Large-span continuous rigid frame construction linear intelligent monitoring method

Abstract

This invention discloses an intelligent monitoring method for the construction alignment of long-span continuous rigid frames, relating to the field of long-span continuous rigid frame construction technology. The method includes the following steps: S1, data acquisition; S2, data processing; S3, multi-level intelligent early warning; S4, construction control; and S5, output of results. This invention utilizes multi-source devices to collect data synchronously, ensuring data real-time performance and integrity. Deep learning removes anomalies, and neural networks construct environmental impact models to correct data, mitigating the impact of environmental factors on alignment data. Digital twins combined with finite element analysis and neural networks predict deformation trends, improving prediction accuracy, and allowing for early prediction of alignment deviation trends and optimization of construction plans. Multi-level intelligent early warning dynamically adjusts thresholds based on multiple factors, accurately issuing early warning information and enhancing the timeliness of risk response. Construction control automatically generates decision suggestions and provides feedback on evaluation results, optimizing the construction process.

Description

Technical Field

[0001] This invention relates to the field of construction technology for long-span continuous rigid frames, specifically to an intelligent monitoring method for the construction alignment of long-span continuous rigid frames. Background Technology

[0002] Long-span continuous rigid frame bridges are characterized by pier-beam integration and continuous main beams. The spans often reach 100-300 meters. During construction, they need to be formed segment by segment through cantilever casting or segmental assembly. The requirements for deformation control are extremely high. During the construction stage, the beam shape is prone to dynamic changes due to factors such as concrete shrinkage and creep, prestressing tension, temperature changes, and load accumulation. Intelligent monitoring is required to make real-time adjustments to ensure closure accuracy.

[0003] For example, the publication number CN102998136A, "A Method for Monitoring the Alignment of a Prestressed Concrete Continuous Rigid Frame Bridge", includes the following steps: a) During the cantilever construction of the main beam, determine the formwork elevation of the main beam and establish an alignment monitoring network for the main beam; b) From before the cantilever construction of the main beam to after the completion of the bridge, measure the alignment of the main beam and measure the stress and temperature of the main beam.

[0004] In existing technologies, it is necessary to obtain data on beam elevation, deflection, and stress during construction to detect and adjust deviations in a timely manner. However, traditional monitoring methods mostly rely on periodic measurements, which result in delayed data updates. This makes it difficult to respond promptly to sudden deformations or environmental changes during construction, and it is also difficult to predict the trend of alignment deviations in advance, which can easily lead to false alarms or missed alarms. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent monitoring method for the construction alignment of long-span continuous rigid frames, so as to solve the problem mentioned in the background art that it is difficult to respond in time to sudden deformations or environmental changes during construction and to predict the trend of alignment deviation in advance.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring method for the construction alignment of long-span continuous rigid frames, comprising the following steps:

[0007] S1. Data Acquisition: Real-time acquisition of data from multiple sensors and different devices to obtain multi-source data and transmit the multi-source data to the data processing center;

[0008] S2. Data Processing: Multi-source data is fused and processed to remove outliers, a coupled model of the influence of environmental factors on structural deformation is constructed, structural deformation measurement data is corrected, and structural deformation trends are predicted.

[0009] In step S2, the data processing includes the following steps:

[0010] S21. Data collection and integration: The collected multi-source data is initially integrated according to the timestamp to form a multi-source heterogeneous dataset. A deep learning model is constructed, and the deep learning model is trained. The real-time collected multi-source data is input into the trained deep learning model, and the identified abnormal data is removed.

[0011] S22. Environmental Impact Correction: Collect historical environmental data and corresponding structural deformation data, use neural network algorithms to establish an environmental factor impact model, input real-time environmental data into the established environmental factor impact model, and correct the real-time collected structural deformation measurement data.

[0012] S23. Structural Deformation Trend Prediction: Construct a digital twin model of bridge construction, perform mechanical analysis on the bridge structure at each construction stage, calculate the mechanical response of the structure, and predict the trend of structural deformation.

[0013] S3, Multi-level Intelligent Early Warning: Dynamically set early warning thresholds, monitor structural deformation and stress parameters in real time, trigger the early warning mechanism when the early warning threshold is exceeded, and issue early warning information according to the early warning level;

[0014] S4. Construction Control: Based on the test and analysis results, generate corresponding decision suggestions, and feed back the effects of the construction adjustments through the monitoring system to analyze the effectiveness of the adjustment measures.

