A monitoring system for long-span multi-tower cable-stayed bridges

Abstract

The application discloses a kind of large-span multi-tower cable-stayed bridges monitoring system, it is related to bridge monitoring technical field, and its technical solution key points are: including establishing benchmark finite element model, through sensing layer is laid in bridge key section, for collecting original monitoring signal, data acquisition and transmission layer carries out digitization processing to original monitoring signal and realizes real-time transmission, while data processing and evaluation layer carries out online quality evaluation to mass data, purifies reconstruction and structure state parameter identification, realizes the visualization of data on visualized application layer and multi-stage early warning visualization management.The application realizes the full coverage of public traffic layer large-span multi-tower cable-stayed bridge, high-precision identification and intelligent management by multi-source fusion sensing, multi-level data purification and star-ground collaborative monitoring, significantly improves the reliability and operation efficiency of structure safety evaluation.

Description

Technical Field

[0001] This invention relates to the field of bridge monitoring technology, and more specifically, to a monitoring system for long-span multi-tower cable-stayed bridges. Background Technology

[0002] Bridge structural health monitoring technology is a crucial means of ensuring the long-term operational safety of large bridges. With the rapid development of transportation infrastructure, long-span bridges, especially multi-tower cable-stayed bridges with combined road and rail structures, face higher demands for structural safety monitoring due to their highly innovative structures, complex stresses, and unique load distributions. Establishing a comprehensive health monitoring system allows for real-time monitoring of the bridge's structural response and development trends under complex operational conditions, providing a scientific basis for bridge operation management, maintenance, and safety assessment.

[0003] Since the mid-to-late 1980s, developed countries in Europe and America have successively established bridge health monitoring systems of various scales. The UK deployed sensors on the Foyle Bridge to monitor the vibration, deflection, and strain of the main girder; Norway, the US, Denmark, and other countries have also successively established health monitoring systems on large bridges such as the Skernsundet cable-stayed bridge, the Sunshine Skyway Bridge cable-stayed bridge, and the Great Belt East suspension bridge. Since the 1990s, my country has successively established structural monitoring systems of different scales on large bridges such as the Tsing Ma Bridge and Kap Shui Mun Bridge in Hong Kong, as well as the Humen Bridge and the Jiangyin Yangtze River Bridge. These systems mainly use traditional point sensors such as accelerometers, strain gauges, and displacement gauges to monitor the response of key bridge sections.

[0004] However, existing technologies still have the following shortcomings: First, traditional monitoring systems mostly use point-based sensor deployment, making it difficult to obtain overall spatial deformation information and continuous strain field distribution of the structure, resulting in a relatively singular monitoring dimension; Second, faced with massive monitoring data, there is a lack of effective online quality assessment and noise reduction mechanisms, leading to low data utilization and difficulty in extracting high-order modal information submerged by noise; Third, for bridge types such as long-span multi-tower cable-stayed bridges with complex load distribution and special structural responses, existing cable force identification and dynamic deformation monitoring are insufficient in accuracy and stability; Fourth, monitoring data is disconnected from physical bridge models, lacking intuitive, lightweight, three-dimensional visualization management methods, making it difficult for monitoring results to effectively guide maintenance decisions.

[0005] Therefore, this invention addresses the structural characteristics and operational needs of long-span multi-tower cable-stayed bridges with both road and rail on the same level. It provides a monitoring system for long-span multi-tower cable-stayed bridges, constructing an intelligent monitoring system capable of achieving full coverage, data purification, high-precision identification, and visualized management, thereby solving the aforementioned problems. Summary of the Invention

[0006] The purpose of this invention is to provide a monitoring system for long-span multi-tower cable-stayed bridges. This invention achieves full coverage, high-precision identification, and intelligent management of long-span multi-tower cable-stayed bridges with both road and rail traffic on the same level through multi-source fusion sensing, multi-level data purification, and satellite-ground collaborative monitoring, which significantly improves the reliability of structural safety assessment and operation and maintenance efficiency.

[0007] The above-mentioned technical objective of the present invention is achieved through the following technical solution: a monitoring system for a long-span multi-tower cable-stayed bridge, comprising a sensing layer, a data acquisition and transmission layer, a data processing and evaluation layer, and a visualization application layer;

[0008] The sensing layer is deployed at key sections of the bridge to collect various raw monitoring signals of the bridge in real time during its operation.

[0009] The data acquisition and transmission layer is connected to the sensing layer and includes multiple data acquisition modules and a communication network; the data acquisition modules are used to receive and digitize the raw monitoring signals acquired by the sensing layer, and the communication network is used to transmit the digitized monitoring data in real time.

[0010] The data processing and evaluation layer is connected to the data acquisition and transmission layer, and includes a data quality evaluation module, a signal processing and calculation module, and a database server connected in sequence. The data quality evaluation module is used to perform online quality evaluation and purification and reconstruction on the received massive monitoring data, and transmit the purified data to the signal processing and calculation module. The signal processing and calculation module is used to identify the structural state parameters of the bridge based on the purified data. The database server is connected to both the data quality evaluation module and the signal processing and calculation module, and is used to store the original monitoring data, intermediate processing results, and the finally identified structural state parameters.

[0011] The visualization application layer is connected to the data processing and evaluation layer, including a building information model-geographic information system visualization platform and a monitoring and operation and maintenance management platform. The visualization application layer is used to receive the structural state parameters output by the data processing and evaluation layer and map them onto the bridge's three-dimensional model to realize the visualization display of monitoring data and multi-level early warning management.

[0012] The present invention is further configured such that: the sensing layer includes a cable force monitoring module, a dynamic deformation monitoring module, an intelligent fiber sensing module, an environmental and load monitoring module, and a synthetic aperture radar interferometric remote sensing monitoring module;

[0013] The cable tension monitoring module is used to collect vibration or strain signals of the stay cables;

[0014] The dynamic deformation monitoring module is used to collect displacement, deflection, and vibration signals of the bridge's main beam and towers.

[0015] The intelligent fiber sensing module is used to collect continuous strain field information along the longitudinal direction of the main beam;

[0016] The environment and load monitoring module is used to collect information on wind speed, temperature, and vehicle load on the bridge deck.

[0017] The synthetic aperture radar interferometric remote sensing monitoring module is used to acquire the overall deformation information of the bridge based on a spaceborne or ground-based synthetic aperture radar interferometric system.

[0018] The present invention is further configured such that: the data quality assessment module includes an online quality assessment unit, an abnormal data classification and processing unit, an effective signal filtering unit, and a signal adaptive decomposition, purification, and reconstruction unit;

[0019] The online quality assessment unit is used to perform packet loss rate checks, range determination, and signal-to-noise ratio analysis on the received monitoring signals.

[0020] The abnormal data classification and processing unit is connected to the online quality assessment unit and is used to distinguish the marked abnormal data into sensor fault data or sudden interference event data.

[0021] The effective signal filtering unit is connected to the abnormal data classification and processing unit, and is used to filter test signals containing structural health status information based on the statistical difference between effective signals and interference noise.

[0022] The adaptive signal decomposition and purification reconstruction unit is connected to the effective signal screening unit and is used to perform empirical mode decomposition on the screened signal, remove noise components, and then reconstruct the purified signal.

[0023] The present invention is further configured such that: the signal processing and calculation module includes a cable force identification unit, a deformation and mode identification unit, a load inversion unit, and a synthetic aperture radar interferometric data processing unit;

[0024] The cable force identification unit is used to input the purified cable acceleration signal into the cable force identification algorithm, and calculate the current cable force value by combining the boundary conditions corrected by the finite element model.

[0025] The deformation and modal identification unit is used to obtain the deflection curve of the main beam based on data from the Global Navigation Satellite System and the connecting pipe, and to identify the strain modes of the structure based on data from long gauge length strain sensors.

