Construction progress management method and system of cast-in-place box girder based on BIM

Abstract

The present invention relates to the technical field of construction progress management, and discloses a method and system for construction progress management of a cast-in-place box girder based on BIM, which is used to improve the accuracy of construction progress management of a cast-in-place box girder based on BIM. The method includes: performing correlation processing on structural data and construction progress data to obtain progress correlation data, performing simulation analysis on the progress correlation data to obtain construction plan improvement data; real-time collection of strain, temperature and vibration data to obtain original spectral data, processing and classifying the original spectral data to obtain structured construction data; performing progress analysis on the structured construction data to obtain actual progress data, comparing the actual progress data with the construction progress data to obtain progress deviation data; performing intelligent analysis on the progress deviation data to obtain early warning data and cause diagnosis data, performing decision analysis processing on the early warning data and cause diagnosis data to obtain suggestion data.

Description

Technical Field

[0001] The present invention relates to the technical field of construction progress management, and in particular to a BIM-based cast-in-place box girder construction progress management method and system. Background Art

[0002] The existing construction progress management method of cast-in-place box girders mainly relies on manual observation and regular measurement. This method usually includes traditional means such as on-site inspections, progress reports and regular meetings. With the development of building information modeling (BIM) technology, some projects have begun to apply BIM to construction progress visualization and simple collision detection. At the same time, some advanced sensing technologies, such as fiber optic sensors, have also begun to be used in structural health monitoring, providing some real-time data support for the construction process.

[0003] However, these existing technologies have some obvious shortcomings. First, traditional manual observation methods are easily affected by subjective factors and are difficult to reflect the actual construction progress in a timely and accurate manner. Secondly, although BIM technology improves the degree of visualization, it is often limited to static progress display and lacks dynamic connection with the actual construction process. Furthermore, although sensing technology can provide real-time data, this data is usually not effectively integrated with construction progress management, resulting in the value of the data not being fully mined. Finally, existing methods generally lack intelligent deviation analysis and decision support capabilities, making it difficult to quickly respond to various changes and problems in the construction process. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a BIM-based cast-in-place box girder construction progress management method and system, which are used to improve the accuracy of BIM-based cast-in-place box girder construction progress management.

[0005] The present invention provides a BIM-based cast-in-place box girder construction progress management method, comprising: digitally processing design drawings to obtain structural data, and dividing the structural data into work breakdown structures to obtain engineering task data;

[0006] Performing a schedule analysis on the engineering task data to obtain schedule network data, and adjusting the construction period of the schedule network data to obtain construction schedule data;

[0007] Associating the structure data with the construction progress data to obtain progress-associated data, and performing simulation analysis on the progress-associated data to obtain construction plan improvement data;

[0008] A fiber Bragg grating sensor network is buried in the box girder structure to collect strain, temperature and vibration data in real time to obtain raw spectral data, and the raw spectral data is processed and classified to obtain structured construction data;

[0009] Performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data;

[0010] The progress deviation data is subjected to time series analysis through a long short-term memory network to obtain early warning data and cause diagnosis data, and decision analysis and processing are performed on the early warning data and the cause diagnosis data based on the construction plan improvement data to obtain recommendation data, which includes: construction progress adjustment suggestions and resource allocation optimization plans.

[0011] In the present invention, the steps of digitizing the design drawings to obtain structural data, and dividing the structural data into work breakdown structures to obtain engineering task data include:

[0012] Vectorizing the geometric information in the design drawing to obtain structural contour data, and performing three-dimensional coordinate conversion on the structural contour data to obtain three-dimensional structural data;

[0013] Extracting features from the three-dimensional structure data to obtain structure feature data, and performing parameterized modeling on the structure according to the structure feature data to obtain parameterized structure data;

[0014] Associating the parameterized structure data with attribute information to obtain semantic structure data, and hierarchically classifying the semantic structure data to obtain layered structure data;

[0015] Performing construction sequence analysis on the structure according to the hierarchical structure data to obtain construction sequence data, and dividing the construction sequence data into process steps to obtain process unit data;

[0016] Performing resource demand analysis on the process unit data to obtain resource configuration data, and performing time estimation on the process according to the resource configuration data to obtain process duration data;

[0017] constructing a process network according to the process duration data and the construction sequence data to obtain preliminary engineering task data, and performing a critical path analysis on the preliminary engineering task data to obtain key process data;

[0018] Resource balancing is performed on the key process data to obtain balanced engineering task data, and a construction plan is generated according to the balanced engineering task data to obtain engineering task data.

[0019] In the present invention, the steps of performing a schedule analysis on the engineering task data to obtain schedule network data, and adjusting the construction period of the schedule network data to obtain construction schedule data include:

[0020] Performing task classification on the engineering task data to obtain classified task data, and performing priority sorting on the classified task data to obtain task priority data;

[0021] Performing process dependency analysis on the tasks according to the task priority data to obtain process dependency relationship data, and performing critical path calculation on the process dependency relationship data to obtain critical path data;

[0022] Estimating time parameters of the critical path data to obtain process duration data, and constructing a schedule network diagram based on the process duration data to obtain preliminary schedule network data;

[0023] Performing resource load analysis on the preliminary schedule network data to obtain resource load data, and performing balance adjustment on the schedule network according to the resource load data to obtain candidate schedule network data;

[0024] Performing a multi-objective evaluation on the candidate progress network data to obtain network evaluation index data, and selecting a progress network to be adjusted according to the network evaluation index data to obtain target progress network data;

[0025] Calculating the total duration of the project according to the target schedule network data to obtain initial duration data, and performing constraint condition analysis on the initial duration data to obtain duration adjustment parameters;

[0026] The target progress network data is compressed or extended according to the construction period adjustment parameters to obtain adjusted progress network data, and the time parameters of the adjusted progress network data are updated to obtain construction progress data.

[0027] In the present invention, the steps of associating the structure data with the construction progress data to obtain progress-associated data, and performing simulation analysis on the progress-associated data to obtain construction scheme improvement data include:

[0028] Performing spatial partition processing on the structural data to obtain structural spatial unit data, and performing construction sequence analysis on the structural spatial unit data to obtain spatial construction sequence data;

[0029] Performing time mapping on the construction progress data according to the spatial construction sequence data to obtain time-space correlation data, and performing conflict detection on the time-space correlation data to obtain spatial conflict data;

[0030] Classifying and counting the spatial conflict data to obtain conflict type data, and generating a conflict resolution method according to the conflict type data to obtain conflict processing data;

[0031] Adjusting the spatiotemporal correlation data according to the conflict processing data to obtain preliminary progress correlation data, and performing resource allocation analysis on the preliminary progress correlation data to obtain resource demand data;

[0032] Performing a balance calculation on the resource demand data to obtain resource allocation scheme data, and revising the preliminary progress-related data according to the resource allocation scheme data to obtain target progress-related data;

[0033] Performing construction process analysis on the target progress-related data to obtain process flow data, and generating a candidate set of construction plans based on the process flow data to obtain candidate construction plan data;

[0034] A multi-objective evaluation is performed on the candidate construction scheme data to obtain scheme evaluation index data, and a recommended scheme is selected based on the scheme evaluation index data to obtain construction scheme improvement data.

[0035] In the present invention, the fiber Bragg grating sensor network is buried in the box girder structure to collect strain, temperature and vibration data in real time to obtain raw spectrum data, and the raw spectrum data is processed and classified to obtain structured construction data, including:

[0036] Performing sensor layout planning on the box girder structure to obtain sensor layout scheme data, and burying a fiber Bragg grating sensor network in the box girder structure according to the sensor layout scheme data to obtain sensor network data;

[0037] Setting signal acquisition parameters for the sensor network data to obtain acquisition parameter data, and performing signal acquisition on the fiber Bragg grating sensor network according to the acquisition parameter data to obtain original spectrum data;

[0038] Performing wavelength demodulation processing on the original spectrum data to obtain wavelength data, and performing temperature compensation calculation on the wavelength data to obtain wavelength data after temperature compensation;

[0039] Performing strain conversion calculation on the temperature-compensated wavelength data to obtain strain data, and performing vibration feature extraction on the strain data to obtain vibration feature data;

[0040] Performing time series merging on the strain data, the wavelength data after temperature compensation and the vibration characteristic data to obtain multi-dimensional time series data, and performing outlier detection on the multi-dimensional time series data to obtain cleaned time series data;

[0041] Performing data standardization processing on the cleaned time series data to obtain standardized data, and performing feature extraction on the standardized data to obtain feature vector data;

[0042] The feature vector data is subjected to data classification processing to obtain classification label data, and the classification label data is combined with the feature vector data to obtain structured construction data.

[0043] In the present invention, the step of performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data includes:

[0044] Performing time series segmentation on the structured construction data to obtain construction stage data, and performing feature extraction on the construction stage data to obtain stage feature data;

[0045] Performing pattern recognition processing on the stage characteristic data to obtain a construction process recognition result, and performing quantitative calculation on the construction progress according to the construction process recognition result to obtain preliminary actual progress data;

[0046] Performing data fusion processing on the preliminary actual progress data to obtain multi-source progress data, and performing consistency verification on the multi-source progress data to obtain verified progress data;

[0047] Performing time standardization processing on the verified progress data to obtain standardized progress data, and performing progress index calculation on the standardized progress data to obtain actual progress index data;

[0048] Synchronously processing the construction progress data according to the actual progress indicator data to obtain aligned progress data, and performing difference calculation between the aligned progress data and the actual progress indicator data to obtain original deviation data;

[0049] Performing statistical analysis on the original deviation data to obtain deviation statistical feature data, and performing threshold judgment on the deviation statistical feature data to obtain deviation type data;

[0050] The deviation type data and the original deviation data are combined and processed to obtain progress deviation data.

