Engineering structure dynamic deformation monitoring system based on multi-sensor fusion

Abstract

The invention relates to the technical field of engineering structure health monitoring, in particular to an engineering structure dynamic deformation monitoring system based on multi-sensor fusion, which comprises a multi-type sensor module, a data acquisition and preprocessing module, a data fusion processing module, a deformation analysis and early warning module and a communication storage module. The multi-type sensor module collects data such as displacement, strain and vibration of an engineering structure; the data acquisition and preprocessing module performs noise reduction on the original data; the data fusion processing module fuses data by adopting algorithms such as Kalman filtering and a D-S evidence theory; the deformation analysis and early warning module constructs a finite element model to analyze deformation and performs graded early warning; and the communication storage module realizes data transmission and storage. The system improves the monitoring precision and reliability through multi-sensor data fusion, and can be applied to dynamic deformation monitoring and safety early warning of engineering structures such as bridges and dams.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engineering structure health monitoring, in particular to an engineering structure dynamic deformation monitoring system based on multi-sensor fusion. BACKGROUND

[0002] With the rapid development of social economy, various large-scale engineering structures such as bridges, dams, high-rise buildings, etc. have emerged in large numbers and are put into use. These engineering structures, under the combined action of natural environmental factors (such as earthquakes, strong winds, temperature changes, rain erosion) and human factors (such as traffic loads, structural aging, changes in use functions), inevitably produce structural deformations. If these deformations cannot be monitored in a timely and accurate manner, it may lead to a decline in structural performance, and even cause serious safety accidents, resulting in huge economic losses and casualties. Therefore, it is of great importance to monitor the dynamic deformation of engineering structures and grasp their health status in order to ensure the safe operation of engineering structures.

[0003] Traditional engineering structure deformation monitoring methods mainly include manual measurement and single-sensor monitoring. Manual measurement such as leveling and total station measurement, although has high measurement accuracy, has the disadvantages of low efficiency, high labor intensity, long monitoring period, and difficulty in realizing real-time monitoring, etc., and cannot capture the dynamic change process of engineering structures in a timely manner. For example, for large bridges, manual measurement method for monitoring their deformation requires a large amount of manpower and time, and can only obtain data at discrete time points, making it difficult to reflect the real-time deformation of bridges under dynamic actions such as traffic loads.

[0004] Single-sensor monitoring, such as using only displacement sensors or strain sensors for monitoring, has limitations in monitoring results due to measurement errors of the sensors themselves and the fact that only one aspect of information of the structure can be obtained. When the sensor is disturbed by the environment (such as electromagnetic interference, temperature change) or fails, the reliability of the monitoring data will be greatly reduced, and the real deformation state of the engineering structure cannot be accurately reflected. For example, in a strong electromagnetic environment, some electronic displacement sensors may have abnormal data, resulting in invalid monitoring results.

[0005] In recent years, with the continuous development of sensor technology, communication technology and computer technology, multi-sensor based engineering structure monitoring systems have gradually been applied. However, most of the existing multi-sensor monitoring systems independently process and analyze the data collected by different types of sensors, do not fully utilize the complementarity between multi-source data, and cannot effectively improve the monitoring accuracy and reliability. At the same time, in terms of data processing, for the large amount of noisy data collected by sensors, there is a lack of efficient data denoising and fusion algorithms, resulting in large errors in data processing results. In addition, the existing monitoring system is not perfect in the early warning mechanism, and cannot make scientific and reasonable hierarchical early warning according to the actual deformation of the engineering structure, making it difficult to meet the actual needs of engineering structure safety monitoring. For example, in some dam monitoring projects, although multiple sensors are installed, due to improper data fusion and analysis methods, potential safety hazards of the dam cannot be found in time, resulting in the inability to take effective maintenance measures in time. Therefore, it is urgent to develop an engineering structure dynamic deformation monitoring system based on multi-sensor fusion to improve the monitoring accuracy, reliability and early warning ability. SUMMARY

[0006] The purpose of the present application is to provide an engineering structure dynamic deformation monitoring system based on multi-sensor fusion, which collects multi-dimensional data of engineering structures through multiple types of sensors, combines advanced data processing and fusion algorithms, realizes high-precision and high-reliability monitoring of engineering structure dynamic deformation, and establishes a scientific and reasonable hierarchical early warning mechanism to ensure the safe operation of engineering structures.