[0026] The load inversion unit is used to combine dynamic weighing system data and structural response data collected by long gauge length sensors to invert the axle load on each lane in real time and establish a spatiotemporal distribution map of vehicles on the bridge deck.

[0027] The synthetic aperture radar interferometric data processing unit is used to perform time-series analysis on multiple radar images to establish a long-term deformation time series of the bridge.

[0028] The present invention is further configured such that: the monitoring and operation and maintenance management platform includes a multi-level early warning unit and a digital-analog linkage unit;

[0029] The multi-level early warning unit is used to trigger yellow, orange, or red early warnings based on the comparison results of the monitored values ​​with theoretical thresholds and historical statistical thresholds, and automatically locate the corresponding component positions in the three-dimensional model;

[0030] The digital-analog linkage unit is used to realize the real-time mapping of monitoring data and building information model components, and supports clicking on the sensor icon on the model to retrieve historical time history curves, spectrum analysis diagrams and equipment file information.

[0031] This invention also provides a monitoring method for long-span multi-tower cable-stayed bridges, comprising the following steps:

[0032] S1. Establish a benchmark finite element model for a long-span multi-tower cable-stayed bridge with both road and rail on the same level, and determine the initial theoretical thresholds for key control sections and monitoring indicators;

[0033] S2. Deploy a hybrid sensor network at key sections of the bridge;

[0034] S3. The data acquisition module collects raw signals according to a preset triggering mechanism and transmits them to the monitoring center in real time through the communication network.

[0035] S4. Perform online quality assessment, purification, and reconstruction of the massive amounts of received monitoring signals;

[0036] S5. High-precision identification of structural state parameters based on the purified signal;

[0037] S6. Acquire multiple radar satellite images, perform refined identification and time series analysis of synthetic aperture radar interferometric signals, and establish a long-term deformation time series of the bridge.

[0038] S7. Compare the identified current structural state parameters with the theoretical threshold and historical statistical threshold set in step S1 to trigger multi-level early warning;

[0039] S8. The processing results from steps S4 to S7 are mapped onto the lightweight 3D model of the bridge in real time to realize the visualization management of monitoring data based on Building Information Modeling-Geographic Information System.

[0040] The present invention is further configured such that the hybrid sensor network deployed in step S2 includes: a long gauge-length fiber optic grating sensor deployed along the longitudinal direction of the main beam, an accelerometer installed on the stay cables, a GPS-BeiDou displacement monitoring station deployed at the top of the tower and the middle of the main beam span, a connecting pipe level deployed along the longitudinal direction of the main beam, and a dynamic weighing system deployed on the bridge deck.

[0041] The present invention is further configured such that step S4 specifically includes:

[0042] S41. Perform online quality assessment on the received signal, including checking the data packet loss rate, determining whether the signal exceeds the reasonable range of the sensor, and identifying whether the signal-to-noise ratio meets the standard.

[0043] S42. Classify the marked abnormal data into sensor fault data or sudden interference event data.

[0044] S43. Using exploratory data analysis methods, the statistical differences between effective signals and interference noise are used to screen out effective test signals that contain information about structural health status.

[0045] S44. Perform empirical mode decomposition on the selected effective signal to decompose the signal into intrinsic mode functions of different frequencies. After removing noise components, reconstruct the signal to obtain the purified signal.

[0046] The present invention is further configured such that the dynamic deformation and modal identification step S5 further includes: based on the data collected by the long gauge length strain sensor, using the dynamic analysis theory of regional sensing, identifying the strain modes of the structure, wherein the strain modes are used to analyze the local damage and low-frequency deformation characteristics of the wide cross-section main beam.

[0047] The present invention is further configured such that step S6 specifically includes:

[0048] S61. To address the issues of synthetic aperture radar signal overlay and low coherence that bridge structures are prone to, a structural semantic information model is introduced to refine the identification of stable point targets on bridges.

[0049] S62. Perform time-series synthetic aperture radar interferometry analysis on the identified point targets to establish a long-term deformation time series of the bridge, which is used to correct the finite element model in step S1 and cross-check with the discrete point data monitored by the global positioning system in step S5.

[0050] In summary, the present invention has the following beneficial effects:

[0051] 1. This invention constructs a composite sensing layer that includes cable tension monitoring, dynamic deformation monitoring, intelligent fiber sensing, environmental load monitoring, and synthetic aperture radar interferometric remote sensing monitoring. This achieves full-scale coverage from local point monitoring to regional distributed monitoring and then to overall deformation monitoring. Long gauge-length fiber optic grating sensors are continuously deployed along the longitudinal direction of the main beam to form a regional distributed sensing network, which solves the problem that traditional point sensors cannot obtain information on the overall strain field of the structure. The introduction of synthetic aperture radar interferometric remote sensing technology makes up for the deficiency of traditional displacement sensors that can only obtain discrete point displacements, providing spatiotemporally continuous data support for the long-term trend deformation analysis of bridges.

[0052] 2. This invention sets up a data quality assessment module in the data processing and evaluation layer, which includes four units: online quality assessment, abnormal data classification and processing, effective signal screening, and signal adaptive decomposition, purification and reconstruction. This forms a complete data "cleaning-screening-purification" processing chain. Through empirical mode decomposition and noise component removal, environmental noise and sensor random errors are effectively eliminated, making the subsequently identified structural state indicators such as cable force and modal parameters more stable and reliable. It can also uncover high-order bridge modal information that has been submerged by noise, providing a high-quality data foundation for structural safety assessment.

[0053] 3. In the signal processing and calculation module, the cable force identification unit, combined with the boundary conditions corrected by the finite element model, improves the accuracy and stability of cable force calculation; the deformation and modal identification unit identifies the strain modes of the structure based on long gauge length strain data. Compared with traditional acceleration modes, strain modes are more sensitive to local damage and low-frequency deformation of the main beam, and are particularly suitable for the complex stress characteristics analysis of wide-section main beams with the same floor as railway and highway; the load inversion unit, combined with dynamic weighing system data and structural response data, realizes the inversion of the spatiotemporal distribution of vehicle loads on the bridge deck, providing load input basis for in-depth analysis of structural mechanical behavior.

[0054] 4. This invention integrates two technical means: ground-based high-precision displacement monitoring (Global Navigation Satellite System, connecting pipe level) and spaceborne synthetic aperture radar interferometric remote sensing monitoring. The synthetic aperture radar interferometric data processing unit solves the technical problems of synthetic aperture radar signal superposition and low coherence in bridge structures by introducing a structural semantic information model. It realizes the fine identification of bridge point targets and temporal deformation analysis. This space-ground collaborative mechanism can not only verify each other, but also realize full-cycle deformation monitoring from millimeter-level instantaneous deformation to centimeter-level long-term trend.

[0055] 5. This invention incorporates a multi-level early warning unit into the monitoring and operation and maintenance management platform, establishing a three-level early warning mechanism of "daily statistical range - design specification limit - abnormal rate of change," realizing progressive alarms from data anomalies to structural risks. The digital-analog linkage unit maps the monitoring data to a lightweight building information model-geographic information system 3D model in real time, realizing intelligent management of "digital-analog linkage." Managers can intuitively retrieve historical time-history curves, spectrum analysis diagrams, and equipment file information by clicking on sensor icons on the model, greatly improving the readability of monitoring data and the convenience of operation and maintenance management.