[0051] In the present invention, the step of performing intelligent analysis on the progress deviation data to obtain early warning data and cause diagnosis data, and performing decision analysis on the early warning data and the cause diagnosis data to obtain suggestion data includes:

[0052] Performing time series analysis on the progress deviation data through a long short-term memory network to obtain deviation trend data, and performing warning level classification according to the deviation trend data to obtain warning level data;

[0053] Performing threshold judgment on the warning level data to obtain trigger warning data, and generating warning information according to the trigger warning data to obtain warning data;

[0054] Extracting features from the progress deviation data to obtain deviation feature data, and classifying the deviation feature data using a decision tree algorithm to obtain deviation cause category data;

[0055] Retrieving a historical case database according to the deviation cause category data to obtain similar case data, and performing correlation analysis on the similar case data to obtain cause correlation data;

[0056] Performing weight calculation on the cause association data to obtain cause weight data, and sorting the deviation causes according to the cause weight data to obtain cause diagnosis data;

[0057] Performing correlation analysis on the early warning data, the cause diagnosis data and the construction plan improvement data to obtain problem feature data, and generating a decision tree based on the problem feature data to obtain decision path data;

[0058] The decision path data is processed for solution generation to obtain candidate solution data, and the candidate solution data is evaluated through a multi-criteria decision-making method to obtain recommendation data, wherein the recommendation data includes: construction schedule adjustment suggestions and resource allocation optimization solutions.

[0059] The present invention also provides a BIM-based cast-in-place box girder construction progress management system, comprising:

[0060] A processing module is used to digitally process the design drawings to obtain structural data, and to divide the structural data into work breakdown structures to obtain engineering task data;

[0061] An analysis module is used to perform a schedule analysis on the engineering task data to obtain schedule network data, and to adjust the duration of the schedule network data to obtain construction schedule data;

[0062] A processing module, used for performing correlation processing on the structural data and the construction progress data to obtain progress correlation data, and performing simulation analysis on the progress correlation data to obtain construction plan improvement data;

[0063] The acquisition module is used to bury a fiber Bragg grating sensor network in the box girder structure, collect strain, temperature and vibration data in real time, obtain raw spectral data, and process and classify the raw spectral data to obtain structured construction data;

[0064] A comparison module, used for performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data;

[0065] The decision-making module is used to perform time series analysis on the progress deviation data through a long short-term memory network to obtain early warning data and cause diagnosis data, and perform decision analysis and processing on the early warning data and the cause diagnosis data based on the construction plan improvement data to obtain recommendation data, wherein the recommendation data includes: construction progress adjustment suggestions and resource allocation optimization plans.

[0066] In the technical solution provided by the present invention, by digital processing of design drawings and work breakdown structure division, the refined management of construction tasks is realized, laying a solid foundation for the subsequent progress plan formulation and optimization. Combined with the three-dimensional visualization characteristics of the BIM model, project managers can intuitively understand the construction progress and spatial layout, greatly improving the accuracy and efficiency of decision-making. Secondly, the fiber Bragg grating sensor network is used for real-time data acquisition, which significantly improves the accuracy and real-time performance of construction process monitoring. The network can simultaneously monitor strain, temperature and vibration data, providing rich data support for comprehensive evaluation of structural status and construction quality. By wavelength demodulation and temperature compensation processing of the original spectral data, the interference of environmental factors is effectively eliminated, ensuring the reliability of monitoring data. Furthermore, this method introduces advanced data processing and analysis technologies, such as LSTM time series analysis, wavelet packet transform feature extraction and random forest classification, which greatly improves the accuracy and prediction ability of progress deviation analysis. The application of these technologies enables the system to detect potential progress problems in a timely manner and provide accurate early warning information. Through multi-level data analysis and decision support, such as association rule mining and case-based reasoning solution generation, a comprehensive and scientific decision-making basis is provided for project managers. In addition, the modular design and data-driven characteristics of this method make it adaptable and scalable, and can be flexibly adjusted according to the characteristics of different projects. In practical applications, this method significantly improves the efficiency of construction progress management, reduces human misjudgment and delays, and can control the progress deviation within 5% of the original plan on average. At the same time, by optimizing resource allocation and construction plans, the overall construction efficiency of the project is improved by 15-20%, which directly reduces the construction cost and the risk of construction delays. Finally, the application of this method also promotes the standardization and digitization of the construction process, provides a systematic way to accumulate valuable engineering experience and knowledge, and is of great significance to improving the management level of the entire construction industry. In general, the present invention not only solves the problems of insufficient real-time performance, low accuracy, and insufficient decision-making support in traditional construction progress management, but also provides new ideas and methods for modern and intelligent engineering project management, which has important practical value and demonstration effect in promoting the digital transformation of the construction industry. With the three-dimensional visualization characteristics, project managers can intuitively understand the construction progress and spatial layout, which greatly improves the accuracy and efficiency of decision-making. Secondly, the fiber Bragg grating sensor network is used for real-time data acquisition, which significantly improves the accuracy and real-time performance of construction process monitoring. The network can simultaneously monitor strain, temperature and vibration data, providing rich data support for comprehensive assessment of structural status and construction quality. By wavelength demodulation and temperature compensation processing of the original spectral data, the interference of environmental factors is effectively eliminated, ensuring the reliability of the monitoring data.Furthermore, this method introduces advanced data processing and analysis techniques, such as LSTM time series analysis, wavelet packet transform feature extraction and random forest classification, which greatly improves the accuracy and predictive ability of schedule deviation analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0068] Figure 1 It is a flow chart of a method for managing the construction progress of a cast-in-place box girder based on BIM in an embodiment of the present invention;

[0069] Figure 2 It is a schematic diagram of a BIM-based cast-in-place box girder construction progress management system in an embodiment of the present invention. DETAILED DESCRIPTION

[0070] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0072] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0073] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , Figure 1 is a flow chart of a method for managing the construction progress of a cast-in-place box girder based on BIM according to an embodiment of the present invention, such as Figure 1 As shown, the following steps are included:

[0074] S101, digitally process the design drawings to obtain structural data, and divide the structural data into work breakdown structures to obtain engineering task data;

[0075] S102, performing a progress plan analysis on the engineering task data to obtain progress network data, and adjusting the construction period of the progress network data to obtain construction progress data;

[0076] S103, performing correlation processing on the structure data and the construction progress data to obtain progress correlation data, and performing simulation analysis on the progress correlation data to obtain construction plan improvement data;

[0077] S104, embedding a fiber Bragg grating sensor network in the box girder structure to collect strain, temperature and vibration data in real time to obtain raw spectral data, and processing and classifying the raw spectral data to obtain structured construction data;

[0078] S105, performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data;

[0079] S106. Perform time series analysis on the progress deviation data through a long short-term memory network to obtain early warning data and cause diagnosis data, and perform decision analysis and processing on the early warning data and cause diagnosis data based on the construction plan improvement data to obtain recommendation data. The recommendation data includes: construction progress adjustment suggestions and resource allocation optimization plans.

[0080] It should be noted that two-dimensional drawings are converted into three-dimensional structural data through vectorization technology. This process includes the extraction of geometric information, the conversion of spatial coordinates, and the identification of structural features. Subsequently, the obtained structural data is divided into a work breakdown structure. This step divides the entire box girder structure into several work units according to the construction sequence and process requirements to form engineering task data. The engineering task data is analyzed for schedule planning, and preliminary schedule network data is constructed through task classification, priority sorting, and process dependency analysis. This network data reflects the logical relationship and time sequence between each work unit. Then, according to the actual situation of the project and resource constraints, the schedule network data is adjusted, including compressing the critical path, balancing resource load, etc., and finally executable construction schedule data is obtained.

[0081] In order to associate the BIM model with the actual construction progress, this method associates the structural data with the construction progress data. This step combines time information with spatial information to generate progress-related data. By simulating and analyzing these associated data, potential construction conflicts, resource allocation problems, etc. can be discovered, thereby obtaining construction plan improvement data and providing a basis for optimizing the construction plan.

[0082] One of the innovations of the present invention is to bury a fiber Bragg grating sensor network in the box girder structure. This sensor network can collect the strain, temperature and vibration data of the structure in real time to form raw spectral data. Through processing steps such as wavelength demodulation, temperature compensation and data standardization, the raw spectral data is converted into structured construction data, providing high-quality real-time data support for subsequent progress analysis.

[0083] By analyzing the progress of structured construction data, we can obtain the actual progress data reflecting the actual construction status. By comparing these data with the construction progress data prepared in the early stage, we can obtain the progress deviation data. These deviation data clearly reflect the difference between the actual construction progress and the planned progress.

[0084] Finally, the progress deviation data is intelligently analyzed, and warning data and cause diagnosis data are generated through time series analysis and pattern recognition technologies. These data are processed through decision analysis and ultimately form recommendation data to provide scientific decision support for project managers.

[0085] For example, in a cast-in-place box girder construction project with a span of 100 meters, three-dimensional structural data containing 1,000 structural units was generated through the digital processing of design drawings. The work breakdown structure divided these units into 50 work packages to form engineering task data. After the schedule analysis, the project duration was initially determined to be 120 days. After resource balancing and critical path compression, the final construction schedule data showed that the optimized duration was 115 days. During the construction process, the fiber optic sensor network buried in the box girder collected data every hour, generating a large amount of raw spectral data. After processing, these data showed that on the 30th day, the actual completion rate of a key node was 85%, while the planned completion rate should be 95%. The system automatically generated progress deviation data, showing that the node was 10% behind. Through the comparison and pattern recognition of historical data, the intelligent analysis system diagnosed that the main cause of the delay was insufficient concrete curing time, and gave specific suggestions for adjusting subsequent processes, including measures such as increasing human resources and optimizing process flow. These suggestions provided an important basis for the project manager to adjust the construction plan in a timely manner, and ultimately completed the project on schedule.

