Construction site project intelligent management method based on digital twinning

Abstract

The invention provides a construction site project intelligent management method based on digital twinning, and the method comprises the steps: collecting construction state data through deploying a sensor array at a construction site, generating twinning input data through space-time alignment and adaptive weighted fusion, and triggering the re-collection calibration when the data fluctuation exceeds a threshold value. Constructing a digital twinborn platform based on the fused data; calculating a progress deviation index to evaluate a construction state; filtering and smoothing short-term fluctuation through a sliding window; predicting a progress trend by using an LSTM model and early warning when a plan is deviated; the emergency degree and the work area contribution degree are comprehensively considered when the scheduling priority is set, a multi-objective optimization algorithm is applied to evaluate resource requirements and dynamically monitor inventory, and finally, an improved ant colony algorithm is adopted to optimize a resource scheduling scheme, so that dynamic management and control of the construction process are realized. The construction management efficiency can be improved, and the resource waste and construction period delay risk can be reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent management of construction engineering, and more particularly, to a construction project intelligent management method based on digital twinning. BACKGROUND

[0002] With the continuous development of the construction industry, the complexity of construction project management is increasing. Traditional construction project management mainly relies on manual inspection and experience judgment, which is not only inefficient, but also easily affected by subjective factors, leading to inaccurate information and decision-making errors. In recent years, with the rapid development of information technology, digital twinning technology has been gradually introduced into construction project management. Digital twinning technology reflects the state and behavior of physical entities in real time by constructing virtual digital models, providing a new way of thinking for construction project management. However, the existing digital twinning technology has some limitations in the application of construction project management.

[0003] The existing digital twinning technology in the application of construction project management mainly focuses on data collection and visualization. By deploying a sensor array on the construction site, collecting construction state data, and transmitting it to the digital twinning platform for visualization, management personnel can intuitively understand the real-time state of the construction site. However, this method still has deficiencies in data processing and analysis. First, the collected data often comes from different sensors with different physical dimensions and timestamps, making it difficult to directly fuse and analyze. Second, the existing digital twinning technology is relatively simple in progress evaluation and resource scheduling, mainly relying on pre-set rules and thresholds, and cannot be dynamically adjusted according to actual conditions. In addition, the existing digital twinning technology also has deficiencies in risk warning and resource optimization, which cannot accurately predict and respond to possible risks in a timely manner, resulting in resource waste and project delay.

[0004] In the implementation of the embodiments of the present application, there are at least the following problems or defects in the prior art: the precision and efficiency of data collection and processing are insufficient, the dynamic and flexibility of progress evaluation and resource scheduling are insufficient, and the accuracy and timeliness of risk warning and resource optimization are insufficient. These problems seriously affect the application effect of digital twinning technology in construction project management, and limit its potential in improving the efficiency and quality of construction project management. SUMMARY

[0005] The present application provides a construction project intelligent management method based on digital twinning, comprising: deploying a sensor array on the construction site to collect construction state data, and performing spatio-temporal alignment processing on the data, generating twin input data through an adaptive weighted fusion algorithm, and triggering data re-sampling and calibration when the fluctuation exceeds the threshold; According to the fusion data, a digital twin platform is constructed, a progress deviation index is calculated to evaluate the construction state, short-term fluctuations are smoothed by using a sliding window filter, a future progress trend is predicted by using an LSTM model, a warning is triggered when deviating from the planned baseline, a twin risk level is divided according to the progress deviation index and the twin resource consumption factor, and a twin scheduling instruction is generated when the risk is high; On the basis of progress evaluation and resource prediction, scheduling priority is set, the urgency is calculated by combining the progress deviation index and the twin resource risk, the importance of the work area is determined by using the work area contribution method, the scheduling resource demand is evaluated by using a multi-objective optimization algorithm, and the material inventory state is dynamically monitored; An improved ant colony optimization algorithm is used to dynamically optimize the resource scheduling scheme.

[0006] Further, a sensor array is deployed on the construction site to collect construction state data, and the data is processed by spatiotemporal alignment. The aligned data is generated by an adaptive weighted fusion algorithm, and data reacquisition and calibration are triggered when fluctuations exceed the threshold. The specific steps are as follows: The sensor array collects data including tower crane inclination parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters; The collected data is standardized to make the data range of different physical quantities consistent; Each standardized data is detected for abnormality by using an isolation forest algorithm to eliminate noise data, and the collected data is processed by spatiotemporal alignment to align different work area data to the same timestamp; After spatiotemporal alignment, the collected data is fused based on a dynamic weight distribution adaptive weighted fusion algorithm. The weight distribution is dynamically adjusted according to the historical correlation between each parameter and the construction progress; The fused data is used as the twin input data at the current time, and a feedback mechanism is set. When the difference between the fused value and the value at the previous time point exceeds the set threshold, the sensor array data collection is restarted and the measurement equipment is calibrated.

[0007] Further, according to the fusion data, a digital twin platform is constructed, a progress deviation index is calculated to evaluate the construction state, short-term fluctuations are smoothed by using a sliding window filter, and the specific steps include: The fusion data is used as the input data for progress evaluation. A progress deviation model is established for the fusion data. A construction progress deviation index is generated by multivariate time series analysis and covariance matrix decomposition. The work area delay state is determined according to the progress deviation index; The construction progress deviation index is dynamically smoothed by using a sliding window filter to eliminate the influence of short-term disturbances; For the work area whose delay state continuously exceeds the limit, the deviation degree of the eigenvalue of the fusion data from the planned threshold is determined to determine the twin resource consumption factor, which is used to analyze the twin cost risk.

[0008] Further, the future progress trend is predicted by the LSTM model, and a warning is triggered when deviating from the planned baseline, and the specific steps are as follows: A progress prediction model is constructed based on historical data, and an LSTM model is applied to predict the progress deviation index at the future time; The input of the progress prediction model is the current and historical progress deviation index and the twin resource consumption factor; If the predicted progress deviation index deviates from the planned baseline threshold continuously, a warning signal is triggered, prompting the risk of time delay.