[0007] The technical solution adopted by the present application to solve its technical problems is: an engineering structure dynamic deformation monitoring system based on multi-sensor fusion, comprising: a multi-type sensor module for collecting displacement, strain and vibration data of the engineering structure; a data acquisition and preprocessing module connected to the multi-type sensor module for filtering and denoising the original data; a data fusion processing module connected to the data acquisition and preprocessing module for fusing the preprocessed data using Kalman filtering algorithm or D-S evidence theory; a deformation analysis and early warning module connected to the data fusion processing module for constructing a structure deformation finite element model based on the fused data and performing hierarchical early warning according to a set threshold; a communication and storage module connected to the deformation analysis and early warning module for remote transmission and storage of data.

[0008] Specifically, the multi-type sensor module includes a fiber grating sensor, a Beidou positioning module, a MEMS acceleration sensor, and an inclination sensor; the fiber grating sensor is used for monitoring structural strain, and the wavelength demodulation accuracy is ±0.1 pm; the Beidou positioning module is used for monitoring structural displacement, and the positioning accuracy is ±2 cm in the horizontal direction and ±3 cm in the elevation direction; the MEMS acceleration sensor is used for monitoring vibration acceleration, and the range is ±10 g; and the inclination sensor is used for monitoring the inclination angle, and the accuracy is ±0.01°.

[0009] Specifically, the data acquisition and preprocessing module adopts a wavelet threshold denoising algorithm to process the original data, a db4 wavelet base function is selected, the decomposition layer number is 3-5 layers, and a soft threshold function is used as the threshold function , wherein x is a noisy signal, is a threshold.

[0010] Specifically, when the data fusion processing module adopts a Kalman filtering algorithm, the state equation is , and the observation equation is , wherein is a state vector at time k, is a state transition matrix, is a process noise covariance matrix, is an observation vector, is an observation matrix, is an observation noise covariance matrix.

[0011] Specifically, the deformation analysis and early warning module sets three levels of early warning thresholds, when the fusion data exceeds a first level of early warning threshold, yellow early warning is performed; when the fusion data exceeds a second level of early warning threshold, orange early warning is performed; and when the fusion data exceeds a third level of early warning threshold, red early warning is performed. The early warning thresholds at each level are determined according to the engineering structure design standard and historical monitoring data.

[0012] Specifically, the communication and storage module adopts 5G and NB-IoT dual-mode communication, 5G is used for real-time data high-speed transmission, and NB-IoT is used for low-power data transmission; and the data storage adopts a time series database, and the storage period is 5-15 minutes.

[0013] Specifically, the power supply module adopts a combination of a solar panel and a lithium battery for power supply, the solar panel power is 120-180 W, and the lithium battery capacity is 60-100 Ah.

[0014] Specifically, when the deformation analysis and early warning module constructs a finite element model, a tetrahedral element is used to divide the grid of the engineering structure, and the element size is determined according to the key parts of the structure, and the element size of the key parts is 0.5-1 m.

[0015] Specifically, the data fusion processing module synchronizes the time of different types of sensor data before data fusion, and uses the NTP protocol to achieve time synchronization, with a synchronization accuracy of ±1ms.

[0016] Specifically, the multi-type sensor module further comprises a laser ranging sensor for monitoring structural crack width changes, with a measurement accuracy of ±0.02mm.

[0017] The beneficial effects of the present application are: Improve monitoring accuracy: By collecting multi-dimensional data of engineering structures through multiple types of sensors and processing them using advanced data fusion algorithms, the complementarity of each sensor data is fully utilized, effectively reducing the measurement error of a single sensor. Compared with traditional single sensor monitoring, the monitoring accuracy is improved by 30%-50%. For example, after fusing the displacement data of the Beidou positioning module with the strain data of the fiber Bragg grating sensor, the actual deformation of the structure can be more accurately reflected.

[0018] Enhance reliability: The application of multi-sensor redundancy design and data fusion technology makes the system able to reduce the influence of faulty sensors on the overall monitoring results through other sensor data and fusion algorithms when a sensor fails or is interfered, ensuring the normal operation of the system, and improving the system reliability by 40%-60%.

[0019] Implement scientific early warning: The established three-level early warning mechanism sets early warning thresholds based on engineering structure design standards and historical monitoring data, and can timely and accurately issue early warning information of different levels according to the deformation degree of the structure, providing sufficient response time for the safety maintenance of engineering structures and effectively avoiding safety accidents.

[0020] Reduce operation and maintenance costs: The power supply module uses a combination of solar power and lithium batteries to reduce dependence on traditional power and reduce long-term operating costs. At the same time, dual-mode communication and flexible data storage methods optimize data transmission and storage strategies, further reducing the operation and maintenance costs of the system.