[0086] By executing the above steps, the refined management of construction tasks is achieved through digital processing of design drawings and division of work breakdown structure, laying a solid foundation for the subsequent progress plan formulation and optimization. Combined with the three-dimensional visualization characteristics of the BIM model, project managers can intuitively understand the construction progress and spatial layout, greatly improving the accuracy and efficiency of decision-making. Secondly, the fiber Bragg grating sensor network is used for real-time data acquisition, which significantly improves the accuracy and real-time performance of construction process monitoring. The network can simultaneously monitor strain, temperature and vibration data, providing rich data support for comprehensive evaluation of structural status and construction quality. By wavelength demodulation and temperature compensation processing of the original spectral data, the interference of environmental factors is effectively eliminated, ensuring the reliability of monitoring data. Furthermore, this method introduces advanced data processing and analysis technologies, such as LSTM time series analysis, wavelet packet transform feature extraction and random forest classification, which greatly improves the accuracy and prediction ability of progress deviation analysis. The application of these technologies enables the system to detect potential progress problems in a timely manner and give accurate early warning information. Through multi-level data analysis and decision support, such as association rule mining and case-based reasoning solution generation, a comprehensive and scientific decision-making basis is provided for project managers. In addition, the modular design and data-driven characteristics of this method make it adaptable and scalable, and can be flexibly adjusted according to the characteristics of different projects. In practical applications, this method significantly improves the efficiency of construction progress management, reduces human misjudgment and delays, and can control the progress deviation within 5% of the original plan on average. At the same time, by optimizing resource allocation and construction plans, the overall construction efficiency of the project is improved by 15-20%, which directly reduces the construction cost and the risk of construction delays. Finally, the application of this method also promotes the standardization and digitization of the construction process, provides a systematic way to accumulate valuable engineering experience and knowledge, and is of great significance to improving the management level of the entire construction industry. In general, the present invention not only solves the problems of insufficient real-time performance, low accuracy, and insufficient decision-making support in traditional construction progress management, but also provides new ideas and methods for modern and intelligent engineering project management, which has important practical value and demonstration effect in promoting the digital transformation of the construction industry. With the three-dimensional visualization characteristics, project managers can intuitively understand the construction progress and spatial layout, which greatly improves the accuracy and efficiency of decision-making. Secondly, the fiber Bragg grating sensor network is used for real-time data acquisition, which significantly improves the accuracy and real-time performance of construction process monitoring. The network can simultaneously monitor strain, temperature and vibration data, providing rich data support for comprehensive assessment of structural status and construction quality. By wavelength demodulation and temperature compensation processing of the original spectral data, the interference of environmental factors is effectively eliminated, ensuring the reliability of the monitoring data.Furthermore, this method introduces advanced data processing and analysis techniques, such as LSTM time series analysis, wavelet packet transform feature extraction and random forest classification, which greatly improves the accuracy and predictive ability of schedule deviation analysis.

[0087] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0088] (1) Vectorize the geometric information in the design drawing to obtain structural contour data, and perform three-dimensional coordinate conversion on the structural contour data to obtain three-dimensional structural data;

[0089] (2) Extracting features from the three-dimensional structure data to obtain structure feature data, and performing parametric modeling on the structure based on the structure feature data to obtain parametric structure data;

[0090] (3) Associating attribute information of parameterized structured data to obtain semantic structured data, and hierarchically classifying the semantic structured data to obtain hierarchical structured data;

[0091] (4) Analyze the construction sequence of the structure according to the hierarchical structure data to obtain construction sequence data, and divide the construction sequence data into process units to obtain process unit data;

[0092] (5) Perform resource demand analysis on process unit data to obtain resource allocation data, and perform time estimation on the process based on the resource allocation data to obtain process duration data;

[0093] (6) Construct a process network based on the process duration data and the construction sequence data to obtain preliminary engineering task data, and perform critical path analysis on the preliminary engineering task data to obtain key process data;

[0094] (7) Perform resource balancing on key process data to obtain balanced engineering task data, and generate a construction plan based on the balanced engineering task data to obtain engineering task data.

[0095] Specifically, the geometric information in the design drawings is first vectorized. Vectorization is the process of converting a raster image into a geometric shape described by a mathematical equation. Through edge detection and contour tracing algorithms, the key contour lines of the box girder structure are extracted to obtain the structural contour data. These data are represented in the form of a two-dimensional coordinate point set. Subsequently, the structural contour data is transformed into three-dimensional coordinates to convert the two-dimensional plane information into three-dimensional space information to obtain three-dimensional structural data. This step is completed by adding elevation information and considering the spatial relationship of the structure. Next, the three-dimensional structural data is feature extracted. Feature extraction is the process of identifying and extracting representative geometric and topological features in the structure. Through algorithms such as edge detection, corner point detection and shape analysis, the key features of the box girder, such as beam height, web thickness, top plate width, etc., are extracted to form structural feature data. Based on these feature data, the structure is modeled using parametric modeling technology. Parametric modeling allows the model to be quickly modified and updated by changing parameter values, which improves the flexibility and adaptability of the model and ultimately obtains parametric structural data.

[0096] The next key step is to associate attribute information with parametric structural data. This process associates non-geometric information, such as material properties, construction requirements, quality standards, etc., with geometric models. By establishing a semantic network, these attribute information are linked to structural elements to form semantic structural data. Semanticization gives data richer meaning and contextual information. Subsequently, the semantic structural data is hierarchically classified, and the structural data is divided into different levels and categories according to standards such as structural functions, construction processes or management requirements to obtain hierarchical structural data. Based on the hierarchical structural data, construction sequence analysis is performed. This step considers factors such as the dependency between structures, construction process requirements and resource constraints, determines the optimal construction sequence of each structural unit, and forms construction sequence data. The construction sequence data is further refined, and the entire construction process is divided into specific process units to obtain process unit data. Each process unit represents an independent construction task.

[0097] Perform resource demand analysis on process unit data to determine the manpower, materials, equipment and other resources required for each process and obtain resource allocation data. According to resource allocation data and historical experience data, estimate the time for each process, calculate the expected duration of the process and form process duration data. Use process duration data and construction sequence data to build a process network. The process network is a graphical representation that shows the logical relationship and time dependency between processes and forms preliminary engineering task data. Perform critical path analysis on these data to identify the key processes that have a decisive impact on the total project duration and obtain key process data. Critical path analysis uses a network diagram algorithm to calculate parameters such as the earliest start time, the latest start time, and the total float time of each process.

[0098] Finally, the key process data is resource balanced. Resource balancing aims to optimize resource allocation and avoid excessive concentration or idleness of resources. By adjusting the start time of non-critical processes, resources can be used more evenly without extending the total construction period, and balanced engineering task data can be obtained. Based on this data, a detailed construction plan is generated, including the start time, end time, required resources and other information of each process, and finally complete engineering task data is formed.

[0099] For example, in a cast-in-place box girder project with a span of 80 meters, 10,000 key coordinate points were extracted from the CAD drawings through vectorization processing to form the structural contour data. After the three-dimensional coordinate conversion, three-dimensional structural data containing 30,000 spatial points were generated. The feature extraction process identified 50 key structural features, such as the top plate thickness of 1.5 meters and the web thickness of 0.5 meters. After parametric modeling, a structural model containing 100 adjustable parameters was formed. Attribute information association added 200 non-geometric attributes to the model, such as concrete strength grade C50, steel bar model HRB400, etc. Hierarchical classification divides the structure into five main levels, including foundation, main beam, diaphragm, prestressed system and auxiliary structure. Construction sequence analysis generated construction sequence data containing 30 main construction steps. The process division was further refined into 150 specific process units. Resource demand analysis showed that the project peak required 50 workers, 3 concrete pump trucks, 2 tower cranes, etc. The time estimation results show that the total duration of the processes on the critical path is 120 days. Critical path analysis identified 7 critical processes, accounting for 4.67% of the total number of processes. After resource balancing, the start time of non-critical processes was adjusted, which increased resource utilization by 15% while keeping the total construction period unchanged. The final construction plan specifies the execution time, resource allocation and quality control requirements of each process in detail.

[0100] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0101] (1) Classify the engineering task data to obtain classified task data, and prioritize the classified task data to obtain task priority data;

[0102] (2) Performing process dependency analysis on the tasks according to the task priority data to obtain process dependency relationship data, and performing critical path calculation on the process dependency relationship data to obtain critical path data;

[0103] (3) Estimate the time parameters of the critical path data to obtain the process duration data, and construct a schedule network diagram based on the process duration data to obtain preliminary schedule network data;

[0104] (4) Perform resource load analysis on the preliminary schedule network data to obtain resource load data, and balance the schedule network based on the resource load data to obtain candidate schedule network data;

[0105] (5) Perform multi-objective evaluation on the candidate progress network data to obtain network evaluation index data, and select the progress network to be adjusted based on the network evaluation index data to obtain the target progress network data;

[0106] (6) Calculate the total duration of the project based on the target schedule network data to obtain the initial duration data, and perform constraint analysis on the initial duration data to obtain duration adjustment parameters;

[0107] (7) The target progress network data is compressed or extended according to the construction period adjustment parameters to obtain the adjusted progress network data, and the time parameters of the adjusted progress network data are updated to obtain the construction progress data.

[0108] Specifically, the task classification process uses the hierarchical analysis method to divide the engineering tasks into different categories according to factors such as structural parts, construction technology and resource requirements, such as foundation engineering, main structure, and auxiliary facilities, to form classified task data. Subsequently, the classified task data is prioritized, and the weighted scoring method is used to assign priority scores to each task, considering factors such as construction logic, technical requirements and resource constraints, to obtain task priority data. According to the task priority data, process dependency analysis is performed. This step uses the precedence diagram method to identify the logical relationship between tasks, including four basic relationships: finish-start (FS), start-start (SS), finish-finish (FF) and start-finish (SF), to generate process dependency data. The critical path calculation is performed on the process dependency data, and the earliest start time, latest start time, total float time and free float time of each process are determined by combining forward calculation and reverse calculation, so as to identify the critical path that has a decisive impact on the total project duration and obtain critical path data.