[0009] Further, according to the progress deviation index and the twin resource consumption factor, the twin risk level is divided, and the twin scheduling instruction is generated when the risk is high, including the following steps: According to the deterioration rate of the progress deviation index and the cumulative amount of the twin resource consumption factor, the twin risk level of each work area is determined, and the twin risk level includes a low risk level, a warning risk level and a high risk level; When the twin risk level is a low risk level, no twin scheduling instruction is generated; When the twin risk level is a warning risk level, the work area is marked as a warning work area, and an inspection enhancement scheme is developed; When the twin risk level is a high risk level, a twin scheduling instruction signal is generated, triggering the automatic material scheduling system to allocate resources; Wherein, the condition of the low risk level is: The deterioration rate is lower than the first deterioration rate threshold, and the cumulative amount of the twin resource consumption factor is lower than the first resource consumption cumulative amount threshold; The condition of the warning risk level is: The deterioration rate reaches or exceeds the first deterioration rate threshold but is lower than the second deterioration rate threshold, and the cumulative amount of the twin resource consumption factor is lower than the first resource consumption cumulative amount threshold; or, the cumulative amount of the twin resource consumption factor reaches or exceeds the first resource consumption cumulative amount threshold but is lower than the second resource consumption cumulative amount threshold, and the deterioration rate is lower than the first deterioration rate threshold; The condition of the high risk level is: the deterioration rate reaches or exceeds the second deterioration rate threshold; or, the cumulative amount of the twin resource consumption factor reaches or exceeds the second resource consumption cumulative amount threshold; or, the deterioration rate reaches or exceeds the first deterioration rate threshold and the cumulative amount of the twin resource consumption factor reaches or exceeds the first resource consumption cumulative amount threshold.

[0010] Further, on the basis of progress evaluation and resource prediction, the scheduling priority is set, the urgency is calculated by combining the progress deviation index and the twin resource risk, and the importance of the work area is determined by the work area contribution degree quantification method, including the following steps: According to the progress evaluation and resource prediction results, the scheduling demand urgency of each work area is quantified, and the urgency index is used to represent the resource demand intensity of the work area at the current time; The relative importance of each work area is quantified by the work area contribution method, a work area importance judgment matrix is constructed, and the spatial relationship between the work area and the critical path is valued; The weight distribution of the judgment matrix is calculated by the coefficient of variation method, and the calculation results are standardized to obtain the weight of each work area as the work area importance index; After the calculation is completed, the key scheduling work area is determined according to the obtained work area importance index value; After obtaining the task urgency and work area importance, the comprehensive scheduling priority index is calculated, and the expression formula is:

[0011] Among them, is the comprehensive scheduling priority index of work area i; is the adjustment parameter, used to adjust the influence weight of the urgency index and the historical resource consumption cumulative amount in the calculation of the comprehensive scheduling priority index; is the urgency index of work area i; represents the twin resource consumption factor cumulative amount of work area i at historical time t; is the decay coefficient, which reflects the degree of decay of the influence of historical data on the current scheduling priority over time; is the importance index of work area i; is the base of the exponential function, used to adjust the calculation scale of the logarithmic term; According to the comprehensive scheduling priority index, the scheduling tasks of each work area are sorted to generate a scheduling task queue, and the task sorting follows the principle of priority from high to low. The first work area in the queue is prioritized for resource allocation.

[0012] Further, a multi-objective optimization algorithm is applied to evaluate the scheduling resource demand and dynamically monitor the material inventory state, including the following steps: Based on the priority and resource gap of each scheduling task, the required concrete volume and steel tonnage are evaluated; According to the type of scheduling task and the requirements of construction process, the material proportioning scheme required by each task is determined, and combined with the warehouse inventory limit and supply chain capacity, the material transportation plan to meet all task demands is calculated; For each scheduling task, analyze the equipment demand, including the number of pump trucks and the occupation time of tower cranes; Based on the priority index, material demand, equipment demand, and manpower demand of the task, and applying a multi-objective particle swarm optimization algorithm for dynamic resource scheduling, the resource scheduling cost is determined.

[0013] Further, the improved ant colony optimization algorithm is used to dynamically optimize the resource scheduling scheme, including the following steps: A multi-objective optimization model is constructed, a multi-objective optimization model is constructed based on resource constraints and task priorities, the objective function of the model includes transportation cost, scheduling response time and resource idle rate, and the resources scheduled include construction materials, transportation vehicles and labor teams; A dynamic pheromone evaporation mechanism is introduced in the standard ant colony algorithm, and when the actual time consumption of the scheduling path exceeds the predicted value, the pheromone concentration of the path is automatically reduced; A resource conflict detection module is set, when multiple tasks compete for the same device, the device allocation path is dynamically adjusted according to the priority index.

[0014] Further, it also includes real-time visual interaction of the digital twin platform: The scheduling scheme is mapped to the BIM model to generate a three-dimensional dynamic deduction view; A bias heat map is embedded in the view interface, and the chroma of the heat map is positively correlated with the progress bias index; When the user clicks on a high-risk work area, the twin resource consumption factor historical curve and the prediction result of the work area are automatically popped up.

[0015] Further, after generating the twin scheduling instruction, a closed-loop verification is started: After the execution of the twin scheduling instruction, the actual progress bias index is calculated by collecting the sensor data of the work area in real time; The actual index and the predicted index are input into a residual analysis model, and if the absolute value of the residual continuously exceeds the tolerance threshold, the parameter recalibration of the digital twin platform is triggered; After recalibration, a new twin scheduling instruction is generated to replace the original instruction until the absolute value of the residual is lower than the set threshold.

