian identification, Pressure sensors arranged at the bridge entrance and exit gates to obtain the load size, drones and their attached cameras, etc.

d. Data transmission: The information collected by various sensors is transferred to the data processing unit through different regional forwarding nodes. The on-site working condition load information and various bridge response information can be remotely transmitted to the data processing unit through wireless communication; The force computing edge device can remotely transmit the calculated cable force or suspension safety status to the full life cycle operation and maintenance management module through wireless communication.

e. Data processing: The various types of information collected from the field are complex and need to be integrated, classified and lightweight to improve the usefulness of the information. This embodiment suggests using a compression algorithm as a lightweight means to perform deviation detection processing and compression filtering simplification on sequential continuous variables, which can not only accurately reflect the actual trend of the data, but also significantly reduce the storage space of information data. The bridge information processed through lightweight data includes real-time working condition load information for finite element numerical analysis, on-site bridge response information for various types of bridges for comparison with numerical analysis results under specific working conditions, and for improving and updating bridges. Related operation and maintenance information of BIM information, etc. The integrated classification is to store specific different types of data in different databases in order to facilitate subsequent data analysis and retrieval.

Step 4. The basic process of digital twin data exchange and optimization module, including:

f. Secondary development plug-in application based on BIM3D API: The data interaction function of BIM3D API can be used by users for secondary development applications. It mainly includes: obtaining the geometry and related parameter data of components, creating or modifying model elements, and creating tools that can be used to quickly Implement UI plug-ins for repeated operation commands and implement functions such as model information sharing. You can try to use the C# language with superior operating capabilities and language features for secondary development of plug-ins, and select appropriate development tools according to the BIM3D software supporting requirements. In turn, the geometric information (component size and coordinate positioning), physical information (elastic modulus, Poisson's ratio, material density, etc.), and unit information (unit type and shape, mesh type, and division fineness, etc.) of the BIM3D three-dimensional model are exported. And the boundary condition processing information of the model, etc., are converted into command flow files corresponding to the finite element software NERAP. This case uses the NERAP version of the command flow standard to define the txt output format, and finally generates a txt document containing the necessary information for finite element analysis.

g. Unit parameters and mesh optimization of the finite element model: Since the geometric parameters and material properties of the bridge structure are constantly changing, the relevant operation and maintenance data need to be updated regularly and reflected in the building information model in a timely manner so that Use secondary development plug-ins to repeatedly convert BIM-related information and update finite element models. On this basis, by comparing the error between the bridge response results of finite element numerical analysis under different working conditions and the on-site measured bridge response, a combination optimization analysis is performed by setting variables such as unit parameters and grid types and combining with intelligent optimization algorithms. Finally determine the variable value range that meets the bridge response error requirements under various working conditions. Different unit types (such as various solid, rod, beam, plate, and shell units) can be selected for different model depth requirements. The applicable simulation ranges of these unit types are different; for grid type and grid size When selecting, the solid unit can choose tetrahedrons or hexahedrons with different numbers of nodes. Whether the mesh size is reasonable or not can be judged by the change range of the stress and strain values of the control part after adjusting the mesh fineness twice. If the stress and strain values change within Within 5%, the fineness value is considered relatively reasonable.

h. Finite element model modification and analysis result data transfer: Since the iterative process of unit parameters and grid optimization requires repeated modification of relevant parameters many times, it is obviously not convenient to modify the unit-related information of the building information model, so it can also be used in Model correction can be achieved by directly modifying unit-related parameters in the finite element software NERAP. Through the optimization process of model geometric information, unit information and related topology information, a more accurate and reliable digital twin model is generated and load simulation analysis is performed, and then error comparison and analysis of the displacement bridge response under specific working conditions are performed. The pre-processing related information input by the finite element model and the model analysis post-processing result information can be exported through the built-in command flow function of the finite element software NERAP to generate an inp file that can be retrieved and used by other safety calculation analysis modules; the solution results are in a res file It can be saved and can be called by the bridge life cycle operation and maintenance management module to perform data reconstruction and platform display of the three-dimensional model stress or displacement cloud diagram. When the bridge's geometric information, topology information and structural analysis-related information (such as unit mesh, material properties, working load and boundary processing information) need to be changed, updated inp files and res files can be generated to overwrite the last output. result.