[0109] The time parameters of the critical path data are estimated using the three-point estimation method, that is, the most optimistic time, the most likely time, and the most pessimistic time of each process are estimated, and then the weighted average is calculated as the expected duration of the process to form the process duration data. Based on the process duration data, the schedule network diagram is constructed using the arrow diagram method or the precedence diagram method to show the logical relationship and time arrangement between the processes, and the preliminary schedule network data is obtained. The preliminary schedule network data is analyzed for resource load, and the demand for various resources in each time period is calculated to obtain resource load data. The resource load analysis uses the resource histogram method to intuitively display the peak and valley values ​​of resource use. According to the resource load data, the schedule network is balanced and adjusted, with the goal of making resource use more balanced without extending the total construction period. The adjustment methods include moving, splitting or parallel processing of non-critical processes, and finally obtaining candidate schedule network data.

[0110] A multi-objective evaluation is conducted on the candidate schedule network data, taking into account multiple objectives such as duration, cost, quality and resource utilization efficiency. The comprehensive scores of each candidate scheme are calculated by combining the analytic hierarchy process and the fuzzy comprehensive evaluation method to form network evaluation index data. According to the network evaluation index data, the scheme with the highest comprehensive score is selected as the schedule network to be adjusted to obtain the target schedule network data. The total duration of the project is calculated based on the target schedule network data to obtain the initial duration data. The constraint analysis of the initial duration data is conducted, taking into account factors such as the contract duration, seasonal construction restrictions, and resource supply cycle, to determine the feasible range and adjustment range of the duration adjustment and obtain the duration adjustment parameters.

[0111] Finally, the target schedule network data is compressed or extended according to the construction period adjustment parameters. Construction period compression adopts the rush work method, which mainly targets the processes on the critical path and shortens the construction period by increasing resource input or adjusting the construction process. Construction period extension mainly targets non-critical processes and appropriately extends the duration of the process within the total floating time to optimize resource utilization. After processing, the adjusted schedule network data is obtained, and then the time parameters are updated, including recalculating the start time, end time and floating time of each process, and finally obtaining the construction progress data.

[0112] For example, in a cast-in-place box girder project with a span of 100 meters, the initial engineering task data contains 200 specific tasks. Through task classification, these tasks are divided into four categories: basic engineering, main structure, prestressed system and auxiliary facilities. After priority sorting, the box girder casting task in the main structure received the highest priority score of 9.5 (out of 10 points). Process dependency analysis shows that the box girder casting process has a finish-start (FS) relationship with its preceding steel bar binding and formwork installation processes. The critical path calculation results show that the project contains 35 critical processes, accounting for 17.5% of the total number of processes. In the time parameter estimation, taking the box girder casting process as an example, its most optimistic time is 5 days, the most likely time is 7 days, and the most pessimistic time is 10 days. The expected duration is 7.2 days calculated by the three-point estimation method. Resource load analysis shows that during the peak period of the project, 3 concrete pump trucks, 2 tower cranes and 80 construction workers are required every day. The multi-objective evaluation uses five evaluation indicators, including the completion rate of the construction period (weight 0.3), the cost control rate (weight 0.25), the quality compliance rate (weight 0.2), the resource utilization rate (weight 0.15) and the safety accident rate (weight 0.1). The evaluation results show that the comprehensive score of the optimal solution is 0.89 (out of 1 point).

[0113] The initial construction period was calculated to be 180 days. Considering that the construction period required by the contract was 175 days, it was necessary to compress the construction period by 5 days. By rushing the 5 processes on the critical path and compressing each process by 1 day, the total construction period was successfully shortened to 175 days. At the same time, the start time of non-critical processes was appropriately postponed, which increased the resource utilization rate from the original 75% to 85%.

[0114] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0115] (1) Perform spatial partitioning on the structural data to obtain structural spatial unit data, and perform construction sequence analysis on the structural spatial unit data to obtain spatial construction sequence data;

[0116] (2) Time mapping the construction progress data according to the spatial construction sequence data to obtain spatiotemporal correlation data, and conflict detection is performed on the spatiotemporal correlation data to obtain spatial conflict data;

[0117] (3) Classify and count the spatial conflict data to obtain conflict type data, and generate conflict resolution solutions based on the conflict type data to obtain conflict resolution data;

[0118] (4) Adjust the spatiotemporal correlation data according to the conflict resolution data to obtain preliminary progress correlation data, and perform resource allocation analysis on the preliminary progress correlation data to obtain resource demand data;

[0119] (5) Balance the resource demand data to obtain resource allocation plan data, and modify the preliminary progress-related data based on the resource allocation plan data to obtain target progress-related data;

[0120] (6) Perform construction process analysis on the target progress-related data to obtain process flow data, and generate a candidate set of construction plans based on the process flow data to obtain candidate construction plan data;

[0121] (7) Perform a multi-objective evaluation on the candidate construction scheme data to obtain scheme evaluation index data, and select a recommended scheme based on the scheme evaluation index data to obtain construction scheme improvement data.

[0122] Specifically, the structural data is first processed by spatial partitioning. This process uses the octree partitioning algorithm to divide the entire box girder structure into multiple spatial units. The octree partitioning algorithm recursively divides the three-dimensional space into eight subspaces until the preset accuracy or depth is reached. The mathematical expression of spatial partitioning is as follows:

[0123] ;

[0124] in, Represents the entire space, represents the i-th subspace, is the weight coefficient. In this way, the structural space unit data is obtained. Subsequently, the construction sequence analysis of the structural space unit data is carried out, and the topological sorting algorithm is used to generate the spatial construction sequence data by considering the structural dependency and construction process requirements. According to the spatial construction sequence data, the construction progress data is time-mapped. This step uses the spatiotemporal mapping function , where s represents the spatial coordinate and t represents the time coordinate. The mapping function can be expressed as:

[0125] ;

[0126] in, and are the spatial and temporal basis functions, respectively, is the weight coefficient, n is the number of basis functions. Through this mapping, spatiotemporal correlation data is obtained. Conflict detection is performed on spatiotemporal correlation data. The spatial interference detection algorithm is used to check the overlap of different construction activities in time and space to obtain spatial conflict data. Spatial conflict data are classified and counted. Cluster analysis methods such as K-means algorithm are used to classify conflicts according to types, severity and other characteristics to obtain conflict type data. According to the conflict type data, a rule-based expert system is used to generate conflict solutions to form conflict processing data.

[0127] According to the conflict handling data, the spatiotemporal correlation data is adjusted. The adjustment process uses heuristic algorithms, such as simulated annealing algorithm, to fine-tune the spatiotemporal position of construction activities while ensuring the construction logic, and obtain preliminary progress correlation data. The preliminary progress correlation data is analyzed for resource allocation, and the linear programming model is used to calculate the demand for various resources in each time period to obtain resource demand data.

[0128] The resource demand data is balanced and a resource balancing algorithm is used, such as the Resource Constrained Project Scheduling Algorithm (RCPSP), with the goal of minimizing the volatility of resource usage. The objective function of the balance calculation can be expressed as:

[0129] ;

[0130] in, Representation Resources r In time t The amount of use, Representation Resources r The average usage of , T is the total number of time periods, and R is the number of resource types. By solving this optimization problem, the resource allocation plan data is obtained. According to the resource allocation plan data, the preliminary progress-related data is corrected to obtain the target progress-related data.

[0131] The construction process analysis is performed on the target progress related data, and the logical relationship between the construction processes is expressed by the process flow chart (PERT chart) to obtain the process flow data. Based on the process flow data, a combination optimization algorithm, such as a genetic algorithm, is used to generate multiple feasible construction plans to form candidate construction plan data.

[0132] Finally, a multi-objective evaluation is conducted on the candidate construction scheme data. The evaluation adopts a combination of the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation, taking into account multiple objectives such as construction period, cost, quality, and safety. The comprehensive score of the evaluation index is calculated as follows:

[0133] ;

[0134] in, For the The weight of the evaluation index, For the Under the indicator The degree of membership of the level, is the fuzzy evaluation matrix element. The scheme evaluation index data is obtained through calculation, and the optimal scheme is selected according to the evaluation results to obtain the construction scheme improvement data.

[0135] For example, in a 120-meter-long cast-in-place box girder project, the spatial partitioning process divides the entire structure into 480 spatial units, each of which is 2m×2m×1m in size. The construction sequence analysis generates spatial construction sequence data containing 60 major construction steps. In the spatiotemporal mapping process, 20 spatial basis functions and 15 temporal basis functions are used for mapping. Conflict detection found 35 potential spatial conflicts, including 10 equipment collision conflicts, 15 working surface overlap conflicts, and 10 insufficient safety distance conflicts. Resource demand analysis shows that 4 concrete pump trucks, 3 tower cranes, and 100 construction workers are required every day during the peak period of the project. Resource balance calculation reduces the standard deviation of resource use from the initial 28.5% to 12.3%. In the multi-objective evaluation, five evaluation indicators are set: construction period (weight 0.3), cost (weight 0.25), quality (weight 0.2), safety (weight 0.15), and environmental impact (weight 0.1). The optimal solution finally selected achieved a comprehensive score of 0.92 (full score 1), which was 15.6% higher than the initial solution. This optimized construction solution shortened the total construction period of the project from the initial 180 days to 165 days, and increased resource utilization from 73% to 88%, providing a scientific and efficient guidance solution for project management.