[0016] The above embodiments of the present application have at least the following beneficial effects: 1. By deploying a sensor array and combining an adaptive weighted fusion algorithm, accurate acquisition and dynamic calibration of construction state data are realized, the problems of data dispersion and error accumulation in traditional construction site monitoring are solved, the real-time and reliability of the input data of the digital twin platform are ensured, and a high-quality data foundation is provided for subsequent analysis.

[0017] 2. By using the progress bias index, the LSTM prediction model and the sliding window filtering technology, dynamic evaluation and trend prediction of the construction progress are realized, the delay risk of the construction period is effectively identified and early warning is given, the problems of response lag and large subjective judgment deviation in manual progress management are solved, and the accuracy and foresight of the construction progress control are improved.

[0018] 3. The resource scheduling strategy based on the multi-objective optimization algorithm and the improved ant colony algorithm dynamically allocates resources in combination with the contribution and priority of the work area, solves the problems of resource conflict and low efficiency in traditional scheduling, realizes the optimal allocation of materials, equipment and manpower, and improves the resource utilization rate and construction collaboration efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0019] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description read in conjunction with the accompanying drawings. In the drawings, several embodiments of the present application are shown by way of example, and not limitation, wherein: Figure 1 A flowchart of a construction site project intelligent management method based on digital twinning provided by an embodiment of the present application. DETAILED DESCRIPTION

[0020] The principles and spirits of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and implement the present application, and do not limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0021] Those skilled in the art know that the embodiments of the present application can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present application can be embodied in the form of a complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0022] It should be noted that any number of elements in the drawings is used for example and not limitation, and any naming is only used for distinction and does not have any limiting meaning.

[0023] The principles and spirits of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and implement the present application, and do not limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art. Figure 1 , Figure 1 A flowchart of a construction site project intelligent management method based on digital twinning provided by an embodiment of the present application. As shown in Figure 1 , a construction site project intelligent management method based on digital twinning includes: S1, deploying a sensor array at the construction site to collect construction state data, and performing spatio-temporal alignment processing on the data, generating twin input data through adaptive weighted fusion algorithm on the aligned data, and triggering data re-sampling and calibration when the fluctuation exceeds the threshold; S2, constructing a digital twin platform according to the fused data, calculating a progress deviation index to evaluate the construction state, smoothing short-term fluctuations using a sliding window filter, predicting future progress trends through an LSTM model, triggering an early warning when deviating from the planned baseline, dividing the twin risk level according to the progress deviation index and the twin resource consumption factor, and generating a twin scheduling instruction when the risk is high; S3, setting scheduling priorities based on progress evaluation and resource prediction, calculating the urgency by combining the progress deviation index and the twin resource risk, determining the importance of work areas through work area contribution quantification methods, and evaluating scheduling resource requirements using a multi-objective optimization algorithm to dynamically monitor material inventory status; S4, using an improved ant colony optimization algorithm to dynamically optimize resource scheduling schemes.

[0024] It should be noted that the present application proposes a digital twin-based intelligent management method for construction projects, aiming to realize real-time monitoring, accurate evaluation and intelligent scheduling of construction state through advanced sensor technology and data analysis algorithm. In this method, first, a sensor array is deployed on the construction site to collect construction state data, including tower crane inclination parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters. These parameters are key indicators for measuring construction safety, quality and progress, and through monitoring, potential problems can be discovered in time and appropriate measures can be taken. The collected data will be processed through spatio-temporal alignment to ensure that data from different sources can be analyzed in the same time frame. Spatio-temporal alignment is an important step in data preprocessing, which solves the problem of inconsistent data collection times and provides a foundation for subsequent data fusion and analysis. Then, the aligned data is fused using an adaptive weighted fusion algorithm to generate twin input data. The adaptive weighted fusion algorithm is an algorithm that dynamically adjusts weights, which allocates weights according to the historical correlation of each parameter with construction progress, thus more accurately reflecting the construction state. When data fluctuations exceed the set threshold, a data re-sampling and calibration mechanism will be triggered to ensure the accuracy and reliability of the data. This mechanism is crucial for ensuring the stable operation of the digital twin platform, as the accuracy of the data directly affects the subsequent progress evaluation and resource scheduling.

[0025] Specifically, the data collected by the sensor array is standardized to make the data range consistent across different physical dimensions, facilitating subsequent analysis and processing. Standardization is a common data preprocessing method that converts data to the same scale, eliminating differences between different dimensions. In this embodiment, the tower crane inclination parameter is in degrees, while the concrete temperature and humidity parameters are in Celsius and percentage. Through standardization, they can be converted to the range of 0 to 1. The anomaly detection uses the Isolation Forest algorithm, which is an unsupervised learning algorithm based on tree structure, capable of effectively identifying abnormal points in the data, thus eliminating noise data. The spatio-temporal alignment process aligns the data of different work areas to the same timestamp, ensuring the synchronization of data in time. The adaptive weighted fusion algorithm with dynamic weight allocation dynamically adjusts the weight according to the historical correlation of each parameter with the construction progress. In this embodiment, if the historical data shows that the correlation between the tower crane inclination parameter and the construction progress is higher, this parameter will be given a higher weight in the fusion process. The progress deviation index is generated through multivariate time series analysis and covariance matrix decomposition, which reflects the deviation between the construction progress and the planned progress. The sliding window filtering is a common signal processing method used to smooth short-term fluctuations, thus more clearly observing long-term trends. The LSTM model is a long short-term memory network, belonging to the deep learning algorithm, used to predict the progress deviation index at future time. The input of this model includes the current and historical progress deviation index and the twin resource consumption factor, which learns the patterns and rules in the historical data to predict the future progress trend. The twin risk level is determined according to the deterioration rate of the progress deviation index and the cumulative amount of the twin resource consumption factor, which divides the risk of the work area into low risk level, warning risk level and high risk level, so as to take corresponding scheduling measures.