[0015] S5. Output of Results: Classify and archive the data from the entire monitoring process according to time, construction stage, and data type, and generate a monitoring report.

[0016] Preferably, in step S1, the data acquisition includes the following steps:

[0017] S11, Multi-source data acquisition: Sensors collect construction-related data in real time;

[0018] S12. Data transmission: Utilizing 5G communication networks and low-power Bluetooth self-organizing network technology, data collected by various sensors and measuring devices are transmitted to the data processing center.

[0019] Preferably, in step S11, the multi-source data acquisition includes fiber optic strain sensors acquiring strain data, MEMS acceleration and displacement sensors acquiring structural vibration and displacement data, intelligent total station measuring target point coordinates, UAVs planning flight routes according to construction progress to acquire structural appearance and overall alignment data, laser scanning equipment acquiring three-dimensional coordinate data, cameras capturing high-definition images, environmental monitoring equipment acquiring environmental parameters, and environmental data acquisition frequency being synchronized with structural monitoring data.

[0020] Preferably, in step S21, training the deep learning model includes labeling the collected multi-source heterogeneous data, distinguishing between normal and abnormal data samples, constructing a training dataset, and using the training dataset to train the deep learning model.

[0021] Preferably, in step S23, the structural deformation trend prediction includes the following steps:

[0022] A1. Construct a digital twin model of the bridge construction process, divide the entire construction process into multiple stages, simulate the loads borne by the bridge at different construction stages, calculate the mechanical response of the structure using the finite element method, and obtain the finite element analysis results.

[0023] A2. Use neural network algorithms to construct a mechanical response analysis model, collect real-time measured data of the mechanical response of the bridge structure, calculate the error between the results and the finite element analysis results, and update and correct the parameters of the mechanical response analysis model.

[0024] A3. Input the current construction stage model status and real-time collected relevant data into the updated mechanical response analysis model to simulate and predict the structural mechanical behavior in the next few construction stages, and obtain the development trend of structural deformation.

[0025] Preferably, in step S3, the multi-level intelligent early warning includes the following steps:

[0026] S31. Collect various data during the bridge construction process, including structural deformation, stress, temperature, humidity, construction progress, hanging basket pre-camber, and prestressing tension value; label the data; and dynamically calculate the warning threshold suitable for the current situation.

[0027] S32. Real-time monitoring of key parameters of structural deformation and stress. When key parameters exceed the dynamically adjusted warning threshold, an early warning mechanism is automatically triggered, and warning information is issued to relevant construction and management personnel according to the warning level.

[0028] Preferably, the warning levels include Level 1, Level 2, and Level 3, and the warning information includes the type of warning, the location where it occurred, the current parameter value, and the corresponding warning level.

[0029] Preferably, in step S4, the decision recommendations include automatically optimizing construction parameters, adjusting the construction sequence, and taking temporary reinforcement measures.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] This invention utilizes multi-source devices to collect data synchronously, and 5G and Bluetooth self-organizing networks to achieve efficient transmission, ensuring data real-time performance and integrity. In the data processing stage, deep learning removes anomalies, neural networks construct environmental impact models and correct data, correcting the impact of environmental factors on linear data. Digital twins, combined with finite element analysis and neural networks, predict deformation trends, improving prediction accuracy, predicting linear deviation trends in advance, and optimizing construction plans. Multi-level intelligent early warning dynamically adjusts thresholds based on multiple factors, accurately issuing early warning information and enhancing the timeliness of risk response. Construction control can automatically generate decision suggestions and provide feedback on evaluation results, optimizing the construction process. Through the integration of multiple technologies, the entire process from data acquisition to construction control is intelligent, improving the accuracy and efficiency of construction linear monitoring and ensuring construction safety and quality. Attached Figure Description

[0032] Figure 1 This is a flowchart of the intelligent monitoring method for the construction alignment of long-span continuous rigid frames according to the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Example: Please refer to Figure 1 This invention provides a technical solution: an intelligent monitoring method for the construction alignment of long-span continuous rigid frames, comprising the following steps:

[0035] S1. Data Acquisition: Real-time acquisition of data from multiple sensors and different devices to obtain multi-source data and transmit the multi-source data to the data processing center;

[0036] Data collection includes the following steps:

[0037] S11. Multi-source data acquisition: Sensors acquire structural strain, displacement, and vibration data in real time. Fiber optic strain sensors acquire strain data multiple times per second. MEMS accelerometers and displacement sensors acquire structural vibration and displacement data at high frequency. Intelligent total stations automatically measure the coordinates of target points at a set frequency. After each measurement, the accuracy of the measurement data is evaluated. If the accuracy does not meet the requirements, the measurement is automatically re-measured. UAVs plan their flight routes according to the construction schedule and acquire structural appearance and overall alignment data. Laser scanning equipment continuously scans during flight, acquiring a large amount of three-dimensional coordinate data per second. Cameras capture high-definition images at certain time intervals. Environmental monitoring equipment synchronously acquires environmental parameters such as temperature, humidity, wind speed, and wind direction. The frequency of environmental data acquisition is synchronized with the structural monitoring data to ensure the time matching of the data.

[0038] S12. Data transmission: Utilizing 5G communication network and low-power Bluetooth self-organizing network technology, data collected by various sensors and measuring devices is transmitted to the data processing center at the construction site in real time and stably. Encryption and verification technologies are used during data transmission to ensure data integrity and accuracy.

[0039] S2. Data Processing: Multi-source data is fused and processed to remove outliers, a coupled model of the influence of environmental factors on structural deformation is constructed, structural deformation measurement data is corrected, and structural deformation trends are predicted.

[0040] Data processing includes the following steps:

[0041] S21. Data Collection and Integration: The collected multi-source data is initially integrated according to timestamps. Each data point is labeled with its precise time of generation, and all data is arranged into a unified time series to ensure consistency in the time dimension, forming a multi-source heterogeneous dataset. A deep learning model is constructed, and a large number of labeled normal and abnormal data samples are used to train the model, enabling it to distinguish between normal and abnormal data. The real-time collected multi-source data is input into the trained deep learning model. The model analyzes the data point by point based on the learned feature patterns, identifying abnormal data points that do not conform to the normal pattern. For example, data that deviates significantly from the normal range due to external interference from sensors, or erroneous data caused by malfunctions in measuring equipment. The identified abnormal data is removed from the dataset to ensure the reliability of subsequent data analysis.

[0042] S22. Environmental Impact Correction: Historical environmental data and corresponding structural deformation data are collected. The historical environmental data covers real-time records of temperature variation range, humidity fluctuation, wind speed and direction under different seasons and weather conditions. At the same time, corresponding structural deformation data are collected. These structural deformation data come from the measured results at different stages of construction. A neural network algorithm is used to establish an environmental factor impact model. Through cross-validation, the historical data is divided into multiple subsets. One part is used as the test set and the rest as the training set in turn. The model is trained and evaluated multiple times to comprehensively evaluate the accuracy and generalization ability of the model. The model parameters are continuously adjusted so that the model can accurately reflect the impact of environmental factors on structural deformation. Real-time environmental data is collected and input into the established environmental factor impact model to calculate the impact value of environmental factors on structural deformation. Based on the calculated impact value, the real-time collected structural deformation measurement data is corrected to eliminate the interference of environmental factors on structural deformation measurement and obtain more accurate actual structural deformation data.

[0043] S23. Structural Deformation Trend Prediction: Construct a digital twin model of bridge construction, perform mechanical analysis on the bridge structure at each construction stage, calculate the mechanical response of the structure, collect real-time measured data of the bridge structure, and predict the trend of structural deformation.

[0044] Structural deformation trend prediction includes the following steps:

[0045] A1. Use 3D software to construct a digital twin model of the bridge construction process. Define the geometry, material properties and connection relationships of each structural component of the bridge in the model. Divide the entire construction process into multiple stages according to the construction schedule. Assign corresponding construction status and time information to the model of each stage. Apply boundary conditions to simulate the loads borne by the bridge in different construction stages. Use the finite element method to perform mechanical analysis on the bridge structure in each construction stage and calculate the mechanical response of the structure, including stress distribution, strain change degree and deformation size.

[0046] A2. Use neural network algorithms to construct a mechanical response analysis model, learn the mapping relationship between finite element analysis results and measured data, collect measured mechanical response data of bridge structures in real time, calculate the error between the results and the finite element analysis results, update and correct the parameters of the mechanical response analysis model, and if the measured deformation is greater than the theoretical deformation, adjust the material parameters and boundary conditions in the mechanical response analysis model through neural network algorithms to make the mechanical response analysis model closer to the actual structural state.