[0021] Strong adaptability: The system uses modular design, which can flexibly configure sensor types and quantities according to the characteristics and monitoring needs of different engineering structures, and is suitable for dynamic deformation monitoring of various engineering structures such as bridges, dams, and high-rise buildings, with wide application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0022] The present application will be further described below in conjunction with the drawings and examples.

[0023] Figure 1 The architecture diagram of the engineering structure dynamic deformation monitoring system based on multi-sensor fusion provided by the present application. DETAILED DESCRIPTION

[0024] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0025] like Figure 1 As shown, the dynamic deformation monitoring system for engineering structures based on multi-sensor fusion of the present invention includes a multi-type sensor module, a data acquisition and preprocessing module, a data fusion processing module, a deformation analysis and early warning module, and a communication and storage module.

[0026] Multiple sensor modules are responsible for collecting multi-dimensional data such as displacement, strain, and vibration of the engineering structure. Specifically, these include: a fiber optic grating sensor, which utilizes the wavelength-strain sensitivity of fiber optic gratings to convert structural strain changes into optical wavelength changes, achieving high-precision strain measurement with a wavelength demodulation accuracy of ±0.1 pm; a BeiDou positioning module, which receives BeiDou satellite signals to achieve real-time monitoring of the three-dimensional displacement of the engineering structure, with positioning accuracy of ±2 cm in the horizontal direction and ±3 cm in the vertical direction; a MEMS accelerometer, which can monitor the acceleration response of the structure under vibration loads in real time, with a measurement range of ±10g; and a tilt sensor, which measures the tilt angle of the structure with an accuracy of ±0.01°. In addition, a laser rangefinder sensor can be configured according to actual needs to monitor changes in the width of structural cracks, with a measurement accuracy of ±0.02 mm.

[0027] The data acquisition and preprocessing module is connected to multiple sensor modules and uses a wavelet thresholding denoising algorithm to process the raw data. db4 is selected as the wavelet basis function, and the number of decomposition levels is set between 3 and 5 levels based on the data characteristics, utilizing a soft thresholding function. Noise reduction processing is performed on noisy signals to effectively remove noise from sensor-collected data and improve data quality.

[0028] The data fusion processing module is connected to the data acquisition and preprocessing module, and can use the Kalman filter algorithm or DS evidence theory to fuse the preprocessed data. When using the Kalman filter algorithm, the state equation is established based on the dynamic characteristics of the engineering structure. and observation equations Through continuous prediction and updating processes, optimal fusion of multi-source data is achieved, thereby improving the accuracy and reliability of the data.

[0029] The deformation analysis and early warning module is connected to the data fusion processing module, constructing a finite element model of structural deformation based on the fused data. Tetrahedral elements are used to mesh the engineering structure, with element sizes controlled between 0.5-1m for key components. Combined with structural material mechanical parameters and boundary conditions, accurate simulation of structural deformation is achieved. Simultaneously, a three-level early warning threshold is set. Based on engineering structural design standards and historical monitoring data, a yellow warning is issued when the fused data exceeds the first-level threshold; an orange warning is issued when it exceeds the second-level threshold; and a red warning is issued when it exceeds the third-level threshold, providing timely evidence for the safety assessment of the engineering structure.

[0030] The communication and storage module connects to the deformation analysis and early warning module, employing dual-mode communication of 5G and NB-IoT. 5G is used for high-speed transmission of real-time data, meeting the real-time upload requirements of large amounts of monitoring data; NB-IoT is used for low-power data transmission, suitable for scenarios with smaller data volumes and lower real-time requirements. Data storage uses a time-series database, with storage periods flexibly set between 5 and 15 minutes, facilitating the management and analysis of monitoring data.

[0031] In addition, the system also includes a power supply module, which uses a combination of solar panels and lithium batteries. The solar panels have a power of 120-180W and the lithium batteries have a capacity of 60-100Ah, which can provide a stable power supply for the system under various environmental conditions and ensure the continuous operation of the system.

[0032] Architecture Diagram Module Description 1. Power supply module It uses a combination of solar panels (120-180W) and lithium batteries (60-100Ah) for power supply, making it suitable for outdoor engineering environments and ensuring continuous system operation.

[0033] 2. Data Acquisition Layer Multiple types of sensors: Fiber Bragg grating sensor: High-precision strain monitoring (±0.1pm); Beidou positioning module: 3D displacement monitoring (horizontal ±2cm, elevation ±3cm); MEMS accelerometer: Vibration acceleration monitoring (range ±10g); Tilt sensor: tilt angle monitoring (accuracy ±0.01°); Laser rangefinder sensor: crack width monitoring (accuracy ±0.02mm).