[0026] Preferably, the construction of the digital twin platform is based on fused data, and a construction progress deviation index is generated through multivariate time series analysis and covariance matrix decomposition. Multivariate time series analysis is a statistical analysis method that considers the interrelationships between multiple time series, which can more accurately reflect the changing trend of construction progress. Covariance matrix decomposition is used to extract the main components of the data, thereby simplifying the model and improving computational efficiency. In the construction process of the LSTM model, the input parameters include the current and historical progress deviation index and the twin resource consumption factor, which are learned and predicted through the training process of the model. The training of the model requires a large amount of historical data, which is used to adjust the parameters of the model so that it can accurately predict future progress trends. When the predicted progress deviation index deviates from the planned baseline threshold, a warning signal is triggered, indicating the risk of delay. The triggering mechanism of the warning signal can be set according to the specific project requirements and risk tolerance, and in this embodiment, a warning can be triggered when the progress deviation index deviates from the planned baseline by more than 10%. The classification of the twin risk level can be determined according to the specific project situation and risk assessment standards, and in this embodiment, specific thresholds for deterioration rate and resource consumption accumulation can be set to more accurately assess the risk level of the work area. In terms of resource scheduling, the scheduling priority can be determined according to the importance and urgency of the work area, and the resource scheduling demand can be evaluated through a multi-objective optimization algorithm, and the material inventory status can be dynamically monitored. The multi-objective optimization algorithm can consider multiple objectives such as cost, time, and resource utilization, and determine the optimal resource scheduling scheme by optimizing these objectives.

[0027] Specifically, the twin scheduling instruction is a dynamic scheduling instruction generated by the digital twin platform based on construction progress evaluation, resource consumption analysis, and risk level classification. In this invention, the digital twin platform processes and analyzes the data collected by the sensor array on the construction site to construct a virtual model highly synchronized with the actual construction process. The twin scheduling instruction is generated based on this virtual model, which reflects the intelligent decision-making of the digital twin platform on construction progress and resource allocation.

[0028] Specifically, the generation process of the twin scheduling instruction is as follows: First, the digital twin platform collects construction state data through the sensor array, including tower crane inclination parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters. After time-space alignment and adaptive weighted fusion processing, the twin input data is generated. Then, the platform calculates the progress deviation index based on these fused data, evaluates the construction state, and predicts the future progress trend through the LSTM model. When the predicted progress deviation index deviates from the planned baseline, the platform will divide the twin risk level according to the progress deviation index and the twin resource consumption factor. For high-risk work areas, the digital twin platform will generate a twin scheduling instruction.

[0029] In this application, the twin scheduling instructions include but are not limited to resource allocation instructions, task priority adjustment instructions, equipment allocation instructions, etc. For example, when the progress deviation index of a certain work area shows that there is a risk of delay in the construction period, the twin scheduling instructions may indicate to increase the number of construction personnel in the work area, adjust the equipment allocation priority, or optimize the material supply plan. These instructions are executed through an automated material scheduling system, thereby achieving dynamic control of the construction process.

[0030] In some embodiments, a sensor array is deployed on the construction site to collect construction state data, and the data is processed for spatio-temporal alignment. The aligned data is generated by an adaptive weighted fusion algorithm, and data re-sampling and calibration are triggered when fluctuations exceed a threshold. The specific steps are as follows: The sensor array collects data including tower crane inclination parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters; The collected data is standardized to ensure consistency in the range of different physical dimensions; Each standardized data is subjected to anomaly detection using the Isolation Forest algorithm to eliminate noise data, and the collected data is subjected to spatio-temporal alignment to align different work area data to the same timestamp; After spatio-temporal alignment, the collected data is fused based on a dynamic weight distribution adaptive weighted fusion algorithm, and the weight distribution is dynamically adjusted according to the historical correlation between each parameter and the construction progress; The fused data is used as the current time's twin input data, and a feedback mechanism is set up. When the difference between the fused value and the previous time point value exceeds a certain threshold, the sensor array data collection is restarted and the measuring equipment is calibrated.

[0031] It should be noted that the present application deploys a sensor array on the construction site to collect construction state data, and processes the data for spatio-temporal alignment to generate twin input data, and triggers data re-sampling and calibration when fluctuations exceed a threshold. The sensor array is a network composed of multiple sensors, used to collect various construction state data in real time. These data include tower crane inclination parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters, which are key indicators for measuring construction safety, quality and progress. Spatio-temporal alignment processing is a time synchronization process of data collected from different sensors at different times, to ensure data consistency and comparability. The adaptive weighted fusion algorithm is an algorithm that dynamically adjusts weights, which allocates weights according to the historical correlation between each parameter and the construction progress, to more accurately reflect the construction state. When the data fluctuation exceeds a certain threshold, the data re-sampling and calibration mechanism is triggered to ensure the accuracy and reliability of the data, which is crucial for the stable operation of the digital twin platform.

[0032] Specifically, the data collected by the sensor array includes tower crane inclination parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters. The tower crane inclination parameters refer to the angle between the tower crane and the vertical direction during operation, which is used to monitor the safety status of the tower crane. The concrete temperature and humidity parameters refer to the changes in temperature and humidity during the curing process, which are used to ensure the quality of the concrete. The construction personnel positioning parameters refer to the location information of the construction personnel on the construction site, which are used for personnel management and safety monitoring. The collected data is standardized to ensure that the data ranges of different physical dimensions are consistent. In this embodiment, the range of the tower crane inclination parameters is converted from 0 to 90 degrees to a standardized value of 0 to 1, and the range of the concrete temperature and humidity parameters is converted from Celsius and percentage to a standardized value. The anomaly detection uses the Isolation Forest algorithm, which is an unsupervised learning algorithm based on tree structure, which can effectively identify abnormal points in the data and eliminate noise data. The spatio-temporal alignment process aligns the data of different work areas to the same timestamp, ensuring the synchronization of the data in time. The adaptive weighted fusion algorithm dynamically adjusts the weights according to the historical correlation between each parameter and the construction progress. In this embodiment, if the historical data shows that the correlation between the tower crane inclination parameters and the construction progress is high, the parameter will be given a higher weight in the fusion process. The fused data is used as the current time's twin input data, and a feedback mechanism is set up. When the difference between the fusion value and the previous time point value exceeds the set threshold, the sensor array data collection and measurement device calibration are restarted to ensure the accuracy and reliability of the data.