[0136] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0137] (1) Planning the sensor layout of the box girder structure to obtain sensor layout plan data, and burying a fiber Bragg grating sensor network in the box girder structure according to the sensor layout plan data to obtain sensor network data;

[0138] (2) Setting signal acquisition parameters for sensor network data to obtain acquisition parameter data, and acquiring signals from the fiber Bragg grating sensor network based on the acquisition parameter data to obtain original spectral data;

[0139] (3) Performing wavelength demodulation processing on the original spectrum data to obtain wavelength data, and performing temperature compensation calculation on the wavelength data to obtain wavelength data after temperature compensation;

[0140] (4) performing strain conversion calculation on the wavelength data after temperature compensation to obtain strain data, and performing vibration feature extraction on the strain data to obtain vibration feature data;

[0141] (5) Merging the strain data, wavelength data after temperature compensation, and vibration characteristic data in time series to obtain multidimensional time series data, and performing outlier detection on the multidimensional time series data to obtain cleaned time series data;

[0142] (6) Perform data standardization on the cleaned time series data to obtain standardized data, and perform feature extraction on the standardized data to obtain feature vector data;

[0143] (7) Performing data classification processing on the feature vector data to obtain classification label data, and combining the classification label data with the feature vector data to obtain structured construction data.

[0144] Specifically, in the BIM-based cast-in-place box girder construction progress management method, the sensor layout of the box girder structure is first planned. This process uses a genetic algorithm to optimize the sensor position, and the objective function is to maximize the monitoring coverage and minimize the number of sensors. The optimization problem can be expressed as:

[0145] ;

[0146] in, To monitor coverage, is the number of sensors, is the maximum number of sensors allowed, and is the weight coefficient. By solving this optimization problem, the sensor layout scheme data is obtained. According to the scheme, a fiber Bragg grating sensor network is buried in the box girder structure to form sensor network data.

[0147] The signal acquisition parameters of the sensor network data are set, including sampling frequency, wavelength range, sensitivity, etc., to obtain acquisition parameter data. According to these parameters, the fiber Bragg grating sensor network is used to collect signals and obtain the original spectral data. The spectral data can be expressed as:

[0148] ;

[0149] in, is the reflected light intensity, is the wavelength of light, is the incident light intensity, is the modulation depth, is the effective refractive index of the optical fiber, is the grating period.

[0150] The original spectrum data is subjected to wavelength demodulation processing, and the center wavelength positioning algorithm, such as Gaussian fitting method, is used to obtain wavelength data. The wavelength data is then subjected to temperature compensation calculation, and the compensation formula is:

[0151] ;

[0152] in, is the Bragg wavelength shift, is the strain sensitivity coefficient, For strain, is the temperature sensitivity coefficient, is the temperature change. The wavelength data after temperature compensation is obtained through this calculation.

[0153] Perform strain conversion calculation on the wavelength data after temperature compensation to obtain strain data. The conversion formula is:

[0154] ;

[0155] in, is the initial Bragg wavelength, is the photoelastic coefficient, is the coefficient of thermal expansion, is the thermo-optical coefficient. Vibration feature extraction is performed on the strain data, and the fast Fourier transform (FFT) algorithm is used to obtain vibration feature data. The strain data, wavelength data after temperature compensation, and vibration feature data are merged in time series to form multi-dimensional time series data. Outlier detection is performed on the multi-dimensional time series data, and the local outlier factor (LOF) algorithm is used to identify and eliminate abnormal data points to obtain cleaned time series data.

[0156] The cleaned time series data is normalized using the Z-score normalization method, and the formula is:

[0157] ;

[0158] in, is the standardized value, is the original value, is the mean, is the standard deviation. Standardized data is obtained after standardization. Feature extraction is performed on the standardized data, and the principal component analysis (PCA) algorithm is used to obtain feature vector data. Finally, data classification is performed on the feature vector data, and the support vector machine (SVM) algorithm is used to divide the data into different construction stages or status categories to obtain classified label data. The classified label data is combined with the feature vector data to finally form structured construction data.

[0159] For example, in a cast-in-place box girder project with a span of 150 meters, during the sensor layout planning process, the genetic algorithm optimization was used to determine the solution of burying 80 fiber Bragg grating sensor points in the box girder, and the monitoring coverage rate reached 95%. The signal acquisition parameters set the sampling frequency to 100Hz and the wavelength range to 1510nm-1590nm. During the wavelength demodulation process, the Gaussian fitting method was used, and the fitting accuracy reached 0.1pm. The temperature compensation calculation uses a temperature sensitivity coefficient of 1.2pm / ℃ and a strain sensitivity coefficient of 0.78pm / με.

[0160] In the strain conversion calculation, the initial Bragg wavelength is 1550nm, the photoelastic coefficient is 0.22, and the thermal expansion coefficient is 0.55× / ℃, the thermo-optical coefficient is 8.6× / ℃. Vibration feature extraction uses 1024-point FFT to obtain spectrum data in the range of 0-50Hz. During the outlier detection process, the LOF threshold is set to 2.5, and 0.5% of the total data volume is identified as outliers. After data standardization, principal component analysis retains the top 10 principal components that explain 95% of the variance. The SVM classifier uses the RBF kernel function to divide the data into 5 construction stage categories with a classification accuracy of 92%. The structured construction data finally generated contains 10 feature dimensions and 1 category label, providing high-quality data support for subsequent construction progress analysis and decision-making.

[0161] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0162] (1) Segment the structured construction data into time series to obtain construction phase data, and extract features from the construction phase data to obtain phase feature data;

[0163] (2) Perform pattern recognition processing on the stage characteristic data to obtain the construction process identification results, and quantify the construction progress based on the construction process identification results to obtain preliminary actual progress data;

[0164] (3) Perform data fusion processing on the preliminary actual progress data to obtain multi-source progress data, and perform consistency check on the multi-source progress data to obtain verified progress data;

[0165] (4) Performing time standardization on the verified progress data to obtain standardized progress data, and performing progress index calculation on the standardized progress data to obtain actual progress index data;

[0166] (5) Synchronize the construction progress data according to the actual progress index data to obtain the aligned progress data, and calculate the difference between the aligned progress data and the actual progress index data to obtain the original deviation data;

[0167] (6) Performing statistical analysis on the original deviation data to obtain deviation statistical feature data, and performing threshold judgment on the deviation statistical feature data to obtain deviation type data;

[0168] (7) The deviation type data and the original deviation data are combined to obtain the progress deviation data.

[0169] Specifically, in the BIM-based cast-in-place box girder construction progress management method, the structured construction data is first segmented into time series. This process uses the dynamic time warping (DTW) algorithm to divide the continuous time series data into different construction stages. The DTW algorithm calculates the similarity between time series and finds the best alignment to achieve data segmentation. After segmentation, the construction stage data is obtained, and then the features of these data are extracted using the wavelet transform method. The wavelet transform can be expressed as:

[0170] ;

[0171] in, is the original signal, is the wavelet function, is the scale parameter, is the translation parameter. Through wavelet transform, the characteristics of each construction stage are extracted to obtain the stage characteristic data.

[0172] The convolutional neural network (CNN) algorithm is used to perform pattern recognition on the stage feature data. The output layer activation function of CNN uses the softmax function, which is defined as:

[0173] ;

[0174] in, For the The probability of the categories, is the input feature, For the The original output scores of each category are obtained. The specific construction process is identified through CNN to obtain the construction process identification results. Based on these results, the construction progress is quantified and the preliminary actual progress data is obtained.

[0175] The preliminary actual progress data is fused using the Dempster-Shafer evidence theory. The fusion rule can be expressed as:

[0176] ;

[0177] in, is the basic probability distribution after fusion, and are the basic probability distributions of the two evidence sources. In this way, the progress data from different sources are integrated to obtain multi-source progress data. The consistency of multi-source progress data is checked, and the verified progress data is obtained using Kendall's W coefficient method.

[0178] The time standardization of the verified progress data is performed using the minimum-maximum normalization method:

[0179] ;

[0180] in, is the standardized value, is the original value, and After standardization, the standardized progress data is obtained, and the progress indicators are calculated, such as completion percentage, progress deviation rate, etc., to obtain the actual progress indicator data.

[0181] The construction progress data is synchronously processed according to the actual progress index data, and the dynamic time warping (DTW) algorithm is used to obtain the aligned progress data. The difference between the aligned progress data and the actual progress index data is calculated to obtain the original deviation data. The difference calculation formula is:

[0182] ;

[0183] in, For the The deviation of time points, To plan progress, For actual progress.

[0184] The original deviation data is statistically analyzed and processed to calculate the mean, standard deviation, skewness, kurtosis and other statistical characteristics to obtain the deviation statistical characteristic data. The deviation statistical characteristic data is threshold judged and the fuzzy C-means clustering algorithm is used to classify the deviation into different types, such as slight deviation, moderate deviation and severe deviation, to obtain the deviation type data.

[0185] Finally, the deviation type data and the original deviation data are combined and processed to form a comprehensive data structure containing deviation value and type information to obtain progress deviation data.

[0186] For example, in a 200-meter-long cast-in-place box girder project, the structured construction data includes 100 days of continuous monitoring data, with 24 hours of data collected every day. Time series segmentation divides the 100-day data into five main construction stages, including foundation construction, main beam pouring, prestressed construction, auxiliary structure construction, and maintenance. Wavelet transform feature extraction uses db4 wavelet with 5 decomposition levels to extract 10 characteristic parameters such as energy and entropy for each stage.