Step 5: Build an intelligent operation and maintenance management platform based on the digital twin.

As shown in Figure 2, the digital twin application module can realize information transfer through the information transfer module and the full life cycle information operation and maintenance management module. The digital twin application module lies in the realization of intelligent functional goals; the full life cycle information operation and maintenance management module It lies in the functions of collection and integration, storage and backup, calling and viewing, and interactive expansion of various types of data; the information transfer module is used for the exchange of application information in the digital twin layer. The above three modules can be used for hierarchical integration to build an intelligent integration platform based on digital twins, thereby realizing intelligent operation and maintenance management of the entire life cycle information of the entire bridge project. As shown in Figure 3, an intelligent operation and maintenance management platform is built based on the hierarchical architecture of the digital twin. This platform integrates the hardware and software foundation of the digital twin and mainly includes the data perception layer, the digital twin data processing layer, and the digital twin application layer. and intelligent construction and operations management.

i. Data perception layer and digital twin data processing layer: The data perception layer can comprehensively perceive and collect the areas required for detection based on different information sources, mainly including various types of IoT sensors, cable force calculation edge devices, and RTK positioning. Technical drones, pedestrian recognition and trajectory tracking modules, intelligent recognition modules for human-machine, material, law, and environment, and ships; among them, the environment and bridge structure monitoring module is based on environmental detectors and DJI drone point cloud scanning, mainly for bridges Detect the internal and external environment and the construction status of the main structure, including the air quality of the surrounding environment, temperature or humidity changes around the bridge, sunshine time under different climate conditions and the real-time construction progress of the bridge structure; intelligent identification module for man-machine, material, law and environment Based on biometric identification technology and satellite positioning technology, intelligent identification and rapid positioning of personnel, equipment, and raw materials entering and exiting the construction site are carried out, and then relevant construction method standards and environmental impact conditions are evaluated, analyzed, and managed as a whole; pedestrian identification and trajectory tracking The module can systematically make full use of the timing and spatial relationships of surveillance videos to optimize and infer the multi-camera network topology. The main function is to design and optimize the pedestrian recognition deep learning model to realize the pedestrian recognition function based on the appearance of pedestrians. It is implemented on the basis of the model. The dynamic delay topology network model between multiple cameras provides a more optimized detection sequence for the search, thereby reducing the blindness of the search; the digital twin data processing layer is based on the bridge digital twin modeling to collect comprehensively perceived data Organize, integrate and exchange, lightweight preprocessing, and finally establish an effective information database according to different module requirements. The specific implementation process of the data perception layer and digital twin data processing layer is as described in steps two and three. The basic implementation process of the intelligent operation and maintenance management platform based on the digital twin hierarchical architecture is introduced below.

j. Digital twin application layer: The digital twin application layer is based on the digital twin application module. Based on the hardware and software implementation requirements of the module's related functions, it conducts integrated analysis based on the received data preprocessing layer information. The digital twin application layer takes the digital twin of the bridge structure as the application main body and can implement BIM system, GIS system, finite element simulation system, database-based comprehensive operation and maintenance management system, preventive assessment system, maintenance plan formulation system and Digital twin expansion functions, etc.