[0033] Preferably, the deployment of the sensor array needs to be optimized according to the specific layout and construction requirements of the construction site to ensure the comprehensiveness and accuracy of data collection. In this embodiment, high-precision inclination sensors are installed at key positions of the tower crane, temperature and humidity sensors are installed in the concrete curing area, and positioning sensors are installed in areas where construction personnel frequently move. The standardization of data can be achieved by calculating the Z-score of each parameter, i.e. each data point is subtracted by the average value of the parameter and then divided by the standard deviation, so that data of different dimensions are converted to the same scale. The parameter settings of the Isolation Forest algorithm include the number of trees and the size of the subsample, which can be adjusted according to the size and complexity of the data to improve the accuracy of anomaly detection. The spatio-temporal alignment process can be achieved by interpolation method. For data points with inconsistent timestamps, the value corresponding to the timestamp is calculated by interpolation, so that the data is synchronized. The weight allocation of the adaptive weighted fusion algorithm can be determined by correlation analysis of historical data. In this embodiment, the correlation coefficient between each parameter and the construction progress is calculated, and the weight is dynamically adjusted according to the size of the correlation coefficient. The threshold of the feedback mechanism can be set according to the actual construction situation and the normal range of data fluctuations. In this embodiment, when the difference between the fusion value and the previous time point value exceeds 10%, the data re-collection and calibration mechanism is triggered.

[0034] In some embodiments, a digital twin platform is constructed according to the fusion data, a progress deviation index is calculated to evaluate the construction state, and a sliding window filter is used to smooth short-term fluctuations, and the specific steps include: The fusion data is used as input data for progress evaluation, a progress deviation model is established for the fusion data, a construction progress deviation index is generated through multivariate time series analysis and covariance matrix decomposition, and a work area delay state is determined according to the progress deviation index; The sliding window filter is used to dynamically smooth the construction progress deviation index and short-term disturbance effects; For the work area whose delay state continuously exceeds the limit, the deviation degree of the characteristic value of the fusion data from the planned threshold is extracted to determine a twin resource consumption factor, which is used to analyze the twin cost risk.

[0035] It should be noted that the present application further refines the construction process of the digital twin platform, especially the calculation of the progress deviation index and the determination of the twin resource consumption factor. The progress deviation index is a key indicator for measuring the deviation between the construction progress and the planned progress, which is generated through multivariate time series analysis and covariance matrix decomposition, and can accurately reflect the real-time state of the construction progress. The sliding window filter is used to smooth short-term fluctuations, making the progress deviation index more stable and facilitating long-term trend observation. The twin resource consumption factor is used to analyze the twin cost risk, which is determined by extracting the deviation degree of the characteristic value of the fusion data from the planned threshold, providing a basis for resource scheduling and risk assessment.

[0036] Specifically, the fusion data is used as input data for progress evaluation, and the calculation of the progress deviation index involves multivariate time series analysis and covariance matrix decomposition. Multivariate time series analysis is a statistical method for analyzing the mutual relationship between multiple time series, which can reveal the potential patterns and trends in construction progress data. Covariance matrix decomposition is a mathematical method for extracting the main components of data, thereby simplifying the model and improving computational efficiency. The progress deviation index is used to determine the work area delay state, i.e., whether the construction progress lags behind the planned progress. The sliding window filter is a signal processing technique that smooths short-term fluctuations by sliding a fixed-size window over the time series and calculating the average value of the data within the window. For the work area whose delay state continuously exceeds the limit, the deviation degree of the characteristic value of the fusion data from the planned threshold needs to be extracted to determine the twin resource consumption factor. The twin resource consumption factor reflects the deviation between resource consumption and planned consumption, and is an important indicator for evaluating resource use efficiency and cost risk.

[0037] Preferably, the calculation of the progress deviation index can be achieved by the following steps: first, collect and organize the fusion data, ensure the integrity and accuracy of the data. Then, apply the multivariate time series analysis method, in this embodiment, the autoregressive moving average model ARMA or the vector autoregressive model VAR, to analyze the time dependence and mutual relationship in the data. Next, through the covariance matrix decomposition, in this embodiment, the principal component analysis PCA, the main components in the data are extracted, and the progress deviation index is obtained. The specific implementation of the sliding window filtering can set a fixed size window, in this embodiment, 30 minutes or 1 hour, and adjust the window size according to the frequency and fluctuation of the actual construction data. The average value or median value of the data in the window is calculated to smooth the short-term fluctuations. The determination of the twin resource consumption factor can be achieved by calculating the difference between the eigenvalue of the fusion data and the planned threshold value, in this embodiment, the percentage difference between the actual resource consumption and the planned resource consumption. These difference values can be further used to evaluate the resource use efficiency and cost risk, and provide data support for resource scheduling and risk warning.

[0038] In some embodiments, the future progress trend is predicted by the LSTM model, and a warning is triggered when deviating from the planned baseline, with the specific steps as follows: A progress prediction model is constructed based on historical data, and the LSTM model is applied to predict the progress deviation index at future time; The input of the progress prediction model is the current and historical progress deviation index and the twin resource consumption factor; If the predicted progress deviation index deviates from the planned baseline threshold continuously, a warning signal is triggered, indicating that there is a risk of delay in the construction period.