[0187] CNN pattern recognition uses a 5-layer convolutional network structure, with a 10×100 feature matrix as input and 20 construction process categories as output. The recognition accuracy rate reaches 95%. In the data fusion processing, three data sources, namely fiber optic sensing, image recognition and manual inspection, are combined, and the confidence after Dempster-Shafer fusion is increased by 15%. Time standardization maps 100 days of data to the range of 0-1. The progress indicator calculation shows that the overall completion rate of the project is 85%, of which the completion rate of the foundation construction and main beam casting phases is 100%, the completion rate of prestressed construction is 90%, the completion rate of auxiliary structure construction is 70%, and the completion rate of the maintenance phase is 65%. After DTW synchronization processing, the average progress deviation calculated is -3.5 days, and the standard deviation is 2.1 days. Fuzzy C-means clustering divides the deviation into three categories: within ±2 days is a slight deviation, ±2 to ±5 days is a moderate deviation, and more than ±5 days is a severe deviation. The final generated progress deviation data showed that 70% of the processes were in the range of slight deviation, 25% were in the range of moderate deviation, and 5% were in the range of severe deviation, providing project managers with a clear overview of the progress status.

[0188] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0189] (1) Perform time series analysis on the progress deviation data through the long short-term memory network to obtain deviation trend data, and divide the warning level according to the deviation trend data to obtain warning level data;

[0190] (2) Performing threshold judgment on the warning level data to obtain trigger warning data, and generating warning information based on the trigger warning data to obtain warning data;

[0191] (3) Extract features from the progress deviation data to obtain deviation feature data, and classify the deviation feature data using a decision tree algorithm to obtain deviation cause category data;

[0192] (4) Search the historical case database based on the deviation cause category data to obtain similar case data, and perform correlation analysis on the similar case data to obtain cause correlation data;

[0193] (5) Calculate the weight of the cause-related data to obtain the cause weight data, and sort the deviation causes according to the cause weight data to obtain the cause diagnosis data;

[0194] (6) Perform correlation analysis on the early warning data, cause diagnosis data, and construction plan improvement data to obtain problem feature data, and generate a decision tree based on the problem feature data to obtain decision path data;

[0195] (7) Performing scheme generation processing on the decision path data to obtain candidate scheme data, and evaluating the candidate scheme data through a multi-criteria decision-making method to obtain recommendation data, wherein the recommendation data includes: construction schedule adjustment suggestions and resource allocation optimization schemes.

[0196] Specifically, the long short-term memory network (LSTM) algorithm is used to perform time series analysis on the progress deviation data. LSTM selectively memorizes and forgets historical information through the mechanism of input gate, forget gate and output gate, thereby obtaining deviation trend data. According to the change rate and amplitude of the deviation trend data, the warning level is divided into three levels: mild, moderate and severe, and the warning level data is obtained.

[0197] The threshold value of the warning level data is judged by using fuzzy logic control method. This method accurately judges the warning level by setting a fuzzy rule set, such as "if the deviation trend rises quickly and lasts for a long time, a high-level warning is triggered" to obtain the triggered warning data. According to the triggered warning data, warning information containing information such as warning level, triggering reason and impact range is generated to form warning data.

[0198] The feature extraction of progress deviation data is performed using the wavelet packet transform (WPT) method. WPT can perform multi-resolution analysis of signals in both the time and frequency domains, extract key features in the deviation data, such as trends, periodicity, and sudden changes, and obtain deviation feature data. Subsequently, the deviation feature data is classified using the decision tree algorithm, which improves the accuracy and interpretability of the classification, and finally obtains the category data of the deviation cause.

[0199] According to the deviation cause category data, the historical case library is retrieved and a content-based recommendation algorithm is used. The algorithm calculates the similarity between the current deviation cause and the historical case, such as cosine similarity, to find the most similar case and obtain similar case data. The similar case data is analyzed for correlation and the frequent pattern mining algorithm FP-Growth is used to obtain the cause correlation data.

[0200] The weight calculation of the cause-related data is carried out, and an improved fuzzy comprehensive evaluation method is adopted. This method combines expert scoring and data statistical analysis to quantitatively evaluate the importance of each cause factor and obtain the cause weight data. According to the cause weight data, the deviation causes are sorted in descending order to generate cause diagnosis data including the main cause, impact degree and possible consequences.

[0201] The association analysis of early warning data, cause diagnosis data and construction plan improvement data is carried out using the improved association rule mining algorithm Apriori. The algorithm gradually generates frequent item sets in an iterative manner, and selects strong association rules based on support and confidence thresholds to obtain problem feature data. The specific process is as follows:

[0202] Data preprocessing: First, the early warning data (such as warning level, triggering cause), cause diagnosis data (such as main deviation cause, impact degree) and construction plan improvement data (such as improvement measures taken previously, effect evaluation) are integrated and encoded to form a unified transaction data set.

[0203] Generate frequent item sets: Use the FP-Growth algorithm, an improved version of the Apriori algorithm, to build FP trees and conditional FP trees to efficiently discover frequent item sets. Set the minimum support threshold to 0.3, which means that the frequency of a certain combination in all data is not less than 30%.

[0204] Generate association rules: Generate association rules based on frequent item sets. Set the minimum confidence threshold to 0.7, which means that when the antecedent appears, the probability of the consequent appearing is not less than 70%.

[0205] Rule screening and interpretation: Screen and interpret the generated association rules. For example, the following rules are found:

[0206] {Advanced warning, improper resource allocation} → {Increase resource investment} (support=0.35, confidence=0.85)

[0207] {Moderate warning, weather impact, temporary protective measures taken before} → {Optimize construction technology} (support=0.32, confidence=0.78).

[0208] Problem feature extraction: Extract problem features based on the selected strong association rules. For example:

[0209] There is a strong correlation between resource allocation and advanced warnings, and they usually need to be resolved by increasing resource input. Even if temporary protective measures are taken, the impact of weather under moderate warnings still needs to be further improved by optimizing construction processes.

[0210] Data fusion: The extracted problem features are integrated with the original warning data, cause diagnosis data and construction plan improvement data to form comprehensive problem feature data.

[0211] For example, in the aforementioned 220-meter span cast-in-place box girder project, a key problem feature obtained by correlation analysis is that there is a strong correlation between improper resource allocation (mainly insufficient concrete supply) under moderate warning and the improvement measures previously taken to increase the number of suppliers (support 0.33, confidence 0.81). This indicates that simply increasing the number of suppliers may not be enough to solve the problem, and a more comprehensive resource allocation optimization solution needs to be considered.

[0212] Furthermore, according to the problem feature data, the C4.5 decision tree algorithm is used to generate a decision tree, and finally the decision path data is obtained.

[0213] Finally, the decision path data is processed for solution generation using the case-based reasoning (CBR) method. CBR extracts solutions from historical successful cases and makes adaptive adjustments through four steps: retrieval, reuse, modification, and retention, to obtain candidate solution data. The candidate solution data is evaluated by multi-criteria through the analytic hierarchy process (AHP), taking into account factors such as implementation difficulty, resource requirements, and expected effects, and finally obtaining priority-ranked recommended data.

[0214] The recommended data obtained mainly include two aspects: construction schedule adjustment suggestions and resource allocation optimization plan. Construction schedule adjustment suggestions include time adjustment of key processes, process sequence optimization, etc.; resource allocation optimization plan includes specific measures such as human resource deployment, equipment use plan adjustment, and material supply chain optimization.

[0215] For example, in a cast-in-place box girder project with a span of 220 meters, LSTM time series analysis showed that the recent progress deviation showed an accelerating growth trend, and it was predicted that the deviation would increase by 3.8 days in the next 7 days. Fuzzy logic control judgment triggered a moderate warning, and the generated warning information included "the main beam pouring progress was delayed, affecting the subsequent prestressing tensioning process." The decision tree classification results showed that the main causes of the deviation were improper resource allocation (confidence 0.85) and weather influence (confidence 0.72). The content-based recommendation algorithm matched 8 similar cases from the historical case library, and the FP-Growth algorithm found that improper resource allocation was strongly associated with insufficient concrete supply (support 0.75, confidence 0.92).

[0216] The final generated recommendation data includes: Construction schedule adjustment suggestions: adjust the time schedule of the main beam pouring process, extend the original 7-day construction period to 9 days to ensure the quality of concrete; adjust the start time of the subsequent prestressing process accordingly. Resource allocation optimization plan: add 5 concrete pump trucks, optimize the selection and supply plan of concrete suppliers; adjust the labor allocation, increase 20% of workers during the main beam pouring period; formulate a detailed weather response plan, including measures such as setting up temporary rain shelters. These suggestions are expected to reduce the progress deviation to within 1.5 days within 15 days, while ensuring construction quality and safety.

[0217] The embodiment of the present invention also provides a BIM-based cast-in-place box girder construction progress management system, such as Figure 2 As shown, the BIM-based cast-in-place box girder construction progress management system specifically includes:

[0218] The processing module 201 is used to digitally process the design drawings to obtain structural data, and divide the structural data into work breakdown structures to obtain engineering task data;

[0219] The analysis module 202 is used to perform a schedule analysis on the engineering task data to obtain schedule network data, and to adjust the duration of the schedule network data to obtain construction schedule data;

[0220] The processing module 203 is used to associate the structure data with the construction progress data to obtain progress-associated data, and to perform simulation analysis on the progress-associated data to obtain construction plan improvement data;

[0221] The acquisition module 204 is used to bury a fiber Bragg grating sensor network in the box girder structure, collect strain, temperature and vibration data in real time, obtain raw spectrum data, and process and classify the raw spectrum data to obtain structured construction data;

[0222] The comparison module 205 is used to perform progress analysis on the structured construction data to obtain actual progress data, and compare the actual progress data with the construction progress data to obtain progress deviation data;

[0223] The decision module 206 is used to perform time series analysis on the progress deviation data through a long short-term memory network to obtain early warning data and cause diagnosis data, and perform decision analysis and processing on the early warning data and the cause diagnosis data based on the construction plan improvement data to obtain recommendation data, which includes: construction progress adjustment suggestions and resource allocation optimization plans.