The specific process includes: The development of the operation and maintenance management system mainly involves steps such as script language selection, server development and database construction. This embodiment proposes to build a free and open source MySQL database for full life cycle operation and maintenance based on the initial information of 6D BIM, on-site comprehensive operation and maintenance information and effective information transferred from the digital twin application module, combined with historical samples of bridge responses under different working conditions. Information database; Apache, which has open source code and supports cross-platform applications, is used as the Web server software in high-configuration computers to facilitate browsing and querying related information on different mobile devices; PHP (Hypertext Preprocessor) is selected as the script for Web development Language, PHP language supports most operating systems and databases. Based on the development combination of PHP+Apache+MySQL, project development and specific function design of dynamic websites can be carried out on the Windows operating system, and finally a comprehensive operation and maintenance management system based on the database can be built. The main contents of the preventive assessment system include intelligent prediction and safety assessment of structural bridge reactions, effective identification of structural damage parts, and mainly involve knowledge bases such as intelligent optimization algorithms, modal parameter identification, and structural damage identification. The maintenance plan system mainly includes the automated generation of maintenance measures or construction plans, etc., and mainly involves knowledge bases such as machine learning and knowledge graph construction. This implementation case recommends setting the corresponding design variables and objective functions based on the res result file, establishing a mapping relationship between changing parameters such as structural parameters or material properties of different components and the deformation of various structural parts, and using intelligent optimization algorithms (such as deep learning CNN) predicts the bridge response of the bridge structure under specific working conditions, and uses finite element analysis results to verify the rationality of parameter settings and the accuracy of deformation prediction. Deep learning CNN is mainly used to process data with a grid-like structure, and has significant advantages for the analysis and recognition of time series real-time information and image data. Using the limit control method within the reasonable range of different bridge response variables, the safety, applicability and durability of the prediction results of the structural bridge response are comprehensively evaluated, and automated reports, maintenance measures, construction suggestions and evaluation reports are generated. Based on the structural damage identification process of the bridge response measured on site, this case recommends a detection method that first designs a modal parameter extraction method to obtain a modal parameter identification model, and then identifies structural damage based on the modal parameters. This detection method can identify structural damage parts and calculate the damage index, and then judge the severity of structural damage based on the damage index, and then obtain relevant detection reports of structural damage, and generate preventive maintenance decisions and health assessment reports for the bridge structure.

On the basis of the correction and improvement of the digital twin, the finite element numerical simulation method can be used to accurately analyze the structural bridge response under on-site working conditions in real time. This is the basic function of the digital twin's "real-time perception"; different engineering models can be established based on historical samples. The bridge response database is the basic function of "benchmark modeling" of the digital twin; the combination of the database and intelligent optimization algorithms for intelligent prediction and safety assessment is the basic function of "accurate prediction" of the digital twin. When there are changes in the bridge operation and maintenance information or bridge structure mechanical performance parameters monitored on site, such as component geometric size errors, changes in material properties, new structural diseases, changes in supports, etc., this information will be adjusted through component details, materials, etc. Manual information input methods such as elastic modulus and stiffness adjustment, component connection methods and spatial position adjustment are reflected in the update process of BIM system information, and then the finite element analysis model is regenerated through the conversion method of secondary development plug-ins; and for on-site working conditions Changes in operation and maintenance information such as structural boundary conditions and structural boundary conditions can be directly applied through the finite element software NERAP for corresponding load application and boundary unit adjustment. After an iterative unit parameter optimization process, information that meets specific conditions is finally transmitted to the comprehensive operation and maintenance system. Perform storage applications.

k. Intelligent construction and operation and maintenance management: Intelligent construction and operation and maintenance management can conduct comprehensive information interaction and integration with the digital twin application layer, fully combine the advantages of computer hardware and software and application development, and build an intelligent construction and operation and maintenance management system that meets the requirements of full life cycle operation and maintenance. Digital twin fusion platform. This digital twin integration platform includes information modules at different stages such as planning and design, construction management, operation and maintenance, etc. It integrates information modules from different dimensions such as 3D building model, 4D construction progress, 5D project quantity and cost, 6D structural analysis, etc., to develop various departments. Modules such as information entry management, collaborative design and communication of each sub-item.

Regarding the information system construction and implementation process of the digital twin fusion platform, it mainly includes the invocation of the database and the function development of the management platform. The related functions of the management platform can be developed through network editing technology, and the collection, storage and analysis of data can be realized through database system development technology. , using API technology to realize lightweight processing of bridge BIM models and convenient display on the Web, realizing the combination of lightweight 3D models and multi-dimensional information at each stage, and pursuing the maximum application of visual information. This embodiment recommends adopting a browser/server network system architecture, using PHP network programming to develop the basic application of the digital twin intelligent operation and maintenance management platform BIM-MAG through unified HTML and Javascript languages, and using MySQL database system development technology Realize the collection, storage and analysis of data, and combine it with the BIM model processed through lightweight technologies such as WebGL, and finally realize the unified storage of multi-dimensional data at different stages on the server side. Clients can send requests to the backend server in the form of domain names through the Web terminals of different devices and retrieve and access specific information. The system can store basic engineering information, operation and maintenance management information, and analysis result information on the server for later use. Customers perform various operations, multi-dimensional browsing and personalized processing on the BIM-MAG management platform. The basic system framework is divided into Web layer, business function layer and data access layer. The Web layer supports users to perform operations such as information query and information entry management; the business function layer is used to implement business and data rules and is mainly responsible for the processing and implementation of business functions. ; The data access layer is used to integrate and store data information and is responsible for database access and invocation, which mainly includes entity three-dimensional model information in the bridge design and construction stage, various monitoring information in the bridge operation and maintenance stage, and result information of digital twin application analysis, etc.