[0039] It should be noted that the present application further refines the construction process of the digital twin platform, especially by predicting the future construction progress trend through the long short-term memory network LSTM model, and triggering a warning when deviating from the planned baseline. The LSTM model is a deep learning algorithm that can effectively process time series data and capture long-term dependencies in the data, thereby accurately predicting the future progress deviation index. The input of the progress prediction model includes the current and historical progress deviation index and the twin resource consumption factor, which can provide rich historical information for the model to help it better learn and predict. When the predicted progress deviation index deviates from the planned baseline threshold continuously, a warning signal is triggered, indicating that there is a risk of delay in the construction period, thereby providing timely decision support for management personnel.

[0040] Specifically, the LSTM model is a special type of recurrent neural network designed to handle long-term dependencies in time series data. It uses a series of gating mechanisms, such as the forget gate, input gate, and output gate, to control the flow of information, effectively learning the temporal dependencies in the data. The input to the progress prediction model includes current and historical progress deviation indices and twin resource consumption factors, where the progress deviation index reflects the deviation between the actual progress and the planned progress, and the twin resource consumption factor reflects the deviation between the actual resource consumption and the planned consumption. These input parameters provide comprehensive historical information to the model, enabling it to more accurately predict future progress trends. The planned baseline threshold is a reference value set based on the construction plan, used to determine whether the predicted progress deviation index is within the normal range. When the predicted progress deviation index consistently deviates from the planned baseline threshold, it indicates that there may be problems with the construction progress, and timely measures need to be taken.

[0041] Preferably, the construction of the LSTM model can be achieved through the following steps: first, collect and organize historical progress deviation indices and twin resource consumption factor data to ensure data integrity and accuracy. Then, divide the data into training and testing sets for model training and validation. Next, construct the LSTM network structure, including the input layer, LSTM layer, and output layer. The input layer receives current and historical progress deviation indices and twin resource consumption factors, the LSTM layer learns the temporal dependencies in the data through gating mechanisms, and the output layer outputs the predicted progress deviation index. During model training, use mean squared error as the loss function, and optimize model parameters through the backpropagation algorithm. When the model performs as expected on the test set, it can be applied to actual progress prediction. The triggering mechanism of the early warning signal can be set according to the specific project requirements and risk tolerance. In this embodiment, the early warning can be triggered when the predicted progress deviation index deviates from the planned baseline by more than 10%.

[0042] In some embodiments, the twin risk level is divided according to the progress deviation index and the twin resource consumption factor, and the twin scheduling instruction is generated when the risk is high, including the following steps: According to the deterioration rate of the progress deviation index and the cumulative amount of the twin resource consumption factor, determine the twin risk level of each work area, which includes low risk level, alert risk level and high risk level; When the twin risk level is low risk level, no twin scheduling instruction is generated; When the twin risk level is alert risk level, mark the work area as an alert work area and develop a strengthened inspection plan; When the twin risk level is high risk level, generate a twin scheduling instruction signal to trigger the automatic material scheduling system to allocate resources; The condition of the low-risk level is: The deterioration rate is lower than a first deterioration rate threshold, and the cumulative amount of the twin resource consumption factor is lower than a first resource consumption cumulative amount threshold; The condition of the alert-risk level is: The deterioration rate reaches or exceeds a first deterioration rate threshold but is lower than a second deterioration rate threshold, and the cumulative amount of the twin resource consumption factor is lower than a first resource consumption cumulative amount threshold; or, the cumulative amount of the twin resource consumption factor reaches or exceeds a first resource consumption cumulative amount threshold but is lower than a second resource consumption cumulative amount threshold, and the deterioration rate is lower than a first deterioration rate threshold; The condition of the high-risk level is: the deterioration rate reaches or exceeds a second deterioration rate threshold; or, the cumulative amount of the twin resource consumption factor reaches or exceeds a second resource consumption cumulative amount threshold; or, the deterioration rate reaches or exceeds a first deterioration rate threshold and the cumulative amount of the twin resource consumption factor reaches or exceeds a first resource consumption cumulative amount threshold.

[0043] It should be noted that the present application further refines the construction process of the digital twin platform, especially by dividing the twin risk level through the progress deviation index and the twin resource consumption factor, and generating a twin scheduling instruction when the risk is high. The twin risk level is determined according to the deterioration rate of the progress deviation index and the cumulative amount of the twin resource consumption factor, and is divided into a low-risk level, an alert-risk level, and a high-risk level. This grading method can help managers more intuitively understand the risk state of each work area and take appropriate measures. When the twin risk level is high, a twin scheduling instruction signal is generated to trigger the automated material scheduling system to allocate resources, thereby ensuring the rationality of construction progress and resource use.

[0044] Specifically, the division of the twin risk level is based on two key parameters: the deterioration rate of the progress deviation index and the cumulative amount of the twin resource consumption factor. The deterioration rate of the progress deviation index refers to the rate of change of the progress deviation index over time, reflecting the speed of deviation of the construction progress from the plan. The cumulative amount of the twin resource consumption factor refers to the cumulative value of the deviation between resource consumption and planned consumption, reflecting resource use efficiency and cost risk. The low-risk level indicates that the progress deviation and resource consumption of the work area are within a controllable range; the alert-risk level indicates that the progress deviation or resource consumption of the work area has approached the critical value and needs to be monitored; the high-risk level indicates that the progress deviation or resource consumption of the work area has exceeded the critical value and needs to be addressed immediately. The thresholds can be determined according to the specific project requirements and risk tolerance, and in the present embodiment, the first threshold of the deterioration rate can be set to 5%, and the second threshold to 10%; the first threshold of the resource consumption cumulative amount can be set to 10%, and the second threshold to 20%.