[0056] 6. This invention is specifically designed for the structural characteristics and operational needs of long-span multi-tower cable-stayed bridges with both road and rail on the same level. Addressing the complex load distribution and unique structural response of this type of bridge, it acquires regional strain information through a long gauge-length sensor network, identifies axle load distribution through a dynamic weighing system, and grasps the modal characteristics of the wide-section main girder through strain modal analysis. This forms a complete technical closed loop of "load identification - response monitoring - condition assessment," providing a replicable technical reference for the design, construction, and operation and maintenance of similar bridges. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the module structure of a monitoring system for a long-span multi-tower cable-stayed bridge according to Embodiment 1 of the present invention;

[0058] Figure 2 This is a data flow diagram of a monitoring system for a long-span multi-tower cable-stayed bridge according to Embodiment 1 of the present invention;

[0059] Figure 3 This is a flowchart illustrating the interaction of a monitoring system for a long-span multi-tower cable-stayed bridge according to Embodiment 1 of the present invention.

[0060] Figure 4 This is a flowchart illustrating the steps of a monitoring method for a long-span multi-tower cable-stayed bridge according to Embodiment 2 of the present invention. Detailed Implementation

[0061] The following is in conjunction with the appendix Figures 1-4 The present invention will be described in further detail below.

[0062] Example 1: A monitoring system for a long-span multi-tower cable-stayed bridge, comprising a sensing layer, a data acquisition and transmission layer, a data processing and evaluation layer, and a visualization application layer.

[0063] In this embodiment, the sensing layer is deployed at key sections of the bridge to collect various raw monitoring signals of the bridge in real time during operation. The sensing layer specifically includes the following modules:

[0064] Cable stress monitoring module: includes an acceleration sensor installed on the surface of the cable and a fiber Bragg grating smart cable embedded in the cable structure. The acceleration sensor is used to collect the vibration signal of the cable, and the fiber Bragg grating smart cable is used to directly sense the strain change of the cable. In this embodiment, the sensor is connected to the nearest data acquisition module through a shielded cable.

[0065] The dynamic deformation monitoring module includes GPS-BeiDou displacement monitoring stations deployed at the top of the tower and mid-span of the main beam, a connecting pipe level instrument deployed longitudinally along the main beam, and a high-precision vibration sensor. The GPS-BeiDou monitoring station transmits positioning data to the data acquisition module via a wireless communication module; the connecting pipe level instrument connects to each measuring point via hydraulic pipelines, and its internal liquid level sensor uploads liquid level data via an RS485 bus.

[0066] Intelligent fiber sensing module: including long gauge-length fiber Bragg grating sensors, which are continuously deployed along the longitudinal direction of the main beam to form a regionally distributed sensing network. The sensors are connected in series through optical fibers and use wavelength division multiplexing technology to transmit strain signals from multiple measuring points on the same optical fiber, which are finally connected to the fiber Bragg grating demodulator.

[0067] Environmental and load monitoring module: This includes an anemometer, temperature sensor, and dynamic weighing system. The anemometer and temperature sensor are connected to the data acquisition module via an analog signal acquisition module; the dynamic weighing system is installed within the bridge deck pavement layer, generating pressure signals when vehicles pass over them, which are then processed by a charge amplifier and transmitted via industrial Ethernet.

[0068] Synthetic Aperture Radar Interferometric Remote Sensing Monitoring Module: Based on a spaceborne or ground-based synthetic aperture radar interferometric system, this module acquires radar satellite images or ground-based radar scan data periodically, processes the data in the background to obtain the overall deformation information of the bridge, and transmits the data directly to the database server of the data processing and evaluation layer without going through the on-site acquisition module.

[0069] In this embodiment, the data acquisition and transmission layer is connected to the sensing layer and includes multiple data acquisition modules and a communication network. The data acquisition modules (data acquisition instruments) are deployed nearby inside the bridge box girder or at the bridge tower. Their input ends are connected to the sensors of the sensing layer through shielded cables, optical fibers, or wireless communication. The data acquisition modules integrate signal conditioning circuits, analog-to-digital converters, and embedded processors, which are responsible for filtering, amplifying, and digitizing the analog signals output by the sensors to form a standard format monitoring data packet.

[0070] The communication network adopts a fiber optic redundant ring network architecture, consisting of industrial Ethernet switches and single-mode optical fibers. The output of the data acquisition module is connected to the nearest industrial Ethernet switch via a network cable. The switches are connected in series in a ring via optical fibers, and finally connected to the monitoring center through the core switch. When an optical fiber is interrupted at a point in the ring network, the network can automatically switch to the backup path within fifty milliseconds to ensure the reliability of data transmission. For remote measurement points, the communication network also includes a 4G / 5G wireless backup link, which automatically switches to wireless transmission mode when the wired network fails.

[0071] In this embodiment, the data processing and evaluation layer is located in the monitoring center computer room and is connected to the data acquisition and transmission layer through a core switch. It includes a data quality evaluation module, a signal processing and calculation module, and a database server connected in sequence.

[0072] The input of the data quality assessment module is connected to the output of the core switch, receiving all raw monitoring data streams from the field in real time. The data quality assessment module further includes:

[0073] Online Quality Assessment Unit: Performs real-time quality checks on received monitoring signals. This unit has a built-in quality assessment algorithm that calculates the packet loss rate frame by frame, determines whether the signal amplitude exceeds the sensor's reasonable range, and calculates whether the signal-to-noise ratio meets the standard. For data that meets the quality requirements, it is tagged "valid" and directly output to the next unit; for data that does not meet the quality requirements, it is tagged "abnormal" and forwarded to the abnormal data classification and processing unit.

[0074] Anomaly Data Classification and Processing Unit: Connected to the online quality assessment unit, this unit receives data tagged with "anomaly". The unit has a built-in classifier that classifies the anomaly data into two categories based on preset rules: one is "bad data" caused by sensor failure (such as continuous constant value or large jump), which is removed and an equipment maintenance work order is generated and sent to the visualization application layer; the other is "special event data" caused by sudden interference events (such as exceeding the limit data at the moment of earthquake or ship collision), which is stored separately as a special event file in a dedicated partition of the database server.

[0075] Effective signal screening unit: Connected to the online quality assessment unit and the abnormal data classification and processing unit, it receives data labeled "effective" and "special event data" retained after classification and processing. This unit uses exploratory data analysis methods to further screen out effective test signals that truly contain structural health status information by utilizing the statistical differences between effective signals and interference noise (including variance, mean shift, and autocorrelation function characteristics).

[0076] The adaptive signal decomposition and purification reconstruction unit is connected to the effective signal filtering unit. It receives the filtered effective signal and has a built-in empirical mode decomposition algorithm to adaptively decompose the non-stationary signal into several intrinsic mode functions. It identifies and removes high-frequency components that represent noise and ultra-low-frequency components that represent trend terms. The remaining intrinsic mode function components are linearly superimposed to reconstruct the purified signal. The purified signal is output to the signal processing and solution module and stored in the "purified data" partition of the database server.

[0077] The input of the signal processing and decoding module is connected to the output of the data quality assessment module (i.e., the signal adaptive decomposition and sanitization reconstruction unit), receiving the sanitized high-quality data. The signal processing and decoding module further includes:

[0078] Cable Force Identification Unit: This unit receives the purified cable acceleration signal and inputs it into the cable force identification algorithm. First, it obtains the cable's natural frequency through a Fast Fourier Transform. Then, combining the corrected parameters such as cable length, linear density, and boundary conditions from the finite element model, it calculates the current cable force value using string vibration theory. The calculation results are stored in real-time in the "Structural Parameters" partition of the database server and simultaneously output to the visualization application layer for display.

[0079] Deformation and Modal Recognition Unit: This unit receives purification data from GPS-BeiDou monitoring stations, connecting pipe levels, and long gauge-length fiber optic grating sensors. It uses a Kalman filter algorithm to fuse GPS and connecting pipe data to obtain the continuous deflection curve of the main beam. Simultaneously, based on the regional strain data collected by the long gauge-length strain sensor, it uses regional sensing dynamic analysis theory to identify the strain modes of the structure (including frequency, damping ratio, and mode shape). The identification results are stored in the database server and output to the visualization application layer.