In addition, a variety of extended functions can be developed based on the intelligent operation and maintenance management platform BIM-MAG, mainly including: scene loading, hierarchical browsing, object access and search and other query functions, as well as various control effects and controls for different objects. Animation demonstration function, as well as various visualization functions such as camera perspective control, dynamic display of interface data, temperature and humidity cloud diagrams, and particle effect diagrams. Building a digital twin intelligent operation and maintenance management platform for large-span complex bridge structures will facilitate the integration and application of information technology in full life cycle management and promote the development of the construction industry from simple three-dimensional informatization to twin digitalization.

Those skilled in the art can easily understand that the above-mentioned large-span hybrid beam suspension bridge with special-shaped towers is only a preferred embodiment of the present invention and is not intended to limit the present invention. Everything is done within the spirit and principles of the present invention. Relevant modifications, equivalent substitutions or improvements, etc., all fall within the protection scope of the present invention.

Claims (10)

1. An urban slow-moving bridge health monitoring and digital twin system, which is characterized by: including a BIM modeling module, a real-life model module, a full life cycle operation and maintenance management module, a digital twin basic module, and a digital twin data conversion and optimization module. Digital twin application module; The BIM modeling module includes unified naming and annotation standard units for component information design classification, a modeling module for creating complex curved surface components of bridges, a modeling module for creating regular components of bridges, and a modeling module for creating bridges with regular components. Model conversion module for converting component information between components; The real-life model module includes a UAV module, an apron module, a live streaming module, a three-dimensional reconstruction module, and a format conversion module; the UAV module integrates RTK positioning and high-resolution cameras; the apron module implements UAV It has a low-battery automatic return-to-home charging function and an operating environment sensing function. The tarmac senses wind speed, rain, and light flight conditions to determine whether flight conditions are met and realize automated operations; the live streaming module uses drones for real-time monitoring and remote viewing at any time. Live work, pictures, and video results are automatically returned and archived, making it convenient to call pictures for three-dimensional reconstruction during real-life modeling; The three-dimensional reconstruction module finally forms a real-life model by adding photos and coordinates, adding image control points and their coordinates, aligning photos, building dense point clouds, generating grids, and generating textures. The real-life model is a collection of multi-angle pictures of physical entities. Perform three-dimensional reconstruction; the format conversion module converts the real-life model in OSGB format into 3D Tiles format; The full life cycle operation and maintenance management module includes a 6D BIM initial information unit for comprehensive information application, a comprehensive operation and maintenance information unit for bridge information update and correction, and a full life cycle information operation and maintenance for digital twin application results management. Management module; The digital twin basic module includes various sensor units used for collecting on-site working conditions and bridge response information during the construction phase, various sensor units used for real-time identification and tracking during the operation and maintenance phase and collecting the load distribution of slow-moving bridges, and various sensor units used for transmitting information. Wireless communication transmission unit, data processing unit for processing information, bridge suspension cable force calculation edge device unit; The digital twin data conversion and optimization module includes a 6D BIM preliminary information extraction and conversion unit including unit information, a working condition information unit used for numerical analysis of finite element structures, and a slow-moving bridge under real-time load conditions. Structural reaction force calculation and early warning unit, structural reaction comparison unit for simulation analysis under different working conditions, unit parameters and grid optimization unit; The digital twin application module includes a digital twin application module that combines finite element analysis and intelligent analysis, and an information transfer module for digital twin layer application information interaction. 2. Implement an urban slow-moving bridge health monitoring and digital twin system according to claim 1, characterized by: 6DBIM initial information unit includes bridge real-life model and structural model information (3D), overall progress planning information (4D), Engineering quantity and cost information (5D), structural analysis related information (6D), of which structural analysis related information (6D) includes unit parameters and mesh division information, material property information, working condition load information, boundary condition processing information and structure Surface crack damage information; The full life cycle information operation and maintenance management module builds a comprehensive operation and maintenance information database through the collection and integration of various types of data, data lightweighting and storage backup, and uniformly updates and stores multi-dimensional data at different stages on the server side for customers to use. The web side retrieves and accesses information, and combines the retrieved information with the BIM lightweight model for viewing and output; The multi-dimensional information of the initial information unit of 6D BIM is added and improved based on the three-dimensional model information. After being lightweight, the multi-dimensional information interacts with the life cycle information operation and maintenance management module. 