[0045] Preferably, the classification of twin risk levels can be achieved by the following steps: first, calculate the deterioration rate of the progress deviation index and the cumulative amount of the twin resource consumption factor of each work area. The deterioration rate can be calculated by time-differentiating the progress deviation index, and the cumulative amount can be calculated by time-integrating the resource consumption deviation. Then, according to the set threshold, the work area is divided into different risk levels. In this embodiment, when the deterioration rate is less than 5% and the cumulative amount is less than 10%, the work area is classified as a low risk level; when the deterioration rate is between 5% and 10% and the cumulative amount is less than 10%, or the cumulative amount is between 10% and 20% and the deterioration rate is less than 5%, the work area is classified as a warning risk level; when the deterioration rate reaches or exceeds 10%, or the cumulative amount reaches or exceeds 20%, or the deterioration rate reaches or exceeds 5% and the cumulative amount reaches or exceeds 10%, the work area is classified as a high risk level. For the work area of high risk level, the twin scheduling instruction signal is generated, triggering the automatic material scheduling system to allocate resources. The specific content of the scheduling instruction can be determined according to the demand of the work area and the availability of resources. In this embodiment, manpower is increased, equipment allocation is adjusted, or material supply plan is optimized.

[0046] In some embodiments, based on progress evaluation and resource prediction, scheduling priority is set, urgency is calculated by combining progress deviation index and twin resource risk, and work area importance is determined by work area contribution quantification method, including the following steps: According to the results of progress evaluation and resource prediction, the urgency of scheduling demand of each work area is quantified, and the urgency index is used to represent the resource demand intensity of the work area at the current time; The relative importance of each work area is determined by the work area contribution quantification method, a work area importance judgment matrix is constructed, and the matrix is valued according to the spatial relationship between the work area and the critical path; The weight distribution of the judgment matrix is calculated by the coefficient of variation method, the calculation result is standardized, and the weight of each work area is obtained as the work area importance index; After the calculation, the key scheduling work area is determined according to the obtained work area importance index value; After obtaining the task urgency and work area importance, the comprehensive scheduling priority index is calculated, and the expression formula is:

[0047] Wherein, is the comprehensive scheduling priority index of work area i; is a regulation parameter, used to adjust the influence weight of the urgency index and the historical resource consumption cumulative amount in the calculation of the comprehensive scheduling priority index; is the urgency index of work area i; represents the cumulative amount of the twin resource consumption factor of work area i at historical time t; is a decay coefficient, representing the degree of decay of the influence of historical data on the current scheduling priority over time; is the importance index of work area i; is the base of the exponential function, used to adjust the calculation scale of the logarithmic term; T is the length of the time window; According to the comprehensive scheduling priority index, the scheduling tasks of each work area are sorted to generate a scheduling task queue, and the task sorting follows the principle of priority from high to low, and the first work area in the queue is preferentially allocated resources.

[0048] It should be noted that the present application further refines the construction process of the digital twin platform, especially how to set the scheduling priority based on progress evaluation and resource prediction, and combine the progress deviation index and the twin resource risk to calculate the urgency, and determine the importance of the work area by the work area contribution quantification method. The setting of scheduling priority is the key link of resource management, which determines which work area should be given priority to resource support under the condition of limited resources. The urgency index is used to represent the resource demand intensity of the work area at the current time, and the work area contribution quantification method is used to determine the relative importance of each work area, which helps to balance efficiency and fairness in resource allocation. The calculation formula of the comprehensive scheduling priority index comprehensively considers the urgency, historical resource consumption cumulative amount and work area importance, which provides a scientific basis for resource scheduling.

[0049] Specifically, the setting of scheduling priority is based on the results of progress evaluation and resource prediction, and the urgency index is used to quantify the resource demand intensity of each work area. The work area contribution quantification method determines the relative importance of each work area by constructing a work area importance judgment matrix and assigning values according to the spatial relationship between the work area and the critical path. The coefficient of variation method is used to calculate the weight distribution of the judgment matrix, and the weight of each work area is obtained through standardization processing, which is used as the work area importance index. In the calculation formula of the comprehensive scheduling priority index, the adjustment parameter is used to adjust the influence weight of the urgency index and the historical resource consumption cumulative amount in the calculation of the comprehensive scheduling priority, and the decay coefficient reflects the degree of decay of the influence of historical data on the current scheduling priority over time. The setting of these parameters needs to be determined according to the specific project requirements and resource management strategies to ensure the rationality and effectiveness of the scheduling decision.

[0050] Preferably, the setting of scheduling priorities can be achieved through the following steps: First, calculate the urgency index for each work area based on the progress assessment and resource forecast results. In this embodiment, the urgency can be determined by analyzing the changing trends of the progress deviation index and resource consumption factor. Next, determine the relative importance of each work area using a work area contribution metric method. In this embodiment, a work area importance judgment matrix can be constructed, and values ​​can be assigned based on the spatial relationship between the work area and the critical path. Then, the weight of each work area can be calculated using the coefficient of variation method. When calculating the comprehensive scheduling priority index, adjustment parameters and attenuation coefficients can be set according to the specific needs of the project. In this embodiment, if the project places more emphasis on urgency, the weight of the urgency index can be increased; if the project wants to consider the impact of historical data more, the attenuation coefficient can be decreased. Finally, the scheduling tasks of each work area are sorted according to the comprehensive scheduling priority index to generate a scheduling task queue, ensuring the rationality and efficiency of resource allocation.

[0051] Adjust parameters The value range is [0,1], and its main function is to balance the weighting of the urgency index and the historical resource consumption accumulation in the calculation of the comprehensive scheduling priority index. When When the overall scheduling priority index approaches 0, it will place greater emphasis on reflecting the importance of the work area, that is, prioritizing the work area's key contribution to the overall project; while when... When the index approaches 1, the overall scheduling priority index focuses more on reflecting the urgency level and historical resource consumption accumulation, prioritizing both the urgency of current resource needs and historical resource consumption. In practical applications, it can be flexibly adjusted according to the specific needs of the project and management strategies. The value of . For example, if the project manager is more concerned with the immediate needs of resources and short-term efficiency, then can be . Set it to a higher value (e.g., 0.7 or 0.8); conversely, if the project manager focuses more on long-term resource optimization and overall project schedule, then... Set it to a lower value (such as 0.2 or 0.3).