[0224] Through the collaborative work of the above modules, the refined management of construction tasks is achieved by digital processing of design drawings and division of work breakdown structure, laying a solid foundation for the subsequent progress plan formulation and optimization. Combined with the three-dimensional visualization characteristics of the BIM model, project managers can intuitively understand the construction progress and spatial layout, greatly improving the accuracy and efficiency of decision-making. Secondly, the fiber Bragg grating sensor network is used for real-time data acquisition, which significantly improves the accuracy and real-time performance of construction process monitoring. The network can simultaneously monitor strain, temperature and vibration data, providing rich data support for comprehensive evaluation of structural status and construction quality. By wavelength demodulation and temperature compensation processing of the original spectral data, the interference of environmental factors is effectively eliminated, ensuring the reliability of monitoring data. Furthermore, this method introduces advanced data processing and analysis technologies, such as LSTM time series analysis, wavelet packet transform feature extraction and random forest classification, which greatly improves the accuracy and prediction ability of progress deviation analysis. The application of these technologies enables the system to detect potential progress problems in a timely manner and give accurate early warning information. Through multi-level data analysis and decision support, such as association rule mining and case-based reasoning solution generation, a comprehensive and scientific decision-making basis is provided for project managers. In addition, the modular design and data-driven characteristics of this method make it adaptable and scalable, and can be flexibly adjusted according to the characteristics of different projects. In practical applications, this method significantly improves the efficiency of construction progress management, reduces human misjudgment and delays, and can control the progress deviation within 5% of the original plan on average. At the same time, by optimizing resource allocation and construction plans, the overall construction efficiency of the project is improved by 15-20%, which directly reduces the construction cost and the risk of construction delays. Finally, the application of this method also promotes the standardization and digitization of the construction process, provides a systematic way to accumulate valuable engineering experience and knowledge, and is of great significance to improving the management level of the entire construction industry. In general, the present invention not only solves the problems of insufficient real-time performance, low accuracy, and insufficient decision-making support in traditional construction progress management, but also provides new ideas and methods for modern and intelligent engineering project management, which has important practical value and demonstration effect in promoting the digital transformation of the construction industry. With the three-dimensional visualization characteristics, project managers can intuitively understand the construction progress and spatial layout, which greatly improves the accuracy and efficiency of decision-making. Secondly, the fiber Bragg grating sensor network is used for real-time data acquisition, which significantly improves the accuracy and real-time performance of construction process monitoring. The network can simultaneously monitor strain, temperature and vibration data, providing rich data support for comprehensive assessment of structural status and construction quality. By wavelength demodulation and temperature compensation processing of the original spectral data, the interference of environmental factors is effectively eliminated, ensuring the reliability of the monitoring data.Furthermore, this method introduces advanced data processing and analysis techniques, such as LSTM time series analysis, wavelet packet transform feature extraction and random forest classification, which greatly improves the accuracy and predictive ability of schedule deviation analysis.

[0225] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the embodiments, a person skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for managing the construction progress of a cast-in-place box girder based on BIM, characterized in that: include: Digitally process the design drawings to obtain structural data, and divide the structural data into work breakdown structures to obtain engineering task data; Performing a schedule analysis on the engineering task data to obtain schedule network data, and adjusting the construction period of the schedule network data to obtain construction schedule data; The structural data and the construction progress data are associated with each other to obtain progress associated data, and the progress associated data is simulated and analyzed to obtain construction plan improvement data, including: spatially partitioning the structural data to obtain structural spatial unit data, and performing construction sequence analysis on the structural spatial unit data to obtain spatial construction sequence data; time mapping the construction progress data according to the spatial construction sequence data to obtain spatiotemporal associated data, and performing conflict detection on the spatiotemporal associated data to obtain spatial conflict data; classifying and counting the spatial conflict data to obtain conflict type data, and generating a conflict resolution method according to the conflict type data to obtain conflict resolution data; and generating conflict resolution data according to the conflict resolution data. The time-space correlation data is adjusted according to the processing data to obtain preliminary progress correlation data, and the preliminary progress correlation data is analyzed for resource allocation to obtain resource demand data; the resource demand data is balanced and calculated to obtain resource allocation scheme data, and the preliminary progress correlation data is corrected according to the resource allocation scheme data to obtain target progress correlation data; the target progress correlation data is analyzed for construction process to obtain process flow data, and a candidate set of construction schemes is generated according to the process flow data to obtain candidate construction scheme data; the candidate construction scheme data is evaluated for multiple objectives to obtain scheme evaluation index data, and a recommended scheme is selected according to the scheme evaluation index data to obtain construction scheme improvement data; Among them, the structural data is processed by spatial partitioning; the octree partitioning algorithm is used to divide the entire box girder structure into multiple spatial units; the octree partitioning algorithm recursively divides the three-dimensional space into eight subspaces until the preset accuracy or depth is reached; the mathematical expression of spatial partitioning is as follows: ； in, Represents the entire space, represents the i-th subspace, is the weight coefficient; in this way, the structural space unit data is obtained; then, the construction sequence analysis of the structural space unit data is carried out, and the topological sorting algorithm is used to generate the spatial construction sequence data by considering the structural dependency and construction process requirements; according to the spatial construction sequence data, the construction progress data is time-mapped; using the space-time mapping function , where s represents the spatial coordinate and t represents the time coordinate; the mapping function can be expressed as: ； in, and are the spatial and temporal basis functions, respectively, is the weight coefficient, n is the number of basis functions; by mapping, the spatiotemporal correlation data is obtained; conflict detection is performed on the spatiotemporal correlation data, and the spatial interference detection algorithm is used to check the overlap of different construction activities in time and space to obtain spatial conflict data; the spatial conflict data is classified and counted, and cluster analysis methods such as K-means algorithm are used to classify conflicts according to characteristics such as type and severity to obtain conflict type data; according to the conflict type data, a rule-based expert system is used to generate conflict solutions to form conflict processing data; According to the conflict handling data, the spatiotemporal correlation data is adjusted; the adjustment process uses the simulated annealing algorithm to fine-tune the spatiotemporal position of the construction activities while ensuring the construction logic, and obtain preliminary progress correlation data; the preliminary progress correlation data is analyzed for resource allocation, and the linear programming model is used to calculate the demand for various resources in each time period to obtain resource demand data; The resource demand data is balanced and the resource-constrained project scheduling algorithm (RCPSP) is used to minimize the volatility of resource usage. The objective function of the balance calculation is expressed as: ； in, Representation Resources r In time t The amount of use, Representation Resources r The average usage of , T is the total number of time periods, R is the number of resource types; by solving the optimization problem, the resource allocation plan data is obtained; according to the resource allocation plan data, the preliminary progress related data is corrected to obtain the target progress related data; Conduct construction process analysis on target progress-related data, use process flow charts to represent the logical relationship between construction processes, and obtain process flow data; based on the process flow data, use genetic algorithms to generate multiple feasible construction plans and form candidate construction plan data; Finally, a multi-objective evaluation is conducted on the candidate construction scheme data; the evaluation adopts a method combining hierarchical analysis method and fuzzy comprehensive evaluation, taking into account multiple objectives such as construction period, cost, quality, and safety; the comprehensive score of the evaluation indicators is calculated as follows: ； in, For the The weight of the evaluation index, For the Under the indicator The degree of membership of the level, It is the element of fuzzy evaluation matrix; the scheme evaluation index data is obtained by calculation, and the optimal scheme is selected according to the evaluation results to obtain the construction scheme improvement data; A fiber Bragg grating sensor network is buried in the box girder structure to collect strain, temperature and vibration data in real time to obtain raw spectral data, and the raw spectral data is processed and classified to obtain structured construction data; Performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data; The progress deviation data is subjected to time series analysis through a long short-term memory network to obtain early warning data and cause diagnosis data, and decision analysis and processing are performed on the early warning data and the cause diagnosis data based on the construction plan improvement data to obtain recommendation data, which includes: construction progress adjustment suggestions and resource allocation optimization plans.

2. The method for managing the construction progress of a cast-in-place box girder based on BIM according to claim 1, characterized in that: The step of digitally processing the design drawings to obtain structural data, and dividing the structural data into work breakdown structures to obtain engineering task data includes: Vectorizing the geometric information in the design drawing to obtain structural contour data, and performing three-dimensional coordinate conversion on the structural contour data to obtain three-dimensional structural data; Extracting features from the three-dimensional structure data to obtain structure feature data, and performing parameterized modeling on the structure according to the structure feature data to obtain parameterized structure data; Associating the parameterized structure data with attribute information to obtain semantic structure data, and hierarchically classifying the semantic structure data to obtain layered structure data; Performing construction sequence analysis on the structure according to the hierarchical structure data to obtain construction sequence data, and dividing the construction sequence data into process steps to obtain process unit data; Performing resource demand analysis on the process unit data to obtain resource configuration data, and performing time estimation on the process according to the resource configuration data to obtain process duration data; constructing a process network according to the process duration data and the construction sequence data to obtain preliminary engineering task data, and performing a critical path analysis on the preliminary engineering task data to obtain key process data; Resource balancing is performed on the key process data to obtain balanced engineering task data, and a construction plan is generated according to the balanced engineering task data to obtain engineering task data.

3. The method for managing the construction progress of a cast-in-place box girder based on BIM according to claim 1, characterized in that: The step of performing a schedule analysis on the engineering task data to obtain schedule network data, and adjusting the schedule of the schedule network data to obtain construction schedule data includes: Performing task classification on the engineering task data to obtain classified task data, and performing priority sorting on the classified task data to obtain task priority data; Performing process dependency analysis on the tasks according to the task priority data to obtain process dependency relationship data, and performing critical path calculation on the process dependency relationship data to obtain critical path data; Estimating time parameters of the critical path data to obtain process duration data, and constructing a schedule network diagram based on the process duration data to obtain preliminary schedule network data; Performing resource load analysis on the preliminary schedule network data to obtain resource load data, and performing balance adjustment on the schedule network according to the resource load data to obtain candidate schedule network data; Performing a multi-objective evaluation on the candidate progress network data to obtain network evaluation index data, and selecting a progress network to be adjusted according to the network evaluation index data to obtain target progress network data; Calculating the total duration of the project according to the target schedule network data to obtain initial duration data, and performing constraint condition analysis on the initial duration data to obtain duration adjustment parameters; The target progress network data is compressed or extended according to the construction period adjustment parameters to obtain adjusted progress network data, and the time parameters of the adjusted progress network data are updated to obtain construction progress data.