3. An urban slow-moving bridge health monitoring and digital twin system according to claim 2, characterized by: The digital twin application module includes finite element numerical analysis results calculated based on real-time load conditions, a working condition bridge response database for intelligent analysis based on historical sample analysis results, an intelligent prediction and safety assessment unit based on the database, and a volume-based The structural damage identification unit of the cumulative neural network and the remaining effective information units of the digital twin application layer are included. Among them, intelligent prediction includes extracting the internal force results of the bridge under similar working conditions at different times, analyzing the internal force changes in key parts of the bridge, and predicting the overall internal force distribution; The information transfer module transfers and transmits the information in the digital twin application module to the life cycle information operation and maintenance management module, and then realizes the integration with the 6D BIM initial information unit through the information viewing and output unit of the life cycle information operation and maintenance management module. information interaction between. 4. An urban slow-moving bridge health monitoring and digital twin system according to claim 3, characterized in that: the remaining effective information units of the digital twin application layer include BIM structures based on optimized finite element model units and grid parameters. Analysis related information is updated to provide the value range of initial unit and grid parameters for the sixth dimension structural analysis information of the initial information unit of 6D BIM; based on the response comparison and error analysis of various types of bridges under different working conditions, the finite element The structural analysis model is used to evaluate boundary type selection and boundary simplification, load simulation effectiveness, and bridge response monitoring layout points; based on the results of finite element analysis and intelligent analysis, preventive maintenance decisions and evaluation documents for the bridge structure are automatically generated; The intelligent analysis optimizes the finite element model boundary conditions, unit parameters and grid information through the working condition response database of historical sample analysis results, so that the finite element results are close to the actual measured values. 5. An urban slow-moving bridge health monitoring and digital twin system according to claim 3, characterized in that: the structural damage identification unit based on the convolutional neural network includes the following steps: Data collection and preprocessing: A data set containing structural surface images is collected by drone close-up photography, and the data set is preprocessed, including image resizing, grayscale, and denoising operations; Divide the data and set up the CNN network architecture model: divide the data set into a training set, a validation set and a test set, set the number and size of the convolutional layers, activation function, pooling layer type and parameters; network training: use training Sets are used to train the CNN. During the training process, the network parameters are updated through the back propagation algorithm to minimize the loss function; Model evaluation: Use an independent test data set to evaluate the performance of the network. The evaluation index is accuracy. Hyperparameter tuning: Based on the performance of the validation set, adjust the network architecture and hyperparameters, including learning rate, batch size, and regularization parameters; Model testing: Use the test set to evaluate the performance of the final model. 6. An urban slow-moving bridge health monitoring and digital twin system according to claim 1, characterized in that: the bridge suspension cable force calculation edge device unit includes a digital twin operating system (DTOS), an acceleration sensor, and a frequency domain data filtering Processing and cable force calculation module, suspension cable stress status wireless communication transmission unit; multiple edge devices are connected through high-speed data communication protocols to form digital twin edge equipment, and the digital twin is created through the mobile client APP, PC client APP or web version. Edge equipment wireless remote management platform provides unified management; The frequency domain data filtering processing and cable force calculation module collects the time domain data of the target cable at preset time intervals, and uses the fast Fourier transform method to convert the time domain data into frequency domain data to construct a spectrogram, The second-order derivative method is used to identify all the wave peak data in the spectrum diagram, and then the cable force of the target cable is calculated through the wave peak data. 7. An urban slow-moving bridge health monitoring and digital twin system according to claim 1, characterized in that: the calculation