[0034] 3. Data Processing Layer Preprocessing: Wavelet thresholding denoising algorithm (db4 basis function, 3-5 level decomposition, soft thresholding function) removes noise from the original data; Data fusion: Time synchronization (accuracy ±1ms) is achieved by using Kalman filtering or DS evidence theory combined with NTP protocol. Deformation analysis and early warning: Finite element model: Tetrahedral element mesh generation, with element size of 0.5-1m for key parts; A three-tiered early warning mechanism is in place: based on the comparison of fused data with thresholds, yellow, orange, and red warnings are triggered.

[0035] 4. Data Management Layer Communication module: Dual-mode 5G (real-time high-speed transmission) and NB-IoT (low-power transmission); Storage module: Time-series database, storage period of 5-15 minutes, supports management of massive monitoring data.

[0036] 5. Application Layer The monitoring center receives and visualizes the data, triggers response mechanisms based on the warning level, and supports decision-making for the safety maintenance of engineering structures.

[0037] Example 1: Dynamic Deformation Monitoring of Large Bridges A large cable-stayed bridge spanning a river was selected as the monitoring object. Thirty fiber optic strain sensors were installed every 8 meters on the main girder to monitor strain changes under vehicle loads and wind forces. Four BeiDou positioning modules were installed at the top of the bridge towers and at both ends of the main girder to monitor the overall displacement of the bridge. Four MEMS accelerometers were installed at the mid-span and quarter-span points of the main girder to monitor vibration acceleration. Four tilt sensors were installed at different heights of the bridge towers to monitor the tilt angle of the towers. Six laser rangefinders were installed at key crack locations on the main girder to monitor crack width changes.

[0038] The data acquisition and preprocessing module employs a wavelet threshold denoising algorithm to process the raw data. The db4 wavelet basis function is selected, and the decomposition level is set to three. A threshold is determined based on the characteristics of the noisy data to denoise the data acquired by each sensor. For example, after denoising the strain data acquired by the fiber Bragg grating sensor, the signal-to-noise ratio is improved by 15 dB.

[0039] The data fusion processing module employs the Kalman filter algorithm for data fusion. State equations and observation equations are established based on the bridge's dynamic characteristics, including the state transition matrix. The observation matrix is ​​determined based on the bridge's mass and stiffness matrices. The measurement principles of each sensor are determined. A Kalman filter algorithm is used to fuse the preprocessed strain, displacement, acceleration, tilt angle, and crack width data to obtain more accurate state information of the bridge structure.

[0040] The deformation analysis and early warning module constructs a finite element model of the bridge based on the fused data. Tetrahedral elements are used to mesh the bridge, with key components having an element size of 0.8m. This is combined with the bridge's material parameters (elastic modulus). The system uses a framework with Poisson's ratio of 0.3 and boundary conditions (fixed pier constraints). The first-level warning thresholds are set as follows: strain change exceeding 60 με, displacement change exceeding 15 mm, acceleration exceeding 0.6 m / s², tilt angle change exceeding 0.6°, and crack width change exceeding 0.2 mm. The second-level warning thresholds are: strain change exceeding 120 με, displacement change exceeding 30 mm, acceleration exceeding 1.2 m / s², tilt angle change exceeding 1.2°, and crack width change exceeding 0.5 mm. The third-level warning thresholds are: strain change exceeding 200 με, displacement change exceeding 50 mm, acceleration exceeding 2 m / s², tilt angle change exceeding 2°, and crack width change exceeding 1 mm. A yellow warning is issued when the strain of the main beam reaches 70 με at a given moment.

[0041] The communication and storage module employs dual-mode communication of 5G and NB-IoT. 5G is used to transmit monitoring data and early warning information to the monitoring center in real time, while NB-IoT is used for data transmission in a low-power mode. Data storage utilizes a time-series database with a storage period of 10 minutes, and is stored on a 2TB hard drive.

[0042] The power supply module uses a 150W solar panel and an 80Ah lithium battery. The solar panel is installed on the side of the bridge tower, and the lithium battery is placed in the equipment box inside the bridge tower to ensure stable power supply to the system.

[0043] Example 2: Dynamic Deformation Monitoring of Hydraulic Dams Monitoring was conducted on a large concrete gravity dam. Forty fiber optic strain sensors were installed every 12 meters along the dam's surface; three BeiDou positioning modules were installed at the top and bottom of the dam; eight MEMS accelerometers were installed at different elevations along the dam; six tilt sensors were installed at the dam abutments and key sections of the dam body; and ten laser rangefinders were installed at crack monitoring points along the dam.