[0080] Load Inversion Unit: Receives vehicle load data from the dynamic weighing system and structural response data from long gauge length sensors. Based on influence line theory and regularized inversion algorithm, this unit inverts the axle load (axle weight, wheelbase) on each lane in real time and establishes a spatiotemporal distribution map of vehicles on the bridge deck. The inversion results are used to correct the load input of the finite element model and are stored in the database server.

[0081] Synthetic Aperture Radar Interferometric Data Processing Unit: This unit periodically receives radar image data from the Synthetic Aperture Radar Interferometric Remote Sensing Monitoring Module via a remote data interface. It performs registration, interferometric processing, and time series analysis on multiple radar images, identifies stable point targets on the bridge, and establishes a long-term deformation time series of the bridge. The processing results are used to correct the finite element model and are also cross-checked with the GPS monitoring data output by the Deformation and Modal Identification Unit.

[0082] The database server adopts a dual-machine hot standby architecture, connecting to the data quality assessment module and the signal processing and calculation module respectively. Data read and write are realized through a high-speed storage area network. The database server internally establishes multiple logical partitions, including: raw data area (stores unprocessed raw monitoring data), purified data area (stores data purified by the quality assessment module), structural parameter area (stores identification results such as cable force, deflection, and modal characteristics), equipment file area (stores sensor parameters, calibration records, and maintenance work orders), alarm record area (stores all early warning event information), and 3D model area (stores bridge BIM model data).

[0083] In this embodiment, the visualization application layer is connected to the data processing and evaluation layer and deployed on the operation station and remote mobile terminal of the monitoring center. It includes a Building Information Modeling-Geographic Information System (BIM-GIS) visualization platform and a monitoring and operation management platform. The input end of the BIM-GIS visualization platform is connected to the database server and reads the data of the structural parameter area in real time through the data interface. The platform has a built-in 3D rendering engine that loads the lightweight bridge BIM model into the GIS map background and performs color rendering on the model according to the real-time data. In this embodiment, the color of the cable is changed according to the magnitude of the cable force and the position of the main beam is adjusted according to the deflection deformation.

[0084] The monitoring and maintenance management platform is integrated with the Building Information Modeling-Geographic Information System (BIM-GIS) visualization platform, providing a human-computer interaction interface. The monitoring and maintenance management platform internally includes:

[0085] Multi-level early warning unit: Connected to the "Structural Parameters" and "Alarm Records" areas of the database server, this unit has a built-in early warning rule engine that reads current cable force, deflection, modal frequency, and other parameters in real time and compares them with preset theoretical thresholds and historical statistical thresholds. When the monitored value exceeds the daily statistical range (in this embodiment, it is set to exceed the historical average ±2 standard deviations), a yellow warning is triggered; when the monitored value exceeds the design specification limit, an orange warning is triggered; when the rate of change of the monitored value is abnormal, i.e., it shows an accelerating growth trend for three consecutive monitoring cycles, a red warning is triggered. After the warning is triggered, the system automatically highlights the corresponding component location on the Building Information Modeling-Geographic Information System visualization platform and stores the alarm information in the "Alarm Records" area of ​​the database server.

[0086] Digital-Model Linkage Unit: Integrated with the Building Information Modeling-Geographic Information System visualization platform, it enables interactive data query. When an operator clicks on any sensor icon on the 3D model, the digital-model linkage unit retrieves the sensor's historical time-history curves, spectrum analysis diagrams, equipment file information, and historical maintenance records from the "Raw Data," "Clean Data," and "Structural Parameters" areas of the database server and displays them next to the model in a pop-up window.

[0087] The data transmission and processing flow of this system includes:

[0088] Data acquisition: The sensors in the sensing layer perceive the bridge status in real time and convert physical quantities into electrical or optical signals. The data acquisition module scans according to a preset sampling frequency. In this embodiment, the high-frequency channel vibration is 100Hz and the low-frequency channel temperature is 0.01Hz. The analog signals are converted into digital quantities and timestamps and sensor IDs are added and packaged into data frames.

[0089] Data Upload: The data acquisition module uploads data frames in real time through the communication network. For data with high real-time requirements (vibration, strain), UDP protocol multicast is used for transmission; for data with high reliability requirements (equipment status, alarm information), TCP protocol point-to-point transmission is used. The core switch distributes the received data stream to the data quality assessment module of the data processing and evaluation layer.

[0090] Data cleaning: The data quality assessment module sequentially processes the raw data through an online quality assessment unit, an abnormal data classification and processing unit, an effective signal screening unit, and a signal adaptive decomposition, purification, and reconstruction unit. This process removes noise, repairs anomalies, and purifies signals to generate high-quality "purified data."

[0091] Parameter identification: The signal processing and calculation module obtains purified data from the data quality assessment module, and performs parallel calculations by the cable force identification unit, deformation and mode identification unit, load inversion unit and synthetic aperture radar interferometric data processing unit to identify the current cable force value, deflection curve, strain mode, load distribution and long-term deformation trend of the structure.

[0092] Data storage: The database server simultaneously receives cleaned data from the data quality assessment module and structural parameters from the signal processing and solution module, stores them according to the preset data table structure, and creates indexes for fast retrieval.

[0093] Visualization: The Building Information Modeling-Geographic Information System (BIM-GIS) visualization platform reads the latest data from the database server in real time and dynamically updates the bridge's 3D model. The monitoring and maintenance management platform's multi-level early warning units continuously monitor data changes; once an early warning is triggered, it immediately displays a notification on the interface and pushes it to the maintenance personnel's mobile devices.

[0094] Closed-loop control: When the monitoring and operation and maintenance management platform detects that the data of a certain measuring point is continuously abnormal, it can adjust the sampling strategy of the corresponding data acquisition module through reverse control commands, change the regular sampling to encrypted sampling, or trigger the video surveillance near the measuring point to capture evidence, so as to realize the closed-loop management of "monitoring-analysis-control".

[0095] Example 2: A monitoring method for long-span multi-tower cable-stayed bridges, comprising the following steps:

[0096] S1. Establish a benchmark finite element model for a long-span multi-tower cable-stayed bridge with both road and rail on the same level, and determine the initial theoretical thresholds for key control sections and monitoring indicators.

[0097] In this embodiment, the design drawings, geological survey reports, material performance parameters (including the elastic modulus of concrete, yield strength of steel, elastic modulus of cables, etc.) and construction monitoring data of the bridge are collected first. The finite element analysis software (Abaqus) is used to establish a three-dimensional finite element model of the entire bridge. The model should accurately simulate the geometric dimensions, boundary conditions (bridge tower foundation constraints, support connection methods), material nonlinearity and geometric nonlinearity characteristics of the bridge.

[0098] Secondly, a full-process construction simulation was conducted, and the model calculation results were compared with the actual construction monitoring data. The model was corrected by adjusting the model parameters (including stiffness correction coefficients and boundary condition simulation methods) so that the model could accurately reflect the actual completed bridge state. The corrected model is the reference finite element model.

[0099] Finally, based on the benchmark finite element model, an operational state simulation analysis is conducted to determine the key control sections (mid-span of the main span, tower-beam junction, top of auxiliary piers, cable anchorage zone, etc.) and the initial theoretical thresholds of each monitoring index. The theoretical thresholds include: the maximum allowable deflection of the main beam under the combined action of design loads, the stress envelope value of the key sections, the range of cable force variation of the stay cables, the fundamental frequency range of the structure, etc. These thresholds will serve as the benchmark for early warning judgment in step S7.

[0100] Step S2: Deploy a hybrid sensor network at key sections of the bridge.

