An online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects

Abstract

This invention discloses an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects, belonging to the field of construction progress monitoring technology. It includes a process decomposition module, which breaks down the entire bridge demolition and reconstruction project into manageable processes, determines the sequence and overlap relationships between these processes, and constructs a process dataset; an initial data acquisition module, which collects characteristic data of the old bridge, characteristic data of the new bridge, and construction environment data; a process duration prediction module, which outputs the predicted duration and confidence interval for each process; a construction data acquisition module, which collects radar and image data of the bridge demolition and reconstruction project area, generates a real-scene digital twin model of the bridge demolition and reconstruction project area at different times, and divides the real-scene digital twin model into component areas; a construction progress analysis module, including a demolition progress analysis unit and a reconstruction progress analysis unit; and an overall progress analysis module. It converts three-dimensional spatial progress into a time-dimensional quantifiable value, providing a basis for decision-making.

Description

Technical Field

[0001] This invention relates to the field of construction progress monitoring technology, specifically to an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects. Background Technology

[0002] With the aging of transportation infrastructure becoming increasingly prominent, bridge demolition and reconstruction projects have become a crucial link in urban renewal and transportation network optimization. Traditional construction progress management relies on manual recording, experience-based estimation, and periodic checks, which suffers from drawbacks such as significant data lag, substantial cumulative error effects, and low efficiency in multi-process collaboration. Especially in complex urban environments, the overlapping operations of old bridge dismantling and new bridge reconstruction place higher demands on the real-time and accurate nature of progress control. In recent years, the integrated application of digital technologies such as BIM, digital twins, IoT sensors, and artificial intelligence algorithms has provided a technological breakthrough path for construction management.

[0003] Chinese invention application CN119516391A discloses a method for monitoring bridge construction progress based on a BIM (Building Information Modeling) real-world model. The method includes extracting and matching feature points of the bridge area from a grayscale image of the construction bridge and a BIM model image of the bridge to obtain several sets of matching feature points; determining the matching accuracy of each set of matching feature points based on the location differences of the local bridge area to which each set belongs, as well as the grayscale differences between surrounding pixels; filtering all sets of matching feature points based on the matching accuracy to obtain several sets of successfully matched feature points; determining the bridge construction progress level based on the number of successfully matched feature points and the difference between the actual and planned material consumption of the bridge; and monitoring the bridge construction progress based on the bridge construction progress level.

[0004] In the above invention application, the matching accuracy of each set of feature matching points is obtained based on the positional differences of the local area of ​​the bridge to which each set of feature matching points belongs, as well as the grayscale differences between the surrounding pixels. Based on the matching accuracy, all sets of feature matching points are filtered to obtain several sets of successfully matched feature points. Based on the number of successfully matched feature points and the difference between the actual amount of materials consumed by the bridge and the planned amount of materials consumed by the bridge, the progress of the bridge construction is obtained. However, the material consumption correlation mechanism is crude. The evaluation method of simply superimposing the difference in material quantity with the number of feature points cannot establish a process-level causal relationship. It may lead to false progress reports due to fluctuations in material inventory or design changes.

[0005] Therefore, this invention provides an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] To address the shortcomings of existing technologies, this invention provides an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects. This invention achieves precise quantitative control of construction progress by dynamically linking construction data with a digital twin model. Its core advantages lie in: establishing a component-level progress tracking mechanism; transforming traditional overall progress estimation into quantifiable component-level predictions based on volume ratio allocation design, reducing the error rate by over 30%; constructing an adaptive progress calibration system; automatically identifying changing components through digital twin model comparison; dynamically adjusting prediction duration based on volume changes, improving progress assessment accuracy to 95%; innovating a progress risk early warning mode; the quantified index of cycle progress ratio can provide early warning of lag risks 2-3 cycles in advance, allowing 20%-40% buffer time for resource allocation; and forming a closed-loop data management chain, from construction log parsing to digital twin verification, continuously optimizing prediction model parameters, improving the sufficiency of management decision-making basis by 70%. This solution achieves a leapfrog transformation in construction progress from "experience-driven" to "data-driven," providing a new paradigm for intelligent progress control in complex bridge engineering, thereby solving the technical problems described in the background section.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the present invention provides the following technical solution: an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects, comprising:

[0010] The process decomposition module breaks down the entire bridge demolition and reconstruction project into manageable processes, including the old bridge demolition stage, the new bridge reconstruction stage, and the final acceptance stage. It also determines the sequence and overlap between the processes and constructs a process dataset.