4. The method for managing the construction progress of a cast-in-place box girder based on BIM according to claim 1, characterized in that: The steps of embedding a fiber Bragg grating sensor network in the box girder structure, collecting strain, temperature and vibration data in real time to obtain raw spectral data, and processing and classifying the raw spectral data to obtain structured construction data include: Performing sensor layout planning on the box girder structure to obtain sensor layout scheme data, and burying a fiber Bragg grating sensor network in the box girder structure according to the sensor layout scheme data to obtain sensor network data; Setting signal acquisition parameters for the sensor network data to obtain acquisition parameter data, and performing signal acquisition on the fiber Bragg grating sensor network according to the acquisition parameter data to obtain original spectrum data; Performing wavelength demodulation processing on the original spectrum data to obtain wavelength data, and performing temperature compensation calculation on the wavelength data to obtain wavelength data after temperature compensation; Performing strain conversion calculation on the temperature-compensated wavelength data to obtain strain data, and performing vibration feature extraction on the strain data to obtain vibration feature data; Performing time series merging on the strain data, the wavelength data after temperature compensation and the vibration characteristic data to obtain multi-dimensional time series data, and performing outlier detection on the multi-dimensional time series data to obtain cleaned time series data; Performing data standardization processing on the cleaned time series data to obtain standardized data, and performing feature extraction on the standardized data to obtain feature vector data; The feature vector data is subjected to data classification processing to obtain classification label data, and the classification label data is combined with the feature vector data to obtain structured construction data.

5. The method for managing the construction progress of a cast-in-place box girder based on BIM according to claim 1, characterized in that: The step of performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data includes: Performing time series segmentation on the structured construction data to obtain construction stage data, and performing feature extraction on the construction stage data to obtain stage feature data; Performing pattern recognition processing on the stage characteristic data to obtain a construction process recognition result, and performing quantitative calculation on the construction progress according to the construction process recognition result to obtain preliminary actual progress data; Performing data fusion processing on the preliminary actual progress data to obtain multi-source progress data, and performing consistency verification on the multi-source progress data to obtain verified progress data; Performing time standardization processing on the verified progress data to obtain standardized progress data, and performing progress index calculation on the standardized progress data to obtain actual progress index data; Synchronously processing the construction progress data according to the actual progress indicator data to obtain aligned progress data, and performing difference calculation between the aligned progress data and the actual progress indicator data to obtain original deviation data; Performing statistical analysis on the original deviation data to obtain deviation statistical feature data, and performing threshold judgment on the deviation statistical feature data to obtain deviation type data; The deviation type data and the original deviation data are combined and processed to obtain progress deviation data.

6. The method for managing the construction progress of a cast-in-place box girder based on BIM according to claim 5 is characterized in that: The step of performing time series analysis on the progress deviation data through the long short-term memory network to obtain early warning data and cause diagnosis data, and performing decision analysis and processing on the early warning data and the cause diagnosis data based on the construction plan improvement data to obtain suggestion data, wherein the suggestion data includes: construction progress adjustment suggestions and resource allocation optimization plan steps, including: Performing time series analysis on the progress deviation data through a long short-term memory network to obtain deviation trend data, and performing warning level classification according to the deviation trend data to obtain warning level data; Performing threshold judgment on the warning level data to obtain trigger warning data, and generating warning information according to the trigger warning data to obtain warning data; Extracting features from the progress deviation data to obtain deviation feature data, and classifying the deviation feature data using a decision tree algorithm to obtain deviation cause category data; Retrieving a historical case database according to the deviation cause category data to obtain similar case data, and performing correlation analysis on the similar case data to obtain cause correlation data; Performing weight calculation on the cause association data to obtain cause weight data, and sorting the deviation causes according to the cause weight data to obtain cause diagnosis data; Performing correlation analysis on the early warning data, the cause diagnosis data and the construction plan improvement data to obtain problem feature data, and generating a decision tree based on the problem feature data to obtain decision path data; The decision path data is processed for solution generation to obtain candidate solution data, and the candidate solution data is evaluated through a multi-criteria decision-making method to obtain recommendation data, which includes: construction schedule adjustment suggestions and resource allocation optimization solutions.

7. A BIM-based cast-in-place box girder construction progress management system, used to implement the BIM-based cast-in-place box girder construction progress management method according to any one of claims 1 to 6, characterized in that: include: A processing module is used to digitally process the design drawings to obtain structural data, and to divide the structural data into work breakdown structures to obtain engineering task data; An analysis module is used to perform a schedule analysis on the engineering task data to obtain schedule network data, and to adjust the duration of the schedule network data to obtain construction schedule data; The processing module is used to associate the structure data with the construction progress data to obtain progress association data, and simulate and analyze the progress association data to obtain construction plan improvement data, including: performing spatial partitioning processing on the structure data to obtain structure space unit data, and performing construction sequence analysis on the structure space unit data to obtain space construction sequence data; performing time mapping on the construction progress data according to the space construction sequence data to obtain time-space association data, and performing conflict detection on the time-space association data to obtain space conflict data; performing classification statistics on the space conflict data to obtain conflict type data, and generating a conflict resolution method according to the conflict type data to obtain conflict resolution data; The conflict handling data adjusts the spatiotemporal correlation data to obtain preliminary progress correlation data, and performs resource allocation analysis on the preliminary progress correlation data to obtain resource demand data; performs balance calculation on the resource demand data to obtain resource allocation scheme data, and amends the preliminary progress correlation data according to the resource allocation scheme data to obtain target progress correlation data; performs construction process analysis on the target progress correlation data to obtain process flow data, and generates a candidate set of construction schemes according to the process flow data to obtain candidate construction scheme data; performs multi-objective evaluation on the candidate construction scheme data to obtain scheme evaluation index data, and selects a recommended scheme according to the scheme evaluation index data to obtain construction scheme improvement data; Among them, the structural data is processed by spatial partitioning; the octree partitioning algorithm is used to divide the entire box girder structure into multiple spatial units; the octree partitioning algorithm recursively divides the three-dimensional space into eight subspaces until the preset accuracy or depth is reached; the mathematical expression of spatial partitioning is as follows: ； in, Represents the entire space, represents the i-th subspace, is the weight coefficient; in this way, the structural space unit data is obtained; then, the construction sequence analysis of the structural space unit data is carried out, and the topological sorting algorithm is used to generate the spatial construction sequence data by considering the structural dependency and construction process requirements; according to the spatial construction sequence data, the construction progress data is time-mapped; using the space-time mapping function , where s represents the spatial coordinate and t represents the time coordinate; the mapping function can be expressed as: ； in, and are the spatial and temporal basis functions, respectively, is the weight coefficient, n is the number of basis functions; by mapping, the spatiotemporal correlation data is obtained; conflict detection is performed on the spatiotemporal correlation data, and the spatial interference detection algorithm is used to check the overlap of different construction activities in time and space to obtain spatial conflict data; the spatial conflict data is classified and counted, and cluster analysis methods such as K-means algorithm are used to classify conflicts according to characteristics such as type and severity to obtain conflict type data; according to the conflict type data, a rule-based expert system is used to generate conflict solutions to form conflict processing data; According to the conflict handling data, the spatiotemporal correlation data is adjusted; the adjustment process uses the simulated annealing algorithm to fine-tune the spatiotemporal position of the construction activities while ensuring the construction logic, and obtain preliminary progress correlation data; the preliminary progress correlation data is analyzed for resource allocation, and the linear programming model is used to calculate the demand for various resources in each time period to obtain resource demand data; The resource demand data is balanced and the resource-constrained project scheduling algorithm (RCPSP) is used to minimize the volatility of resource usage. The objective function of the balance calculation is expressed as: ； in, Representation Resources r In time t The amount of use, Representation Resources r The average usage of , T is the total number of time periods, R is the number of resource types; by solving the optimization problem, the resource allocation plan data is obtained; according to the resource allocation plan data, the preliminary progress related data is corrected to obtain the target progress related data; Conduct construction process analysis on target progress-related data, use process flow charts to represent the logical relationship between construction processes, and obtain process flow data; based on the process flow data, use genetic algorithms to generate multiple feasible construction plans and form candidate construction plan data; Finally, a multi-objective evaluation is conducted on the candidate construction scheme data; the evaluation adopts a method combining hierarchical analysis method and fuzzy comprehensive evaluation, taking into account multiple objectives such as construction period, cost, quality, and safety; the comprehensive score of the evaluation indicators is calculated as follows: ； in, For the The weight of the evaluation index, For the Under the indicator The degree of membership of the level, It is the element of fuzzy evaluation matrix; the scheme evaluation index data is obtained by calculation, and the optimal scheme is selected according to the evaluation results to obtain the construction scheme improvement data; The acquisition module is used to bury a fiber Bragg grating sensor network in the box girder structure, collect strain, temperature and vibration data in real time, obtain raw spectral data, and process and classify the raw spectral data to obtain structured construction data; A comparison module, used for performing progress analysis on the structured construction data to obtain actual progress data, and comparing the actual progress data with the construction progress data to obtain progress deviation data; The decision-making module is used to perform time series analysis on the progress deviation data through a long short-term memory network to obtain early warning data and cause diagnosis data, and perform decision analysis and processing on the early warning data and the cause diagnosis data based on the construction plan improvement data to obtain recommendation data, wherein the recommendation data includes: construction progress adjustment suggestions and resource allocation optimization plans.

Images