[0044] The wavelet threshold denoising algorithm of the data acquisition and preprocessing module uses the db4 wavelet basis function and sets the number of decomposition layers to 4 to denoise the original data, effectively removing environmental noise and sensor noise.

[0045] The data fusion processing module employs the DS evidence theory for data fusion. First, the basic probability allocation function for each sensor's data is determined. Then, multi-source data is fused according to the DS synthesis rules. For example, for displacement and strain data, effective data fusion is achieved by calculating their confidence levels under different evidence combinations.

[0046] The deformation analysis and early warning module constructs a finite element model of the dam, using tetrahedral element meshing, with key components having an element size of 0.6m. This is combined with the dam's material parameters (elastic modulus, etc.). The system uses a boundary condition (Poisson's ratio 0.2) and fixed constraints on the dam foundation. The first-level warning threshold is set as follows: strain change exceeding 80 με, displacement change exceeding 20 mm, acceleration exceeding 0.8 m / s², tilt angle change exceeding 0.8°, and crack width change exceeding 0.3 mm. The second-level warning threshold is set as follows: strain change exceeding 150 με, displacement change exceeding 40 mm, acceleration exceeding 1.5 m / s², tilt angle change exceeding 1.5°, and crack width change exceeding 0.8 mm. The third-level warning threshold is set as follows: strain change exceeding 300 με, displacement change exceeding 60 mm, acceleration exceeding 2.5 m / s², tilt angle change exceeding 2.5°, and crack width change exceeding 1.5 mm. When the displacement of a certain part of the dam reaches 35 mm, the system issues an orange warning.

[0047] The communication and storage module uses the 5G network for real-time data transmission, NB-IoT is used for low-power data transmission, the storage cycle is set to 12 minutes, and a 3TB hard drive is used to store data.

[0048] The power supply module uses a 180W solar panel and a 100Ah lithium battery. The solar panel is installed in an open area on top of the dam, and the lithium battery is installed in the monitoring equipment room of the dam to ensure continuous power supply to the system.

[0049] Example 3: Dynamic Deformation Monitoring of High-Rise Buildings A super high-rise office building with 60 floors and a height of 280 meters was selected as the monitoring object. Four fiber optic strain sensors were installed every five floors in the core wall of the building, for a total of 48 sensors, to monitor strain changes in the core under wind load, structural weight, and personnel movement loads. Two Beidou positioning modules were installed on the roof and on the 30th-floor refuge floor to monitor displacement at different heights. Two MEMS accelerometers were installed at the junction of each corridor and stairwell, for a total of 120 sensors, to capture the building's vibration acceleration in real time. One tilt sensor was installed at the top of each of the four corner frame columns to monitor the building's tilt angle. Additionally, eight laser rangefinders were installed at locations on the building's exterior facade where cracks might appear to monitor changes in crack width.

[0050] The data acquisition and preprocessing module employs a wavelet threshold denoising algorithm to process the raw data, using the db4 wavelet basis function and setting the decomposition level to 5. Threshold parameters are finely adjusted to suit the characteristics of different types of sensor data. For example, for high-frequency vibration data acquired by a MEMS accelerometer, optimizing the threshold effectively removes environmental vibration interference noise, achieving a signal effective component retention rate of 92%.

[0051] The data fusion processing module employs the Kalman filter algorithm for data fusion. Based on the dynamic characteristics of high-rise buildings, state equations and observation equations are established. The state transition matrix... The determination is made by comprehensively considering the influence of dynamic forces such as the building's mass matrix, stiffness matrix, and wind loads; observation matrix. Precise calculations are performed based on the measurement principles and installation locations of each sensor. The pre-processed strain, displacement, acceleration, tilt angle, and crack width data are fused using a Kalman filter algorithm, significantly improving the accuracy and stability of the data.

[0052] The deformation analysis and early warning module constructs a finite element model of the high-rise building based on the fused data. Tetrahedral elements are used to mesh the building structure, with the element size set to 0.5m in key areas such as the core tube and frame columns. This is combined with the building's material parameters (elastic modulus of concrete). ), elastic modulus of steel The system sets the following warning thresholds: Level 1: strain change exceeding 40 με, displacement change exceeding 8 mm, acceleration exceeding 0.4 m / s², tilt angle change exceeding 0.4°, and crack width change exceeding 0.1 mm; Level 2: strain change exceeding 80 με, displacement change exceeding 15 mm, acceleration exceeding 0.8 m / s², tilt angle change exceeding 0.8°, and crack width change exceeding 0.3 mm; Level 3: strain change exceeding 150 με, displacement change exceeding 25 mm, acceleration exceeding 1.5 m / s², tilt angle change exceeding 1.5°, and crack width change exceeding 0.5 mm. A yellow warning is issued when the strain of a core tube reaches 50 με.