[0047] A3. Input the current construction stage model status and real-time collected relevant data into the updated mechanical response analysis model to simulate and predict the structural mechanical behavior of the next few construction stages, obtain the development trend of structural deformation, and present the prediction results by drawing curves of structural deformation changing with time or construction stage, providing a scientific and accurate basis for construction decisions.

[0048] S3, Multi-level Intelligent Early Warning: Dynamically set early warning thresholds, monitor structural deformation and stress parameters in real time, trigger the early warning mechanism when the early warning threshold is exceeded, and issue early warning information according to the early warning level;

[0049] Multi-level intelligent early warning includes the following steps:

[0050] S31. Collect various data during bridge construction, including structural deformation, stress, temperature, humidity, construction progress, hanging basket camber, and prestressing tension value. Based on historical data and expert experience, annotate the data, define samples for different levels of early warning and corresponding construction adjustment measures. Based on the current construction stage, structural condition, environmental factors, and historical data, dynamically calculate an early warning threshold suitable for the current situation. In the early stages of construction, when the structure is relatively stable, the early warning threshold can be set more leniently; as construction progresses and the structure becomes more complex and the risk increases, the early warning threshold should be tightened accordingly.

[0051] S32. Real-time monitoring of key parameters of structural deformation and stress. When key parameters exceed the dynamically adjusted warning threshold, an early warning mechanism is automatically triggered. Warning information is issued to relevant construction and management personnel according to the warning level. The warning levels include Level 1, Level 2 and Level 3. Level 1 warning is for minor anomalies, with parameters close to the threshold; Level 2 warning is for moderate anomalies, with parameters reaching the threshold; and Level 3 warning is for severe anomalies, with parameters exceeding the threshold. The warning information should clearly indicate the type of warning, the location of the occurrence, the current parameter value and the corresponding warning level.

[0052] S4. Construction Control: Based on the test and analysis results, corresponding decision suggestions are generated. These suggestions include automatically optimizing construction parameters, such as adjusting the specific value of the hanging basket pre-camber, changing the order and magnitude of prestressing tension, adjusting the construction sequence, such as advancing or postponing certain construction procedures, and taking temporary reinforcement measures. The effects of the construction adjustments are fed back through the monitoring system. The monitoring system collects data such as structural deformation and stress in real time and transmits them to the data processing center. Data analysis algorithms are used to evaluate the effects of the adjustments, compare the data before and after the adjustments, and analyze the effectiveness of the adjustment measures.

[0053] S5. Output of Results: The data from the entire monitoring process will be classified and archived according to time, construction stage, and data type, and stored in a reliable database. Detailed monitoring reports will be generated regularly, including structural status assessment, construction adjustment status, and future trend prediction, which will be easy to access and archive.

[0054] This invention acquires data from multiple sources, including sensors, a smart total station, a drone, and environmental monitoring equipment. Sensors collect real-time data on structural strain, displacement, and vibration; the smart total station measures coordinates at different frequencies; the drone collects appearance and alignment data; and the environmental monitoring equipment simultaneously collects parameters such as temperature, humidity, wind speed, and wind direction. Using 5G and Bluetooth self-organizing network technologies, the multi-source data is transmitted to a processing center. Deep learning algorithms are used to fuse the multi-source data, identify and remove outliers, and signal processing techniques are employed for noise reduction. An environmental factor influence model is constructed to correct deformation data, eliminating the impact of environmental factors on the data. A digital twin model is then built, and finite element analysis and machine learning are combined to simulate structural forces. The system learns from behavior, updates the model in real time based on measured data, dynamically adjusts the early warning threshold, automatically issues warnings and provides decision suggestions when parameters are abnormal, and automatically adjusts construction strategies based on monitoring and analysis results. The data processing center uses data analysis algorithms to evaluate the effects of the adjustments, compares the data before and after the adjustments, analyzes the effectiveness of the adjustment measures, and further optimizes the parameters of the digital twin model and construction control strategies based on the evaluation results, forming a closed-loop control system from monitoring, analysis, adjustment to re-monitoring, continuously improving the accuracy of construction alignment control, classifying and archiving monitoring data, and regularly generating reports to present structural status, construction adjustments, and future trends.