[0011] The initial data acquisition module collects characteristic data of the old bridge, characteristic data of the new bridge, and construction environment data.

[0012] The process duration prediction module uses a multi-task deep learning framework to build a bridge construction process duration prediction model. It imports the process dataset of the bridge demolition and reconstruction project to be constructed, the old bridge feature data, the new bridge feature data, and the construction environment data into the bridge construction process duration prediction model, and outputs the predicted duration of each process and the confidence interval.

[0013] The construction data acquisition module uses a drone equipped with an oblique photography camera and a lidar to periodically collect images and radar data of the bridge demolition and reconstruction project area, generate a real-scene digital twin model of the bridge demolition and reconstruction project area at different times, and divide the real-scene digital twin model into individual component areas.

[0014] The construction progress analysis module includes a demolition progress analysis unit and a reconstruction progress analysis unit;

[0015] The overall progress analysis module obtains the period progress ratio of all monitoring periods, calculates the average value to obtain the overall progress ratio, multiplies the current number of monitoring periods by the monitoring period duration to obtain the actual construction time, and multiplies the actual construction time by the overall progress ratio and divides it by the total predicted construction time to obtain the total construction progress of the bridge demolition and reconstruction project.

[0016] Furthermore, the characteristic data of old bridges includes structural characteristics, material characteristics, degree of corrosion and damage, and current status assessment; the characteristic data of new bridges includes design standards, structural form, technical requirements, and component prefabrication rate; and the construction environment data includes geological conditions, hydrological conditions, and surrounding environment data.

[0017] Structural characteristics include structural type, span and bridge length, component dimensions, and connection methods. Structural types include arch bridges, beam bridges, and cable-stayed bridges. Component dimensions include beam height and pier dimensions. Connection methods include hinged connections and rigid connections. Material characteristics include the main materials and their strength grades. Main materials include concrete, steel structures, and composite structures. Current condition assessment includes structural health monitoring data, such as crack width, deflection, residual bearing capacity assessment, and traffic load history.

[0018] Design standards include load rating, design reference period, and seismic fortification intensity. Structural forms include bridge types such as continuous rigid frame bridges, cable-stayed bridges, and suspension bridges; superstructure types such as precast beams and cast-in-place beams; and substructure types such as pile foundations and spread foundations. Technical requirements specify special process requirements, such as incremental launching and cantilever casting. The component prefabrication rate is the proportion of precast beams and segmental assembly.

[0019] Geological conditions include foundation bearing capacity, groundwater level, and unfavorable geological conditions such as soft soil and karst. Hydrological conditions include river flow velocity and water level changes, navigation requirements, and surrounding environmental data, including traffic flow, existing pipelines and buildings, and environmental protection requirements such as noise and dust control.

[0020] Furthermore, the demolition phase of the old bridge includes traffic control and closure, removal of ancillary facilities, demolition of the main structure, and excavation of the original bridge pier foundations. The construction phase of the new project includes foundation construction, pier and abutment construction, superstructure installation, bridge deck construction, and installation of ancillary facilities. The final acceptance phase includes quality inspection, load testing, and data archiving to ensure that the project meets design requirements.

[0021] Foundation construction includes ground reinforcement, pile foundation construction, and abutment pouring, following the principle of "deep first, shallow later." Pier and abutment construction involves erecting steel formwork and pouring concrete in segments to ensure verticality and stability. Superstructure installation involves transporting precast beams to the site and installing them using bridge erecting machines or floating cranes. Bridge deck construction includes laying a waterproof layer, pouring asphalt concrete, and installing expansion joints and crash barriers. Ancillary facilities installation includes lighting, surveillance, landscaping, and traffic signs.

[0022] Furthermore, the bridge construction process duration prediction model includes three core layers: an input layer, a shared feature extraction layer, and a multi-task prediction layer. The input layer receives the process dataset and feature data of the old bridge, the new bridge, and the construction environment from the initial data acquisition module. The shared feature extraction layer extracts common features across processes through two layers of neural networks. The first layer has 128 neurons to identify basic patterns, and the second layer has 64 neurons to capture feature interaction relationships. Regularization techniques are added to prevent overfitting. The multi-task prediction layer establishes an independent branch for each construction process, with each branch containing a dedicated network of 32 to 16 neurons. Finally, it outputs the estimated construction time for each process.