[0053] The communication and storage module employs dual-mode communication of 5G and NB-IoT. The 5G network is responsible for rapidly transmitting large amounts of real-time monitoring data to the monitoring center, while NB-IoT is used to reduce power consumption in low-data-volume transmission scenarios (such as device status information reporting). Data storage utilizes a time-series database with a storage period of 5 minutes, employing a 4TB solid-state drive to meet the storage requirements of massive monitoring data for ultra-high-rise buildings.

[0054] The power supply module uses a 120W solar panel and a 60Ah lithium battery. The solar panel is installed in an unused area on the building's roof, while the lithium battery is placed in the building's equipment room. An intelligent charging and discharging management system ensures that the solar panel charges the lithium battery when there is sufficient sunlight, prioritizing solar power for the system. During periods of insufficient sunlight or at night, the lithium battery provides a stable power supply, guaranteeing uninterrupted 24-hour operation.

[0055] Example 4: Dynamic Deformation Monitoring of Large Sports Venues Taking a large-scale comprehensive sports stadium as an example, the stadium adopts a large-span spatial steel structure with a roof span of 180 meters. At key nodes of the steel truss, one fiber optic strain sensor is installed every 6 meters, for a total of 50 sensors, to monitor the strain of the steel truss under conditions of personnel gathering, equipment loads, and temperature changes. Five Beidou positioning modules are installed at the four corners and center of the roof to monitor roof displacement changes. One MEMS accelerometer is installed every 10 meters on the lower chord of the steel truss, for a total of 30 sensors, to monitor the vibration acceleration of the structure. One tilt sensor is installed at the top of each column, for a total of 20 sensors, to monitor the tilt angle of the columns. Simultaneously, 12 laser rangefinders are installed at the welds of the steel components to monitor changes in crack width at the welds.

[0056] The data acquisition and preprocessing module employs a wavelet thresholding denoising algorithm, selecting the db4 wavelet basis function and setting the decomposition level to 4 layers. For fiber optic strain sensor data that is significantly affected by temperature, the threshold is adaptively adjusted to effectively eliminate noise interference caused by temperature drift, reducing the measurement error of the strain data to ±3με.

[0057] The data fusion processing module employs DS evidence theory for data fusion. Based on the characteristics and historical data of each sensor, a basic probability allocation function is determined. For example, for displacement and acceleration data, by analyzing their correlation under different operating conditions, evidence weights are rationally allocated, and then the multi-source data is fused according to DS synthesis rules, thus improving the reliability of data fusion.

[0058] The deformation analysis and early warning module constructs a finite element model of the stadium based on the fused data. Tetrahedral elements are used to mesh the steel structure, with the element size set to 0.8m at key locations such as steel truss nodes. This is combined with the material parameters of the steel structure (elastic modulus). The system uses Poisson's ratio (0.3) and boundary conditions (fixed constraint at the bottom of the column) to set the following warning thresholds: Level 1: strain change exceeding 50 με, displacement change exceeding 10 mm, acceleration exceeding 0.5 m / s², tilt angle change exceeding 0.5°, and crack width change exceeding 0.15 mm; Level 2: strain change exceeding 100 με, displacement change exceeding 20 mm, acceleration exceeding 1 m / s², tilt angle change exceeding 1°, and crack width change exceeding 0.4 mm; Level 3: strain change exceeding 200 με, displacement change exceeding 30 mm, acceleration exceeding 2 m / s², tilt angle change exceeding 2°, and crack width change exceeding 0.8 mm. When a displacement of 18 mm is detected at a certain part of the roof, the system issues an orange warning.

[0059] The communication and storage module employs dual-mode communication of 5G and NB-IoT. 5G enables real-time, high-speed transmission of monitoring data, while NB-IoT is used for low-power communication of the device. Data storage utilizes a time-series database with a storage period of 8 minutes, and is stored on a 3TB hard drive.

[0060] The power supply module is equipped with a 160W solar panel and an 80Ah lithium battery. The solar panel is installed on the roof of the stadium, while the lithium battery is placed in the equipment room of the stadium to provide a stable power supply for the monitoring system and ensure the normal operation of the system during various events and activities.