[0055] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent monitoring of the alignment of a long-span continuous rigid frame construction, characterized in that: The method comprises the following steps: S1, data acquisition: real-time acquisition of data of multiple sensors and different devices, acquisition of multi-source data, and transmission of the multi-source data to a data processing center; S2, data processing: fusion processing of the multi-source data, elimination of abnormal data, construction of a coupling model of environmental factors on structural deformation, correction of structural deformation measurement data, and prediction of structural deformation trend; In step S2, the data processing comprises the following steps: S21, data collection and integration: preliminary integration of the collected multi-source data according to timestamps, formation of a multi-source heterogeneous data set, construction of a deep learning model, training of the deep learning model, input of real-time collected multi-source data into the trained deep learning model, and elimination of identified abnormal data; In step S21, the training of the deep learning model comprises labeling of the collected multi-source heterogeneous data, distinguishing of normal data and abnormal data samples, construction of a training data set, and training of the deep learning model using the training data set; S22, environmental influence correction: collection of historical environmental data and corresponding structural deformation data, establishment of an environmental factor influence model using a neural network algorithm, input of real-time environmental data into the established environmental factor influence model, and correction of real-time collected structural deformation measurement data; S23, structural deformation trend prediction: construction of a digital twin model of bridge construction, mechanical analysis of the bridge structure at each construction stage, calculation of the mechanical response of the structure, and prediction of the trend of structural deformation; In step S23, the structural deformation trend prediction comprises the following steps: A1, construction of a digital twin model in the bridge construction process, division of the entire construction process into multiple stages, simulation of the load borne by the bridge at different construction stages, calculation of the mechanical response of the structure using a finite element method, and obtaining of finite element analysis results; A2, construction of a mechanical response analysis model using a neural network algorithm, real-time acquisition of mechanical response measurement data of the bridge structure, calculation of the error between the mechanical response measurement data and the finite element analysis results, and updating and correction of parameters of the mechanical response analysis model; A3, input of the model state of the current construction stage and the real-time collected related data into the updated mechanical response analysis model, simulation and prediction of the mechanical behavior of the structure at future construction stages, and obtaining of the development trend of structural deformation; S3, multi-level intelligent early warning: dynamic setting of a warning threshold, real-time monitoring of structural deformation and stress parameters, triggering of a warning mechanism when the warning threshold is exceeded, and issuance of warning information according to a warning level; S4, construction regulation: generation of corresponding decision suggestions according to the detection analysis results, feedback of the effects of the construction adjustment through a monitoring system, and analysis of the effectiveness of the adjustment measures; S5, achievement output: classification and archiving of data of the entire monitoring process according to time, construction stage, and data type, and generation of a monitoring report.

2. The long-span continuous rigid frame construction alignment intelligent monitoring method according to claim 1, characterized in that: In step S1, the data acquisition comprises the following steps: S11, multi-source data acquisition: real-time acquisition of construction-related data by sensors; S12, data transmission: using 5G communication network and Bluetooth low energy ad hoc network technology, the data collected by various sensors and measuring devices are transmitted to the data processing center.

3. The long-span continuous rigid frame construction alignment intelligent monitoring method according to claim 2, characterized in that: In step S11, the multi-source data collection includes collecting strain data by fiber Bragg grating strain sensors, collecting structural vibration and displacement data by MEMS acceleration and displacement sensors, measuring target point coordinates by intelligent total station, collecting structure appearance and overall linear data by unmanned aerial vehicle according to construction progress planning flight route, obtaining three-dimensional coordinate data by laser scanning device, shooting high-definition images by camera, and collecting environmental parameters by environmental monitoring device. The frequency of environmental data collection is synchronized with the structural monitoring data.

4. The long-span continuous rigid frame construction alignment intelligent monitoring method according to claim 1, characterized in that: In step S3, the multi-level intelligent early warning includes the following steps: S31, collecting various data in the bridge construction process, including structural deformation, stress, temperature, humidity, construction progress, hanging basket camber, prestress tension value, labeling the data, and dynamically calculating the early warning threshold suitable for the current situation; S32, real-time monitoring of key parameters of structural deformation and stress, automatically triggering the early warning mechanism when the key parameters exceed the dynamically adjusted early warning threshold, and issuing early warning information to relevant construction personnel and management personnel according to the early warning level.

5. The long-span continuous rigid frame construction alignment intelligent monitoring method according to claim 4, characterized in that: The early warning level includes first, second and third levels, and the early warning information includes the type of early warning, the location of occurrence, the current parameter value and the corresponding early warning level.

6. The long-span continuous rigid frame construction alignment intelligent monitoring method according to claim 1, characterized in that: In step S4, the decision suggestion includes automatically optimizing construction parameters, adjusting construction sequence, and taking temporary reinforcement measures.

Images