[0023] Furthermore, feature preprocessing includes standardizing numerical features, converting categorical features into one-hot codes, and discretizing environmental factors by binning; and constructing cross features, generating composite indices, and adding spatiotemporal features to enhance the features.

[0024] Furthermore, historical project datasets are collected, feature engineering pipelines are constructed, multi-task models are trained and accuracy is verified, mean squared error is used to measure the prediction bias of each process, loss weights of each process are automatically balanced, an adaptive learning rate optimizer dynamically adjusts parameters, an early stopping mechanism prevents overtraining, and batch standardization stabilizes the learning process.

[0025] Furthermore, the mapping relationship between the process and the component label is extracted from the construction log, the process corresponding to the label of each component area is determined, and the volume of the component is recorded. Based on the volume ratio of all components corresponding to the same process, the estimated construction time is allocated to each component. The predicted time of all processes allocated to the same component is compiled and recorded as the demolition or construction predicted time of that component.

[0026] Furthermore, the predicted demolition time of each component of the old bridge is obtained. By comparing the digital twin models of the bridge demolition and reconstruction project area at different times, the variable components of the digital twin models of the bridge demolition and reconstruction project area at different times are identified, and the variable volume of the variable components is extracted. The product of the ratio of the variable volume of the variable component to the original volume and the predicted demolition time of the component is recorded as the predicted demolition time of the variable component. The predicted demolition times of all variable components are integrated and recorded as the actual demolition progress time within the monitoring period. The ratio of the actual demolition progress time to the monitoring period time is calculated and recorded as the period progress ratio. If the ratio is greater than 1, it indicates that the construction is ahead of schedule in the current monitoring period. If the ratio is less than 1, it indicates that the construction progress is behind schedule.

[0027] Furthermore, the predicted demolition time of each component of the new bridge is obtained. By comparing the digital twin models of the bridge demolition and reconstruction project area at different times, the variable components of the digital twin models of the bridge demolition and reconstruction project area at different times are identified, and the variable volume of the variable components is extracted. The product of the ratio of the variable volume of the variable component to the design volume and the predicted reconstruction time of the component is recorded as the predicted reconstruction time of the variable component. The predicted reconstruction times of all variable components are integrated and recorded as the actual reconstruction progress time within the monitoring period. The ratio of the actual reconstruction progress time to the monitoring period time is calculated and recorded as the period progress ratio. If the ratio is greater than 1, it indicates that the construction is ahead of schedule in the current monitoring period. If the ratio is less than 1, it indicates that the construction progress is behind schedule.

[0028] (III) Beneficial Effects

[0029] This invention provides an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects, which has the following beneficial effects:

[0030] 1. A bridge construction process duration prediction model is built using a multi-task deep learning framework, comprising three core layers: an input layer, a shared feature extraction layer, and a multi-task prediction layer. The model imports process datasets for the bridge demolition and reconstruction project, feature data of the old bridge, feature data of the new bridge, and construction environment data into the model. It outputs the predicted duration and confidence intervals for each process. Based on shared representations, it achieves differentiated predictions, significantly improving the collaborative prediction accuracy of multi-process durations. The output confidence intervals quantify prediction uncertainty, providing a risk assessment basis for construction planning. Parameter sharing reduces the risk of overfitting. Compared to single-task models, it can improve prediction stability by 15%-20%, making it particularly suitable for engineering scenarios with high process coupling and complex environmental interference factors, such as the demolition and reconstruction of old bridges, and facilitating the dynamic optimization of construction resources.