[0061] Example 5: Dynamic Deformation Monitoring of Railway Bridges A simply supported beam bridge on a high-speed railway was selected as the monitoring object, with a span of 32 meters. One fiber optic strain sensor was installed every 4 meters on the side of each beam, eight sensors per beam, for a total of 16 sensors across the bridge, to monitor strain changes in the beam under dynamic train loads. Three BeiDou positioning modules were installed at each end of the beam and at mid-span to monitor beam displacement. Four MEMS accelerometers were installed near the mid-span and supports to monitor vibration acceleration. Four tilt sensors were installed on the top of the piers to monitor their tilt angles. Six laser rangefinders were installed at the diaphragm connections to monitor changes in diaphragm crack width.

[0062] The data acquisition and preprocessing module employs a wavelet thresholding denoising algorithm, using the db4 wavelet basis function and setting the decomposition level to 3. To address the strong interference signals generated by passing trains, the threshold function is optimized, effectively removing the impact of train wheel-rail noise on sensor data and improving the signal-to-noise ratio.

[0063] The data fusion processing module employs the Kalman filter algorithm for data fusion. Based on the dynamic characteristics of railway bridges under train dynamic loads, accurate state equations and observation equations are established. (State transition matrix) Considering the impact of factors such as train speed and axle load on bridge structure, the observation matrix... Based on the sensor's installation location and measurement accuracy, the optimal fusion of multi-source data is achieved through the Kalman filter algorithm.

[0064] The deformation analysis and early warning module constructs a finite element model of the railway bridge based on the fused data. Tetrahedral elements are used to mesh the bridge, with key components having an element size of 0.6m. This is combined with the bridge's material parameters (elastic modulus of concrete). The system uses Poisson's ratio of 0.2 and boundary conditions (fixed pier constraints, simply supported beam constraints) to set the following warning thresholds: Level 1: strain change exceeding 70 με, displacement change exceeding 12 mm, acceleration exceeding 0.7 m / s², tilt angle change exceeding 0.7°, and crack width change exceeding 0.2 mm; Level 2: strain change exceeding 140 με, displacement change exceeding 25 mm, acceleration exceeding 1.4 m / s², tilt angle change exceeding 1.4°, and crack width change exceeding 0.5 mm; Level 3: strain change exceeding 250 με, displacement change exceeding 40 mm, acceleration exceeding 2.5 m / s², tilt angle change exceeding 2.5°, and crack width change exceeding 1 mm. A yellow warning is issued when the strain of a beam reaches 80 με.

[0065] The communication and storage module employs dual-mode communication of 5G and NB-IoT. 5G is used for real-time transmission of large amounts of monitoring data when trains pass by, while NB-IoT is used for low-volume device status monitoring data transmission during normal operation. Data storage utilizes a time-series database with a storage period of 12 minutes, and is stored on a 2TB hard drive.

[0066] The power supply module uses a 140W solar panel and a 70Ah lithium battery. The solar panel is installed on the side of the bridge pier, and the lithium battery is placed in the equipment box on the bridge pier to provide stable power to the monitoring system and ensure the safety monitoring of the railway bridge.

[0067] Example 6: Dynamic Deformation Monitoring of Urban Subway Tunnels Monitoring was conducted on a subway tunnel section in a certain city. One fiber optic strain sensor was installed every 5 meters at the tunnel's arch crown, arch waist, and arch bottom, with three sensors installed per section, totaling 60 sections and 180 sensors. These sensors were used to monitor strain changes in the tunnel lining under soil pressure, train vibration, and other effects. Eight BeiDou positioning modules were installed every 200 meters at the tunnel entrances, exits, and the middle section to monitor the overall displacement of the tunnel. Twenty-four MEMS accelerometers were installed every 50 meters inside the tunnel to monitor vibration acceleration. Four tilt sensors were installed at the tunnel's connecting passages to monitor the tilt angle of the connecting passages. Ten laser rangefinders were installed at locations prone to cracking in the tunnel lining to monitor crack width changes.

[0068] The data acquisition and preprocessing module employs a wavelet threshold denoising algorithm, using the db4 wavelet basis function and setting the decomposition level to 4. To address the complex electromagnetic interference within the tunnel, deep denoising processing is performed on the sensor data, effectively improving data quality.

[0069] The data fusion processing module employs DS evidence theory for data fusion. Based on the characteristics of tunnel monitoring data, it rationally determines the basic probability allocation function for each sensor's data and fuses multi-source data through DS synthesis rules, thereby enhancing the reliability and accuracy of the data.