and early warning unit for slow-moving bridge structural reaction force under real-time load conditions includes a cross-camera Pedestrian recognition and trajectory tracking module, anonymous pedestrian weight detection module, finite element numerical analysis module and result output module, finite element result error evaluation module, and threshold alarm module; The cross-camera pedestrian recognition and trajectory tracking module uses the timing and spatial relationships of surveillance videos to optimize and infer the multi-camera network topology design and optimize the pedestrian recognition deep learning model, thereby realizing the pedestrian recognition function based on the appearance of pedestrians. In the pedestrian recognition deep learning model Based on this, the dynamic delay topology network model between multiple cameras is implemented to provide a more optimized detection sequence for the search, thereby reducing the blindness of the search. 8. An urban slow-traffic bridge health monitoring and digital twin system according to claim 1, characterized in that: the update and iterative implementation process of the unit parameters and grid optimization units includes two levels: through comprehensive operation and maintenance information The unit updates on-site operation and maintenance information and performs unit parameter and grid optimization and iterative update of unit information through comparative analysis of bridge response errors under different working conditions. Among them, the update content of the comprehensive operation and maintenance information unit includes the collection and import of on-site working condition information. , collection and comparative analysis of various bridge response information, comprehensive operation and maintenance information of geometric parameters and material properties obtained through on-site appearance measurement and component testing. 9. A method for implementing an urban slow-traffic bridge health monitoring and digital twin system according to claim 1, characterized by comprising the following steps: Step 1: Reconstruct the real-life model of the bridge and its surrounding environment to reflect the real physical entities; Step 2: Establish a BIM three-dimensional information model of the bridge structure; Step 3: Build the digital twin base layer, which specifically includes collecting the overall status data of the bridge, transmitting the data to the data processing unit and lightweight data processing; Step 4: Convert and optimize the digital twin data; Step 5: Build an intelligent operation and maintenance management platform based on the digital twin hierarchical architecture. Specifically, through the information transfer module, the information transfer and comprehensive application between the digital twin application module and the full life cycle information operation and maintenance management module are realized. . 10. An urban slow-traffic bridge health monitoring and digital twin system according to claim 9, characterized in that: the intelligent operation and maintenance management platform combines a digital twin application module, an information transfer module, and a full life cycle information operation and maintenance module. The management module is layered, specifically divided into data perception layer, digital twin data processing layer, digital twin application layer and intelligent construction and operation and maintenance management layer; The data sensing layer includes various types of IoT sensors, bridge suspension cable force calculation edge devices, RTK positioning technology drones, environment and bridge structure monitoring modules, human-machine, material, law and environmental intelligent identification modules, and pedestrian identification and trajectory tracking modules; The digital twin data processing layer includes bridge digital twin modeling, precise real-life three-dimensional modeling modules, on-site load condition response data, mechanical performance testing data and other effective operation and maintenance data, and collects, organizes, integrates and exchanges the data. , lightweight preprocessing, and finally establish an effective information database according to different module requirements; The digital twin application layer includes the bridge structure digital twin, BIM system, GIS system, finite element simulation system, database-based comprehensive operation and maintenance management system, preventive assessment system based on intelligent early warning analysis, knowledge map-based And machine learning maintenance plan formulation system and other digital twin expansion applications, damage identification of structural surface cracks based on convolutional neural network; The intelligent construction and operation and maintenance management layer includes multi-dimensional information such as three-dimensional models, construction progress information, project quantity and cost information, and structural analysis divided into different information dimensions, and multi-stage information such as planning and design, construction management, and operation and maintenance divided into different construction stages. , as well as functional modules such as information entry and management, collaborative communication of various sub-items, and preventive maintenance decision-making; The output end of the data sensing layer is connected to the input end of the digital twin data processing layer, the output end of the digital twin data processing layer is connected to the input end of the digital twin application layer, and the output end of the digital twin application layer is connected to the intelligent construction and The input terminal of the operation and maintenance management layer is connected.