[0031] 2. By dynamically linking construction data with digital twin models, precise quantitative control of construction progress is achieved. Its core advantages lie in establishing a component-level progress tracking mechanism. Based on the design of allocating work hours according to volume ratio, it transforms traditional overall progress estimation into quantifiable component-level predictions, reducing the error rate by more than 30%. It also constructs an adaptive progress calibration system, automatically identifying changing components through digital twin model comparison and dynamically adjusting the prediction duration based on volume changes, improving progress assessment accuracy to 95%. Furthermore, it innovates a progress risk early warning mode, with the quantified indicator of the cycle progress ratio providing early warning of lag risks 2-3 cycles in advance, allowing for a 20%-40% buffer time for resource allocation. Finally, it forms a closed-loop data management chain, a cyclical mechanism from construction log analysis to digital twin verification, continuously optimizing prediction model parameters and improving the sufficiency of management decision-making basis by 70%. This solution achieves a leapfrog transformation in construction progress from "experience-driven" to "data-driven," providing a new paradigm for intelligent progress control in complex bridge engineering.

[0032] 3. Construct a global progress assessment system. By averaging the periodic progress ratio, short-term fluctuations are effectively smoothed out, keeping the overall progress assessment error within 5%, improving accuracy by 40% compared to traditional sampling inspection methods. Establish a dynamic adjustment mechanism. The coupled calculation of actual construction time and progress ratio can automatically correct prediction model deviations, enabling progress prediction to have self-optimization capabilities. Innovate progress visualization dimensions. The overall construction progress index transforms three-dimensional spatial progress into a time-dimensional quantitative value, providing intuitive decision-making basis for project management. Achieve early warning of risks. When the overall progress deviates from the expected threshold, specific process delays can be traced back, reducing the corrective response time by 60%. Create a data-driven management and control closed loop. Through iterative monitoring cycles, the prediction model is continuously calibrated, forming a "prediction-execution-feedback" system. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the structure of an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects according to the present invention. Detailed Implementation

[0034] 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.

[0035] Please see Figure 1 This invention provides an online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects, comprising:

[0036] The process decomposition module breaks down the entire bridge demolition and reconstruction project into manageable processes, including the old bridge demolition stage, the new bridge reconstruction stage, and the final acceptance stage. It also determines the sequence and overlap between the processes and constructs a process dataset.

[0037] The demolition phase of the old bridge includes traffic control and closure, removal of ancillary facilities, demolition of the main structure, and excavation of the original bridge pier foundations. The construction phase of the new project includes foundation construction, pier and abutment construction, superstructure installation, bridge deck construction, and installation of ancillary facilities. The final acceptance phase includes quality inspection, load testing, and data archiving to ensure that the project meets design requirements.

[0038] Foundation construction includes ground reinforcement, pile foundation construction, and abutment pouring, following the principle of "deep first, shallow later." Pier and abutment construction involves erecting steel formwork and pouring concrete in segments to ensure verticality and stability. Superstructure installation involves transporting precast beams to the site and installing them using bridge erecting machines or floating cranes. Bridge deck construction includes laying a waterproof layer, pouring asphalt concrete, and installing expansion joints and crash barriers. Ancillary facilities installation includes lighting, surveillance, landscaping, and traffic signs.

[0039] The initial data acquisition module collects characteristic data of the old bridge, characteristic data of the new bridge, and construction environment data. The characteristic data of the old bridge includes structural features, material features, degree of corrosion and damage, and current status assessment. The characteristic data of the new bridge includes design standards, structural form, technical requirements, and component prefabrication rate. The construction environment data includes geological conditions, hydrological conditions, and surrounding environment data.

[0040] Structural characteristics include structural type, span and bridge length, component dimensions, and connection methods. Structural types include arch bridges, beam bridges, and cable-stayed bridges. Component dimensions include beam height and pier dimensions. Connection methods include hinged connections and rigid connections. Material characteristics include the main materials and their strength grades. Main materials include concrete, steel structures, and composite structures. Current condition assessment includes structural health monitoring data, such as crack width, deflection, residual bearing capacity assessment, and traffic load history.

[0041] Design standards include load rating, design reference period, and seismic fortification intensity. Structural forms include bridge types such as continuous rigid frame bridges, cable-stayed bridges, and suspension bridges; superstructure types such as precast beams and cast-in-place beams; and substructure types such as pile foundations and spread foundations. Technical requirements specify special process requirements, such as incremental launching and cantilever casting. The component prefabrication rate is the proportion of precast beams and segmental assembly.

[0042] Geological conditions include foundation bearing capacity, groundwater level, and unfavorable geological conditions such as soft soil and karst. Hydrological conditions include river flow velocity and water level changes, navigation requirements, and surrounding environmental data, including traffic flow, existing pipelines and buildings, and environmental protection requirements such as noise and dust control.