[0101] This step, based on the key control sections determined in step S1, deploys a multi-source fusion hybrid sensor network to ensure the comprehensiveness and redundancy of monitoring data. First, long gauge length fiber optic grating sensors are deployed along the key sections in the longitudinal direction of the main beam (including L / 4, L / 2, and 3L / 4, where L is the main span). During deployment, the sensors should be close to the lower edge of the main beam or the surface of the web, fixed with special adhesives or mechanical clamps, and connected in series with fiber optic jumpers to form a regionally distributed sensor network. The gauge length of each long gauge length sensor is determined according to the height of the main beam section and monitoring requirements, usually 1 to 2 meters.

[0102] Secondly, an accelerometer is installed on each cable about 2 to 5 meters from the bridge deck. For longer cables, a measuring point can be added in the middle of the cable. The accelerometer is fixed to the cable surface with a special clamp to ensure that the sensor vibrates at the same frequency as the cable. At the same time, fiber optic grating smart cables are pre-embedded or installed in some key cables to directly measure the cable strain and cross-check the results with the frequency method measurement of the accelerometer.

[0103] Next, GPS-BeiDou displacement monitoring stations were set up at the top of the bridge towers and at the mid-span of the main beam. The reference station was set up on stable bedrock near the bridge site. Connecting pipe leveling points were set up at regular intervals (100 to 200 meters) along the longitudinal direction of the main beam. The measuring points were connected by hydraulic pipelines to form a static leveling system. High-precision vibration sensors were set up at the bridge supports and expansion joints to monitor the dynamic response of the structure.

[0104] Then, a dynamic weighing system is installed on the bridge deck's driving lanes, typically using piezoelectric film or quartz sensors, embedded in the pavement layer to cover the entire lane width. Three-dimensional ultrasonic anemometers are installed at the top of the bridge towers and the cantilever ends of the main beams, and temperature sensors (air temperature sensors and internal structural temperature sensors) are installed inside and outside the main beams.

[0105] Finally, artificial corner reflectors were installed near the bridge site to serve as stable targets for synthetic aperture radar interferometric remote sensing monitoring. The corner reflectors should be oriented towards the direction of the radar satellite's passage to ensure that they form high-brightness point targets in the radar image.

[0106] Step S3: The data acquisition module collects the raw signal according to the preset triggering mechanism and transmits it to the monitoring center in real time through the communication network.

[0107] This step is responsible for converting the physical quantities sensed by the sensing layer into digital signals and reliably transmitting them to the data processing center. First, the sampling parameters of the data acquisition module are set. For dynamic response signals (vibration, dynamic strain, dynamic deflection), high-frequency sampling is used, and the sampling frequency is determined according to the structural modal analysis requirements, usually from 50Hz to 200Hz. For static response signals (temperature, hydrostatic level, static strain), low-frequency sampling is used, and the sampling frequency is usually from 0.01Hz to 0.1Hz. For dynamic weighing systems, event-triggered sampling is used. When a vehicle passes by, high-frequency sampling is automatically started to record the pressure waveform when the wheel passes by.

[0108] Secondly, configure the triggering mechanism of the data acquisition module. In addition to the conventional continuous and timed acquisition, a linkage triggering mechanism is set: when the dynamic weighing system detects a heavy vehicle (axle load exceeding the set threshold) passing by, it immediately triggers the vibration and strain sensors to perform encrypted acquisition (sampling frequency increased to 500Hz, duration 30 seconds) to obtain structural response details under vehicle load. When the GPS-BeiDou monitoring station detects a sudden displacement change (such as exceeding 5mm), it triggers the strain sensors near the cross-section to perform high-frequency acquisition.

[0109] Then, the data acquisition module filters, amplifies, and converts the acquired analog signals to digital data to generate digital monitoring data. Each data packet contains the sensor ID, timestamp (synchronized with the central server via NTP protocol), channel number, and original sampled value.

[0110] Finally, the data acquisition module uploads data packets in real time through the communication network. For data with high real-time requirements (vibration, strain), UDP protocol multicast is used to ensure low latency; for data with high reliability requirements (equipment status, alarm information), TCP protocol point-to-point transmission is used to ensure no data loss. The communication network adopts a fiber optic redundant ring network. When a link is interrupted, the network automatically switches to the backup path to ensure the continuity of data transmission. For remote measurement points, the 4G / 5G wireless backup link automatically takes over data transmission when the wired network fails.

[0111] Step S4: Perform online quality assessment, purification, and reconstruction on the received massive monitoring signals.

[0112] This step aims to extract high-quality, effective signals from massive amounts of raw data, providing a reliable data foundation for subsequent parameter identification. It specifically includes the following sub-steps:

[0113] S41. Perform online quality assessment on the received signal, including checking the data packet loss rate, determining whether the signal exceeds the reasonable range of the sensor, and identifying whether the signal-to-noise ratio meets the standard.

[0114] The data quality assessment module in the data processing and evaluation layer receives all raw monitoring data streams from the field in real time and performs online quality assessment on each frame of data. The assessment includes:

[0115] Packet loss rate check: Determine whether there is packet loss based on the continuity of data packet sequence numbers. If the continuous packet loss exceeds the set threshold of 10 packets, the data quality for that period is deemed unqualified and marked as "communication abnormal".

[0116] Range determination: Check whether the signal amplitude exceeds the sensor's reasonable range. If it continuously exceeds the upper limit of the range or falls below the lower limit, the sensor is considered to be faulty and marked as "over-range abnormality".

[0117] Signal-to-noise ratio (SNR) analysis: Calculate the signal-to-noise ratio (SNR) of the signal. For vibration signals, the effective frequency band energy can be extracted by bandpass filtering and compared with the out-of-band noise energy. If the SNR is lower than the set threshold of 20dB, the signal is determined to be severely polluted by noise and marked as "low SNR".

[0118] For data that meets the quality standards, it is labeled "valid" and directly output to step S43. For data that does not meet the quality standards, it is labeled "abnormal" and forwarded to step S42.

[0119] S42. Classify the marked abnormal data into sensor fault data or sudden interference event data.

[0120] The abnormal data classification and processing unit receives data tagged with "abnormal" and performs intelligent classification based on preset rules:

[0121] Sensor fault data: Data with characteristics such as continuous constant value (no change in multiple consecutive frames of data), large jump (change in adjacent frames exceeds the physical possibility range), periodic abnormal pulses, etc., are judged as "bad data" caused by sensor fault or acquisition channel fault. Such data is directly rejected. At the same time, the system automatically generates equipment maintenance work order and sends it to the visualization application layer to notify maintenance personnel to check the sensor or acquisition module.

[0122] Sudden interference event data: Data whose amplitude briefly exceeds the limit but whose waveform shape conforms to physical laws (seismic waves, impact response) is judged as "special event data" caused by sudden interference events. Although such data is abnormal, it contains important structural response information. Therefore, it is not removed as bad data, but is tagged as "special event" and stored separately in a dedicated partition of the database server as an event archive.

[0123] S43. Using exploratory data analysis methods, the statistical differences between effective signals and interference noise are used to screen out effective test signals that contain information about structural health status.

[0124] The valid signal filtering unit receives data labeled "valid" and "special event data" that has been classified and retained. It then uses exploratory data analysis methods to further filter out valid test signals that truly contain information about the structural health status. Specific methods include:

[0125] Statistical characteristic analysis: Calculate the mean, variance, skewness, kurtosis and other statistical measures of the data, and compare them with the historical statistical distribution. If the statistical measures deviate from the historical distribution by more than 3 times the standard deviation, it may be a new structural response feature and should be retained; if the deviation is caused by sudden noise, it should be removed.

[0126] Correlation analysis: For multiple sensors of the same type at the same cross section, calculate the correlation coefficient between them. If the correlation coefficient of a certain sensor with other sensors decreases significantly, it may be that the performance of the sensor has degraded and further analysis is required.