[0070] The deformation analysis and early warning module constructs a finite element model of the subway tunnel based on the fused data. Tetrahedral elements are used to mesh the tunnel, with key components having an element size of 0.5m. This is combined with the material parameters of the tunnel lining (elastic modulus of concrete). The system uses Poisson's ratio (0.2) and boundary conditions (soil constraints) to set the following warning thresholds: Level 1: strain change exceeding 60 με, displacement change exceeding 10 mm, acceleration exceeding 0.6 m / s², dip angle change exceeding 0.6°, and crack width change exceeding 0.15 mm; Level 2: strain change exceeding 120 με, displacement change exceeding 20 mm, acceleration exceeding 1.2 m / s², dip angle change exceeding 1.2°, and crack width change exceeding 0.4 mm; Level 3: strain change exceeding 200 με, displacement change exceeding 30 mm, acceleration exceeding 2 m / s², dip angle change exceeding 2°, and crack width change exceeding 0.8 mm. A yellow warning is issued when the strain at the crown of a section reaches 70 με.

[0071] The communication and storage module employs dual-mode communication of 5G and NB-IoT. 5G is used for real-time transmission of large amounts of monitoring data, while NB-IoT is used for low-power data transmission. Data storage utilizes a time-series database with a storage period of 15 minutes, and is stored on a 3TB hard drive.

[0072] The power supply module uses a 180W solar panel and a 100Ah lithium battery. The solar panel is installed at a suitable location near the tunnel entrance and exit, and the lithium battery is placed in the equipment box inside the tunnel. Through reasonable power supply design, the monitoring system can be made to operate stably in the complex environment of the subway tunnel.

[0073] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic deformation monitoring system for engineering structures based on multi-sensor fusion, characterized in that, include: Multiple types of sensor modules are used to collect displacement, strain, and vibration data of engineering structures; The data acquisition and preprocessing module is connected to the multi-type sensor module and is used to filter and reduce noise in the raw data; The data fusion processing module is connected to the data acquisition and preprocessing module and is used to fuse the preprocessed data using the Kalman filter algorithm or DS evidence theory. The deformation analysis and early warning module is connected to the data fusion processing module and is used to construct a structural deformation finite element model based on the fused data and to provide graded early warnings according to a set threshold. The communication and storage module is connected to the deformation analysis and early warning module and is used for remote data transmission and storage.

2. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: The various sensor modules include fiber optic grating sensors, BeiDou positioning modules, MEMS accelerometers, and tilt sensors. The fiber optic grating sensor is used to monitor structural strain, and its wavelength demodulation accuracy is ±0.1 pm; The Beidou positioning module is used to monitor structural displacement, with a positioning accuracy of ±2cm in the horizontal direction and ±3cm in the vertical direction. The MEMS accelerometer is used to monitor vibration acceleration, with a range of ±10g. The tilt sensor is used to monitor the tilt angle with an accuracy of ±0.01°.

3. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: The data acquisition and preprocessing module uses a wavelet thresholding algorithm to process the raw data. The wavelet basis function is db4, the decomposition level is 3-5 levels, and a soft thresholding function is used. Where x is a noisy signal, The threshold value is used.

4. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: When the data fusion processing module uses the Kalman filter algorithm, the state equation is: The observation equation is ,in Let k be the state vector at time k. Here is the state transition matrix. The process noise covariance matrix is... For the observation vector, For the observation matrix, To observe the noise covariance matrix.

5. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: The deformation analysis and early warning module is equipped with three levels of early warning thresholds; A yellow alert is issued when the fused data exceeds the first-level warning threshold. An orange alert will be issued when the level-two warning threshold is exceeded. A red alert will be issued when the threshold for a Level 3 warning is exceeded. The warning thresholds at each level are determined based on engineering structural design standards and historical monitoring data.

6. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: The communication and storage module adopts dual-mode communication of 5G and NB-IoT. 5G is used for real-time high-speed data transmission, and NB-IoT is used for low-power data transmission. Data storage adopts a time-series database with a storage period of 5-15 minutes.

7. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: It also includes a power supply module, which uses a combination of solar panels and lithium batteries for power supply. The solar panel has a power of 120-180W and the lithium battery has a capacity of 60-100Ah.

8. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: When constructing the finite element model, the deformation analysis and early warning module uses tetrahedral elements to mesh the engineering structure. The element size is determined according to the key parts of the structure, and the element size of the key parts is 0.5-1m.

9. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: Before data fusion, the data fusion processing module performs time synchronization on data from different types of sensors, using the NTP protocol to achieve time synchronization with an accuracy of ±1ms.

10. The dynamic deformation monitoring system for engineering structures based on multi-sensor fusion according to claim 1, characterized in that: The multi-type sensor module also includes a laser rangefinder sensor for monitoring changes in the width of structural cracks, with a measurement accuracy of ±0.02 mm.

Images