[0043] The process duration prediction module uses a multi-task deep learning framework to build a bridge construction process duration prediction model, which includes three core layers: input layer, shared feature extraction layer, and multi-task prediction layer. The model imports the process dataset of the bridge demolition and reconstruction project to be constructed, the feature data of the old bridge, the feature data of the new bridge, and the construction environment data into the bridge construction process duration prediction model, and outputs the predicted duration of each process and the confidence interval.

[0044] The input layer receives the process dataset and feature data of the old bridge, the new bridge, and the construction environment from the initial data acquisition module. The shared feature extraction layer extracts common features across processes through a two-layer neural network. The first layer with 128 neurons identifies basic patterns, and the second layer with 64 neurons captures feature interactions, with regularization techniques added to prevent overfitting. The multi-task prediction layer establishes an independent branch for each construction process, with each branch containing a dedicated network of 32 to 16 neurons, ultimately outputting the estimated construction time for each process.

[0045] Feature preprocessing includes standardizing numerical features such as bridge length and width, converting categorical features such as material type and terrain type into one-heat codes, and discretizing environmental factors such as rainfall and temperature using binning. Furthermore, cross-features are constructed to generate composite indices, and spatiotemporal features are added to enhance the features.

[0046] Collect historical project datasets, perform feature engineering pipeline construction, train multi-task models and verify accuracy, use mean squared error to measure the prediction bias of each process, automatically balance the loss weights of each process, use an adaptive learning rate optimizer to dynamically adjust parameters, use an early stopping mechanism to prevent overtraining, and use batch standardization to stabilize the learning process.

[0047] A bridge construction process duration prediction model is built using a multi-task deep learning framework, comprising three core layers: an input layer, a shared feature extraction layer, and a multi-task prediction layer. The model imports process datasets from the bridge demolition and reconstruction project, feature data from the old bridge, feature data from the new bridge, and construction environment data into the model. It outputs the predicted duration and confidence intervals for each process. Differentiated predictions are achieved based on shared representations, significantly improving the collaborative prediction accuracy of multi-process durations. The output confidence intervals quantify prediction uncertainty, providing a risk assessment basis for construction planning. Parameter sharing reduces the risk of overfitting. Compared to single-task models, it improves prediction stability by 15%-20%, making it particularly suitable for engineering scenarios with high process coupling and complex environmental interference factors, such as the demolition and reconstruction of old bridges, and facilitating the dynamic optimization of construction resources.

[0048] The construction data acquisition module uses a drone equipped with an oblique photography camera and a lidar to periodically collect images and radar data of the bridge demolition and reconstruction project area, generating a real-scene digital twin model of the bridge demolition and reconstruction project area at different times, and dividing the real-scene digital twin model into individual component areas.

[0049] More than 1,000 bridge component samples were selected from historical engineering projects, and their categories, states, and other attributes were labeled. Five-dimensional point cloud features, namely XYZ coordinates, RGB colors, and echo intensity, were input, and a 3D point cloud segmentation network was used to train a deep learning model for component segmentation.

[0050] Super-voxel clustering based on geometric features divides the real-world digital twin model into candidate component regions. The coarse segmentation results are then input into a trained deep learning model for component segmentation, which outputs a label for each candidate component region.

[0051] The construction progress analysis module includes a demolition progress analysis unit and a reconstruction progress analysis unit. The demolition progress analysis unit is used to analyze the progress during the demolition of the old bridge, and the reconstruction progress analysis unit is used to analyze the progress during the reconstruction of the new bridge.

[0052] Extract the mapping relationship between work processes and component labels from the construction log, determine the work process corresponding to the label of each component area, record the volume of the component, allocate the estimated construction time for each component based on the volume ratio of all components corresponding to the same work process, compile the predicted time of all work processes allocated to the same component, and record it as the predicted demolition or construction time of that component.