[0127] Time-frequency feature analysis: Perform short-time Fourier transform or wavelet transform on the signal to extract time-frequency features. If the time-frequency features of the signal do not match the modal features of the structural theory, it may be noise interference, which should be removed.

[0128] S44. Perform empirical mode decomposition on the selected effective signal to decompose the signal into intrinsic mode functions of different frequencies. After removing noise components, reconstruct the signal to obtain the purified signal.

[0129] The adaptive signal decomposition, purification, and reconstruction unit performs in-depth purification processing on the filtered effective signals. This unit incorporates an empirical mode decomposition algorithm, and the processing flow is as follows:

[0130] Adaptive decomposition: This method adaptively decomposes a non-stationary signal into several eigenmode functions. The decomposition process is based on the local characteristic time scale of the signal, filtering layer by layer from high frequency to low frequency, ultimately obtaining eigenmode function components of different frequency components from high to low and a residual term.

[0131] Noise component identification: Analyze the power spectral density and autocorrelation function of each intrinsic mode function component. Generally, the intrinsic mode function components representing random noise have the characteristics of wide bandwidth and low autocorrelation; the intrinsic mode function components representing structural response have the characteristics of narrow bandwidth and high autocorrelation. Identify the noise-dominated intrinsic mode function components and remove them.

[0132] Trend separation: Identify ultra-low frequency intrinsic mode components (slow deformation caused by temperature) that represent long-term trends, and decide whether to retain them based on analysis requirements;

[0133] Signal reconstruction: The remaining intrinsic mode function components are linearly superimposed to reconstruct the purified signal. The reconstructed signal retains the structural response characteristics of the original signal while eliminating noise interference, providing high-quality data input for subsequent parameter identification.

[0134] The purified signal is output to the signal processing and calculation module in step S5 on the one hand, and stored in the "purified data" partition of the database server on the other hand.

[0135] Step S5: Perform high-precision identification of structural state parameters based on the purified signal.

[0136] This step utilizes the purified, high-quality data to identify the structural state parameters of the bridge using advanced algorithms. This step specifically includes the following sub-steps:

[0137] Step S51: Perform spectral analysis on the purified cable acceleration signal to identify the cable's natural frequency. Combine this with the corrected cable length, linear density, and boundary condition parameters from the finite element model to calculate the current cable force value. The cable force identification unit receives the purified cable acceleration signal and uses the frequency method to identify the cable force.

[0138] Spectrum analysis: A fast Fourier transform is performed on the purified acceleration time-history signal to obtain the power spectral density function. A peak-picking algorithm is then used to identify the natural frequencies of the cable. To improve frequency identification accuracy, peak averaging or polynomial fitting methods can be used to refine the spectrum.

[0139] Order identification: Based on the vibration characteristics of the cable, the order of each frequency is identified. Usually, the fundamental frequency of the cable can be determined by the difference between adjacent order frequencies.

[0140] Cable force calculation: Substitute the identified natural frequencies into the cable force calculation formula. For short cables considering bending stiffness, use the accurate formula that takes boundary conditions into account; for long cables, a simplified string vibration formula can be used.

[0141]

[0142] Where T is the cable force. Let L be the linear density and L be the cable length. For the nth order frequency, during the calculation, parameters such as cable length and linear density are adopted using the values ​​corrected by the finite element model in step S1 to improve the calculation accuracy;

[0143] Results verification: For cables equipped with fiber Bragg grating smart cables, the frequency method identification results are compared and verified with the strain calculated cable force directly measured by the fiber Bragg grating. If the difference exceeds the set threshold of 5%, the algorithm parameters are corrected.

[0144] Step S52: The purified global navigation satellite system data, connecting pipe data and long gauge length strain data are fused to obtain the deflection curve of the main beam, and the strain modes of the structure are identified based on the long gauge length strain data.

[0145] The deformation and modal recognition unit receives purification data from GPS-BeiDou monitoring stations, connecting pipe levels, and long gauge-length fiber optic grating sensors, and performs fusion processing and modal recognition.

[0146] Deflection curve fusion: Using the Kalman filter algorithm, high-frequency low-precision displacement data from GPS-BeiDou monitoring stations and low-frequency high-precision displacement data from the connecting pipe level are fused to obtain high-precision, high-frequency dynamic deflection curves at each measuring point of the main beam. The state equation of the Kalman filter is based on the structural dynamics model, and the observation equation is based on the measurement equations of the two sensors. The optimal estimate is obtained through iterative updates.

[0147] Strain mode identification: Based on the regional strain data collected by long gauge length strain sensors, the strain modes of the structure are identified using regional sensing dynamic analysis theory. Compared with traditional acceleration modes, strain modes are more sensitive to local damage and low-frequency deformation of the main beam. The identification steps include: performing a fast Fourier transform on the purified strain time history data to obtain the strain frequency response function; identifying the modal frequencies and damping ratios of the strain modes using frequency domain decomposition or random subspace method; and constructing strain mode shapes based on the relative relationships of strain amplitudes at each measuring point.

[0148] Result verification: Compare the identified modal frequencies with the calculation results of the finite element model in step S1. If the difference is large, it may be that the structure has been damaged or the boundary conditions have changed, triggering further analysis.

[0149] Step S53: Perform inversion analysis on the vehicle load data collected by the dynamic weighing system and the structural response data collected by the long gauge length sensor to identify the axle load distribution on each lane and establish a spatiotemporal distribution map of vehicles on the bridge deck. The load inversion unit receives the vehicle load data from the dynamic weighing system and the structural response data from the long gauge length sensor, and performs load inversion:

[0150] Axle load identification: The raw pressure waveform recorded by the dynamic weighing system is analyzed. Through peak detection and waveform matching algorithms, the number of axles, axle load, wheelbase and speed of each vehicle are identified. Combined with timestamps, a vehicle passage record for each vehicle is established.

[0151] Lane matching: Based on the time difference of the vehicle passing through the dynamic weighing sensors in each lane, the lane and lateral position of the vehicle are determined.

[0152] Construction of a spatiotemporal load distribution map: All identified vehicle information is spatiotemporally aggregated to establish a spatiotemporal distribution map of vehicle loads on the bridge deck. The map uses time as the horizontal axis and the longitudinal position of the bridge as the vertical axis, with color intensity representing load magnitude, intuitively displaying the spatiotemporal evolution of vehicle loads on the bridge;

[0153] Load input correction: Input the actual vehicle load obtained by inversion into the finite element model in step S1, compare the structural response calculated by the model with the actual monitored structural response to verify the accuracy of the model. At the same time, the actual load distribution data can be used for structural fatigue assessment and remaining life prediction.

[0154] Step S6: Acquire multi-view radar satellite images, perform refined identification and time-series analysis of synthetic aperture radar interferometry signals, and establish a long-term deformation time series of the bridge. This step utilizes spaceborne or ground-based synthetic aperture radar interferometry to obtain overall deformation information of the bridge, complementing ground-based point monitoring. This step specifically includes the following sub-steps:

[0155] Step S61: To address the issues of synthetic aperture radar signal overlay and low coherence that bridge structures are prone to, a structural semantic information model is introduced to refine the identification of stable point targets on the bridge.

[0156] To address the issues of synthetic aperture radar signal overlay and low coherence in bridge structures (especially steel trusses and cables), a structural semantic information model is introduced for refined identification, including:

[0157] Image selection: Acquire multiple radar satellite images (Sentinel-1, TerraSAR-X, COSMO-SkyMed, etc.) covering the bridge area, with a time span covering the entire monitoring period and no fewer than 20 images;

[0158] Preprocessing: Preprocessing operations such as image registration, interferogram generation, flattening removal, and terrain phase removal are performed on the images;

[0159] Structural semantic information introduction: By combining the bridge's BIM model and design drawings, a structural semantic information model of the bridge is established to clarify the spatial location and geometric characteristics of each component of the bridge (main beam, tower, cable, pier); in synthetic aperture radar interferometry processing, this semantic information is used as a priori constraints to guide the identification and screening of point targets;

[0160] Refined Point Target Identification: A joint analysis method of coherent and incoherent information is employed to identify stable point targets on the bridge. For flat areas such as the bridge deck, a high-coherence point target identification algorithm is used; for complex structures such as cables and railings, a permanent scatterer identification algorithm based on amplitude deviation is used. The identified point targets should cover the main components of the bridge, and their density should meet the requirements of deformation analysis.