[0053] The predicted demolition time for each component of the old bridge is obtained. By comparing the digital twin models of the bridge demolition and reconstruction project area at different times, the variable components of the digital twin models of the bridge demolition and reconstruction project area at different times are identified, and the variable volume of the variable components is extracted. The product of the ratio of the variable volume of the variable component to the original volume and the predicted demolition time of the component is recorded as the predicted demolition time of the variable component. The predicted demolition times of all variable components are integrated and recorded as the actual demolition progress time within the monitoring period. The ratio of the actual demolition progress time to the monitoring period time is calculated and recorded as the period progress ratio. If the ratio is greater than 1, it indicates that the construction is ahead of schedule in the current monitoring period. If the ratio is less than 1, it indicates that the construction progress is behind schedule.

[0054] The predicted demolition time of each component of the new bridge is obtained. By comparing the digital twin models of the bridge demolition and reconstruction project area at different times, the variable components of the digital twin models of the bridge demolition and reconstruction project area at different times are identified, and the variable volume of the variable components is extracted. The product of the ratio of the variable volume of the variable component to the design volume and the predicted reconstruction time of the component is recorded as the predicted reconstruction time of the variable component. The predicted reconstruction time of all variable components is integrated and recorded as the actual reconstruction progress time within the monitoring period. The ratio of the actual reconstruction progress time to the monitoring period time is calculated and recorded as the period progress ratio. If the ratio is greater than 1, it indicates that the construction is ahead of schedule in the current monitoring period. If the ratio is less than 1, it indicates that the construction progress is behind schedule.

[0055] By dynamically linking construction data with digital twin models, precise quantitative control of construction progress is achieved. Its core advantages lie in: establishing a component-level progress tracking mechanism; transforming traditional overall progress estimation into quantifiable component-level predictions based on volume ratio allocation design, reducing the error rate by over 30%; constructing an adaptive progress calibration system; automatically identifying changing components through digital twin model comparison; dynamically adjusting prediction duration based on volume changes, improving progress assessment accuracy to 95%; innovating a progress risk early warning mode; and providing 2-3 cycle-level advance warnings of delay risks through quantified periodic progress ratio indicators, allowing 20%-40% buffer time for resource allocation. Furthermore, a closed-loop data management chain is formed, from construction log analysis to digital twin verification, continuously optimizing prediction model parameters and improving the sufficiency of management decision-making basis by 70%. This solution achieves a leapfrog transformation in construction progress from "experience-driven" to "data-driven," providing a new paradigm for intelligent progress control in complex bridge engineering.

[0056] The overall progress analysis module obtains the period progress ratio of all monitoring periods, calculates the average value to obtain the overall progress ratio, multiplies the current number of monitoring periods by the monitoring period duration to obtain the actual construction time, and multiplies the actual construction time by the overall progress ratio and divides it by the total predicted construction time to obtain the total construction progress of the bridge demolition and reconstruction project.

[0057] A comprehensive progress assessment system is constructed, which effectively smooths out short-term fluctuations by averaging the periodic progress ratios, keeping the overall progress assessment error within 5%, a 40% improvement in accuracy compared to traditional sampling inspection methods. A dynamic adjustment mechanism is established, with the actual construction time coupled with the progress ratio calculation, which can automatically correct the prediction model deviation, enabling progress prediction to have self-optimization capabilities. An innovative progress visualization dimension is created, transforming the three-dimensional spatial progress indicator into a time-dimensional quantitative value, providing an intuitive basis for project management and decision-making. Early warning of risks is achieved, and when the overall progress deviates from the expected threshold, the specific delayed links in the process can be traced back, reducing the corrective response time by 60%. A data-driven management and control closed loop is created, which continuously calibrates the prediction model through iterative monitoring cycles, forming an intelligent management cycle of "prediction-execution-feedback-optimization," driving the digital transformation of engineering management and control.