[0161] Step S62: Perform time-series synthetic aperture radar interferometry analysis on the identified point targets to establish a long-term deformation time series of the bridge. This time series is used to correct the finite element model in step S1 and is cross-checked with the discrete point data monitored by the Global Positioning System in step S5. The time-series synthetic aperture radar interferometry analysis on the identified point targets to establish a long-term deformation time series of the bridge includes:

[0162] Phase unwrapping: The interferogram is unwrapped to recover the absolute phase value of each point target;

[0163] Atmospheric phase correction: Utilizing the low-frequency spatial characteristics of atmospheric phase, phase errors caused by atmospheric delay are removed through filtering or meteorological data-based methods.

[0164] Deformation time series inversion: Using permanent scatterer interferometry or small baseline set technology, multiple images are jointly inverted to obtain the deformation time series of each point target along the radar line of sight. The time series is based on the acquisition time of the first image and displays the cumulative change of deformation value period by period.

[0165] Result Conversion and Verification: Convert the line-of-sight deformation into vertical and horizontal deformation (requires orbital data from at least two different perspectives). Compare and verify the deformation results identified by synthetic aperture radar interferometry with the discrete point data monitored by GPS in step S52 after spatial interpolation. If the trends are consistent and the error is within the allowable range (e.g., ±3mm), the reliability of the two methods is verified. If there is a systematic deviation, analyze the cause and correct the processing parameters.

[0166] Step S7: Compare the identified current structural state parameters with the theoretical thresholds and historical statistical thresholds set in Step S1 to trigger multi-level early warnings. This step assesses and issues early warnings based on the identification results of Steps S5 and S6.

[0167] First, a three-level early warning rule engine is established, including a yellow warning: triggered when the monitored value exceeds the daily statistical range. The daily statistical range is dynamically updated based on historical monitoring data, and the historical mean and standard deviation are usually calculated using the sliding window method. The warning threshold is the mean ± k times the standard deviation (k is 2 or 3). In this embodiment, when the cable force of a certain cable deviates from the historical mean by more than 2 times the standard deviation, a yellow warning is triggered, indicating "abnormal fluctuation in cable force, attention is recommended";

[0168] Orange alert: Triggered when the monitored value exceeds the design specification limit. The design specification limit is derived from the theoretical threshold determined in step S1 based on the design specification and finite element analysis. In this embodiment, when the main beam deflection exceeds the design allowable value (L / 500), an orange alert is triggered, indicating "deflection exceeds the limit, and there is a safety hazard in the structure";

[0169] Red Alert: Triggered when the rate of change of the monitored value is abnormal, indicating that the structure may suffer rapid damage. The abnormal rate of change is determined by calculating the derivative or difference of the monitored value. In this embodiment, when the strain of a certain section continues to increase monotonically for a short period of time (1 hour) and the rate of increase exceeds the set threshold (10με / hour), a red alert is triggered, indicating that "the structure may suffer rapid damage and an emergency inspection is recommended."

[0170] Secondly, the coordinated response after the warning is triggered includes warning location: after the warning is triggered, the system automatically highlights and flashes the corresponding component location on the Building Information Modeling-Geographic Information System visualization platform, and pops up a warning information window to display the warning type, trigger time, current value, threshold, change trend graph, etc.

[0171] Information push: Warning information is sent to the mobile terminals of relevant managers and technicians in real time via SMS, APP push, email, etc.

[0172] Encrypted data acquisition: For the monitoring points that trigger the early warning and their adjacent areas, the system automatically adjusts the sampling strategy of the data acquisition module, changing regular sampling to encrypted sampling (increasing the sampling frequency by 5-10 times) to obtain more detailed response data and provide support for cause analysis;

[0173] Video linkage: If video surveillance is installed on site, the system will automatically call the cameras near the warning area to capture and record images of the warning components, recording the situation on site.

[0174] Step S8: The processing results from steps S4 to S7 are mapped in real time onto a lightweight 3D bridge model, enabling visualized management of monitoring data based on Building Information Modeling (BIM) and Geographic Information System (GIS). This step uses digital-model linkage technology to transform abstract data into intuitive graphical displays, improving the readability of monitoring data and the convenience of operation and maintenance management.

[0175] First, a lightweight 3D model of the bridge is created. Based on the bridge's design drawings and BIM model, a lightweight model suitable for loading on web and mobile devices is generated through lightweight model processing (mesh simplification, texture compression, and instantiated rendering). The model should include all the main components of the bridge (main beams, towers, cables, piers, supports, and expansion joints), and data binding interfaces should be reserved for each component and sensor.

[0176] Secondly, real-time mapping between monitoring data and the 3D model is achieved, including color mapping: continuous monitoring values ​​such as cable force, stress, and deflection are mapped to the colors of model components. In this embodiment, a color gradient range is set: green represents a safe state (monitored value < 0.5 times the threshold), yellow represents a state of concern (0.5 times the threshold ≤ monitored value < 0.8 times the threshold), orange represents a warning state (0.8 times the threshold ≤ monitored value < threshold), and red represents an over-limit state (monitored value ≥ threshold). The model components dynamically change color according to real-time data, intuitively displaying the safety status distribution of the entire bridge.

[0177] Deformation mapping: Deflection monitoring data is mapped to the geometric deformation of the main beam. Based on the deflection values ​​at each measuring point, the vertical displacement of the main beam at each location is calculated using an interpolation algorithm, and the alignment of the main beam in the 3D model is updated in real time, intuitively displaying the deformation of the bridge under load.

[0178] Chart overlay: Real-time data charts are overlaid at key locations in the model. In this embodiment, the cable force time history curve is displayed near the cable anchorage zone, the deflection spectrum is displayed at the mid-span of the main beam, and the wind speed rose is displayed at the top of the bridge tower.

[0179] Finally, interactive data query and maintenance management are implemented, including sensor query: when an operator clicks on any sensor icon on the 3D model, the system retrieves detailed information about the sensor from the "raw data", "cleaned data" and "structural parameters" areas of the database server, including: real-time readings, historical time-history curves, spectrum analysis diagrams, equipment parameters (model, range, accuracy), installation location photos, calibration records, historical maintenance work orders, etc., and displays them in a pop-up window next to the model.

[0180] Maintenance work order management: When an equipment maintenance work order is generated in step S42, or an alarm is triggered in step S7, the system marks the corresponding location on the 3D model and generates a maintenance work order. The work order includes a problem description, location information, suggested measures, and processing time limit, and can be assigned to maintenance personnel online. After completing the on-site handling, the maintenance personnel can upload the processing results and photos via mobile device to update the work order status, forming a closed-loop management system.

[0181] Automatic report generation: The system supports automatic generation of monitoring reports by day, week, month, and year, including structural status parameter statistics, early warning event summaries, equipment operating status, and maintenance work order completion status. Reports can be directly overlaid on the 3D model or exported as PDF or Excel format.

[0182] This specific embodiment is merely an explanation of the present invention and is not intended to limit the invention. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they are within the scope of the claims of the present invention.