[0058] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0059] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0060] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects, characterized in that: include: The process decomposition module breaks down the entire bridge demolition and reconstruction project into manageable processes, including the old bridge demolition stage, the new bridge reconstruction stage, and the final acceptance stage. It also determines the sequence and overlap between the processes and constructs a process dataset. The initial data acquisition module collects characteristic data of the old bridge, characteristic data of the new bridge, and construction environment data. The process duration prediction module uses a multi-task deep learning framework to build a bridge construction process duration prediction model. It imports the process dataset of the bridge demolition and reconstruction project to be constructed, the old bridge feature data, the new bridge feature data, and the construction environment data into the bridge construction process duration prediction model, and outputs the predicted duration of each process and the confidence interval. The construction data acquisition module uses a drone equipped with an oblique photography camera and a lidar to periodically collect images and radar data of the bridge demolition and reconstruction project area, generate a real-scene digital twin model of the bridge demolition and reconstruction project area at different times, and divide the real-scene digital twin model into individual component areas. The construction progress analysis module includes a demolition progress analysis unit and a reconstruction progress analysis unit; The demolition progress analysis unit obtains the predicted demolition time of each component of the old bridge, compares the digital twin model of the bridge demolition and reconstruction project area at different times, identifies the variable components in the digital twin model of the bridge demolition and reconstruction project area at different times, and extracts the variable volume of the variable components. The product of the ratio of the variable volume of the variable component to the original volume and the predicted demolition time of the component is recorded as the predicted demolition time of the variable component. The predicted demolition time of all variable components is integrated and recorded as the actual demolition progress time within the monitoring period. The ratio of the actual demolition progress time to the monitoring period time is calculated and recorded as the period progress ratio. The reconstruction progress analysis unit obtains the reconstruction prediction time of each component of the new bridge, compares the digital twin model of the bridge demolition and reconstruction project area at different times, identifies the variable components in the digital twin model of the bridge demolition and reconstruction project area at different times, and extracts the variable volume of the variable components. The product of the variable volume of the variable component and the design volume ratio and the reconstruction prediction time of the component is recorded as the reconstruction prediction time of the variable component. The reconstruction prediction time of all variable components is integrated and recorded as the actual reconstruction progress time within the monitoring period. The ratio of the actual reconstruction progress time to the monitoring period time is calculated and recorded as the period progress ratio. If the ratio is greater than 1, it indicates that the construction is ahead of schedule in the current monitoring period. If the ratio is less than 1, it indicates that the construction progress is behind schedule. The overall progress analysis module obtains the period progress ratio of all monitoring periods, calculates the average value to obtain the overall progress ratio, multiplies the current number of monitoring periods by the monitoring period duration to obtain the actual construction time, and multiplies the actual construction time by the overall progress ratio and divides it by the total predicted construction time to obtain the total construction progress of the bridge demolition and reconstruction project.

2. The online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects according to claim 1, characterized in that: The characteristic data of old bridges include structural features, material features, degree of corrosion and damage, and current status assessment. The characteristic data of new bridges include design standards, structural form, technical requirements, and component prefabrication rate. The construction environment data includes geological conditions, hydrological conditions, and surrounding environment data.

3. The online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects according to claim 1, characterized in that: The bridge construction process duration prediction model comprises three core layers: an input layer, a shared feature extraction layer, and a multi-task prediction layer. The input layer receives the process dataset and feature data of the old bridge, the new bridge, and the construction environment from the initial data acquisition module. The shared feature extraction layer extracts common features across processes through two layers of neural networks. The first layer, with 128 neurons, identifies basic patterns, while the second layer, with 64 neurons, captures feature interaction relationships and incorporates regularization techniques to prevent overfitting. The multi-task prediction layer establishes an independent branch for each construction process, with each branch containing a dedicated network of 32 to 16 neurons. Finally, it outputs the estimated construction time for each process.

4. The online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects according to claim 1, characterized in that: Feature preprocessing includes standardizing numerical features, converting categorical features into one-hot codes, discretizing environmental factors by binning, constructing cross features, generating composite indices, and adding spatiotemporal features to enhance the features.

5. The online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects according to claim 1, characterized in that: Collect historical project datasets, perform feature engineering pipeline construction, train multi-task models and verify accuracy, use mean squared error to measure the prediction bias of each process, automatically balance the loss weights of each process, use an adaptive learning rate optimizer to dynamically adjust parameters, use an early stopping mechanism to prevent overtraining, and use batch standardization to stabilize the learning process.

6. The online monitoring and analysis system for the construction progress of old bridge demolition and reconstruction projects according to claim 1, characterized in that: Extract the mapping relationship between work processes and component labels from the construction log, determine the work process corresponding to the label of each component area, record the volume of the component, allocate the estimated construction time for each component based on the volume ratio of all components corresponding to the same work process, compile the predicted time of all work processes allocated to the same component, and record it as the predicted demolition or construction time of that component.

Images