Claims

1. A monitoring system for long-span multi-tower cable-stayed bridges, characterized in that: It includes a sensing layer, a data acquisition and transmission layer, a data processing and evaluation layer, and a visualization application layer; The sensing layer is deployed at key sections of the bridge to collect various raw monitoring signals of the bridge in real time during its operation. The data acquisition and transmission layer is connected to the sensing layer and includes multiple data acquisition modules and a communication network; the data acquisition modules are used to receive and digitize the raw monitoring signals acquired by the sensing layer, and the communication network is used to transmit the digitized monitoring data in real time. The data processing and evaluation layer is connected to the data acquisition and transmission layer, and includes a data quality evaluation module, a signal processing and calculation module, and a database server connected in sequence. The data quality evaluation module is used to perform online quality evaluation and purification and reconstruction on the received massive monitoring data, and transmit the purified data to the signal processing and calculation module. The signal processing and calculation module is used to identify the structural state parameters of the bridge based on the purified data. The database server is connected to both the data quality evaluation module and the signal processing and calculation module, and is used to store the original monitoring data, intermediate processing results, and the finally identified structural state parameters. The visualization application layer is connected to the data processing and evaluation layer, including a building information model-geographic information system visualization platform and a monitoring and operation and maintenance management platform. The visualization application layer is used to receive the structural state parameters output by the data processing and evaluation layer and map them onto the bridge's three-dimensional model to realize the visualization display of monitoring data and multi-level early warning management.

2. The monitoring system for a long-span multi-tower cable-stayed bridge according to claim 1, characterized in that: The sensing layer includes a cable force monitoring module, a dynamic deformation monitoring module, an intelligent fiber sensing module, an environmental and load monitoring module, and a synthetic aperture radar interferometric remote sensing monitoring module. The cable tension monitoring module is used to collect vibration or strain signals of the stay cables; The dynamic deformation monitoring module is used to collect displacement, deflection, and vibration signals of the bridge's main beam and towers. The intelligent fiber sensing module is used to collect continuous strain field information along the longitudinal direction of the main beam; The environment and load monitoring module is used to collect information on wind speed, temperature, and vehicle load on the bridge deck. The synthetic aperture radar interferometric remote sensing monitoring module is used to acquire the overall deformation information of the bridge based on a spaceborne or ground-based synthetic aperture radar interferometric system.

3. The monitoring system for a long-span multi-tower cable-stayed bridge according to claim 1, characterized in that: The data quality assessment module includes an online quality assessment unit, an abnormal data classification and processing unit, an effective signal filtering unit, and a signal adaptive decomposition, purification, and reconstruction unit. The online quality assessment unit is used to perform packet loss rate checks, range determination, and signal-to-noise ratio analysis on the received monitoring signals. The abnormal data classification and processing unit is connected to the online quality assessment unit and is used to distinguish the marked abnormal data into sensor fault data or sudden interference event data. The effective signal filtering unit is connected to the abnormal data classification and processing unit, and is used to filter test signals containing structural health status information based on the statistical difference between effective signals and interference noise. The adaptive signal decomposition and purification reconstruction unit is connected to the effective signal screening unit and is used to perform empirical mode decomposition on the screened signal, remove noise components, and then reconstruct the purified signal.

4. The monitoring system for a long-span multi-tower cable-stayed bridge according to claim 1, characterized in that: The signal processing and calculation module includes a cable force identification unit, a deformation and mode identification unit, a load inversion unit, and a synthetic aperture radar interferometric data processing unit. The cable force identification unit is used to input the purified cable acceleration signal into the cable force identification algorithm, and calculate the current cable force value by combining the boundary conditions corrected by the finite element model. The deformation and modal identification unit is used to obtain the deflection curve of the main beam based on data from the Global Navigation Satellite System and the connecting pipe, and to identify the strain modes of the structure based on data from long gauge length strain sensors. The load inversion unit is used to combine dynamic weighing system data and structural response data collected by long gauge length sensors to invert the axle load on each lane in real time and establish a spatiotemporal distribution map of vehicles on the bridge deck. The synthetic aperture radar interferometric data processing unit is used to perform time-series analysis on multiple radar images to establish a long-term deformation time series of the bridge.

5. The monitoring system for a long-span multi-tower cable-stayed bridge according to claim 1, characterized in that: The monitoring and operation management platform includes a multi-level early warning unit and a digital-analog linkage unit; The multi-level early warning unit is used to trigger yellow, orange, or red early warnings based on the comparison results of the monitored values ​​with theoretical thresholds and historical statistical thresholds, and automatically locate the corresponding component positions in the three-dimensional model; The digital-analog linkage unit is used to realize the real-time mapping of monitoring data and building information model components, and supports clicking on the sensor icon on the model to retrieve historical time history curves, spectrum analysis diagrams and equipment file information.

6. A monitoring method for long-span multi-tower cable-stayed bridges, applied to a monitoring system for long-span multi-tower cable-stayed bridges according to any one of claims 1-5, characterized in that: Includes the following steps: S1. Establish a benchmark finite element model for a long-span multi-tower cable-stayed bridge with both road and rail on the same level, and determine the initial theoretical thresholds for key control sections and monitoring indicators; S2. Deploy a hybrid sensor network at key sections of the bridge; S3. The data acquisition module collects raw signals according to a preset triggering mechanism and transmits them to the monitoring center in real time through the communication network. S4. Perform online quality assessment, purification, and reconstruction of the massive amounts of received monitoring signals; S5. High-precision identification of structural state parameters based on the purified signal; S6. Acquire multiple radar satellite images, perform refined identification and time series analysis of synthetic aperture radar interferometric signals, and establish a long-term deformation time series of the bridge. S7. Compare the identified current structural state parameters with the theoretical threshold and historical statistical threshold set in step S1 to trigger multi-level early warning; S8. The processing results from steps S4 to S7 are mapped onto the lightweight 3D model of the bridge in real time to realize the visualization management of monitoring data based on Building Information Modeling-Geographic Information System.

7. The monitoring method for a long-span multi-tower cable-stayed bridge according to claim 6, characterized in that: The hybrid sensor network deployed in step S2 includes: long gauge-length fiber optic grating sensors deployed along the longitudinal direction of the main beam, accelerometers installed on the stay cables, global navigation satellite system displacement monitoring stations deployed at the top of the tower and the middle of the main beam span, a connecting pipe level deployed along the longitudinal direction of the main beam, and a dynamic weighing system deployed on the bridge deck.

8. The monitoring method for a long-span multi-tower cable-stayed bridge according to claim 6, characterized in that: Step S4 specifically includes: S41. Perform online quality assessment on the received signal, including checking the data packet loss rate, determining whether the signal exceeds the reasonable range of the sensor, and identifying whether the signal-to-noise ratio meets the standard. S42. Classify the marked abnormal data into sensor fault data or sudden interference event data. S43. Using exploratory data analysis methods, the statistical differences between effective signals and interference noise are used to screen out effective test signals that contain information about structural health status. S44. Perform empirical mode decomposition on the selected effective signal to decompose the signal into intrinsic mode functions of different frequencies. After removing noise components, reconstruct the signal to obtain the purified signal.

9. A monitoring method for a long-span multi-tower cable-stayed bridge according to claim 6, characterized in that: The dynamic deformation and modal identification step S5 further includes: based on the data collected by the long gauge length strain sensor, using the dynamic analysis theory of regional sensing, identifying the strain modes of the structure, which are used to analyze the local damage and low-frequency deformation characteristics of the wide cross-section main beam.

10. A monitoring method for a long-span multi-tower cable-stayed bridge according to claim 6, characterized in that: Step S6 specifically includes: S61. To address the issues of synthetic aperture radar signal overlay and low coherence that bridge structures are prone to, a structural semantic information model is introduced to refine the identification of stable point targets on bridges. S62. Perform time-series synthetic aperture radar interferometry analysis on the identified point targets to establish a long-term deformation time series of the bridge, which is used to correct the finite element model in step S1 and cross-check with the discrete point data monitored by the global positioning system in step S5.

Images