[0052] Adjust parameters This is a time decay factor used to reflect the degree to which the influence of historical data on the current scheduling priority decays over time, and its value ranges from (0,1). In this invention, exponential decay is recommended because it more naturally reflects the changes in the importance of historical data over time. In implementation, if... A value of 0.9 indicates that the weight of historical data decreases by 10% after each time step. In practical applications, this can be adjusted based on the specific needs of the project and the dynamic nature of the data. The value of the parameter. If the project manager wants to pay more attention to the recent resource consumption, the parameter can be set to a smaller value (e.g., 0.7 or 0.8); if the project manager wants to consider the resource consumption over a longer period of time, the parameter can be set to a larger value (e.g., 0.9 or 0.95).

[0053] In some embodiments, a multi-objective optimization algorithm is applied to evaluate the scheduling resource demand and dynamically monitor the material inventory status, including the following steps: Based on the priority of each scheduling task and the resource gap, the required concrete volume and steel tonnage are evaluated; According to the type of scheduling task and the construction process requirements, the material proportioning scheme required by each task is determined, and combined with the warehouse inventory limit and supply chain capacity, the material transportation plan that meets the demand of all tasks is calculated; For each scheduling task, the equipment demand is analyzed, including the number of pump trucks and the tower crane occupation time; Based on the priority index of the task and the material demand, equipment demand, and manpower demand, a multi-objective particle swarm optimization algorithm is applied for dynamic resource scheduling to determine the resource scheduling cost.

[0054] It should be noted that the present application further refines the construction process of the digital twin platform, especially the application of a multi-objective optimization algorithm to evaluate the scheduling resource demand and dynamically monitor the material inventory status. The multi-objective optimization algorithm is an optimization method that can consider multiple objectives simultaneously, used to balance different demands and constraints in resource scheduling. The evaluation of scheduling resource demand is based on the priority of each scheduling task and the resource gap to determine the specific resources required, such as concrete volume and steel tonnage. Dynamic monitoring of material inventory status ensures reasonable allocation and timely supply of resources, avoiding resource waste and project delay.

[0055] Specifically, the multi-objective optimization algorithm is used to evaluate the scheduling resource demand, which considers multiple objectives such as cost, time, and resource utilization. The priority of the scheduling task is determined according to the importance and urgency of the work area, while the resource gap refers to the difference between the resources required to complete the task and the currently available resources. In this embodiment, concrete volume and steel tonnage are common resource demands in construction tasks, and their specific values depend on the scale of the task and the construction process requirements. The material proportioning scheme is determined according to the type of scheduling task and the construction process requirements, which specifies the specific proportion of each material. Warehouse inventory limits and supply chain capacity are important factors affecting material transportation planning, which determine the supply capacity and transportation efficiency of materials. Equipment demand analysis includes determining the number of pump trucks and tower crane occupation time required for each scheduling task, which directly affects construction progress and resource allocation.

[0056] ​​Preferably, the construction of the multi-objective optimization algorithm can be realized by the following steps: first, define the objective function of the optimization problem, such as minimizing the resource scheduling cost, maximizing the resource utilization rate, etc. in this embodiment. Then, according to the priority of the scheduling task and the resource gap, determine the amount of resources required by each task, such as the amount of concrete and the number of tons of steel. Then, combined with the warehouse inventory limit and the supply chain capacity, develop a material transportation plan that meets the demand of all tasks. For equipment demand analysis, according to the construction process requirements and task types, determine the number of pump trucks and tower crane occupation time required by each task. Finally, a multi-objective particle swarm optimization algorithm is applied for dynamic resource scheduling, which finds the optimal solution by simulating the behavior of a particle swarm. During the algorithm running process, according to the priority index of the task, material demand, equipment demand, and manpower demand, dynamically adjust the resource allocation scheme to ensure the rationality and efficiency of resource scheduling.

[0057] In some embodiments, an improved ant colony optimization algorithm is used to dynamically optimize the resource scheduling scheme, including the following steps: A multi-objective optimization model is constructed based on resource constraints and task priorities, and the objective function of the model includes transportation cost, scheduling response time, and resource idle rate. The resources scheduled include construction materials, transportation vehicles, and labor teams. A dynamic pheromone evaporation mechanism is introduced into the standard ant colony algorithm, which automatically reduces the pheromone concentration of the scheduling path when the actual time consumption of the path exceeds the predicted value. A resource conflict detection module is set up to dynamically adjust the device allocation path according to the priority index when multiple tasks compete for the same device.

[0058] It should be noted that the present application further refines the construction process of the digital twin platform, especially using an improved ant colony optimization algorithm to dynamically optimize the resource scheduling scheme. Ant colony optimization algorithm is an optimization algorithm that simulates the foraging behavior of ants, which finds the optimal path through the transmission and update of pheromones. The improved ant colony optimization algorithm introduces a dynamic pheromone evaporation mechanism based on the standard ant colony algorithm, which automatically reduces the pheromone concentration of the scheduling path when the actual time consumption of the path exceeds the predicted value, thereby improving the adaptability and efficiency of the algorithm. In addition, a resource conflict detection module is set up to dynamically adjust the device allocation path according to the priority index when multiple tasks compete for the same device, ensuring the rationality and fairness of resource allocation.

[0059] Specifically, the improved ant colony optimization algorithm is a swarm intelligence-based optimization method that seeks the optimal solution by simulating the behavior of ants releasing and sensing pheromones during the process of finding food. In resource scheduling, pheromones represent the quality of paths, and ants choose paths based on pheromone concentration. The dynamic pheromone evaporation mechanism is an improvement measure that automatically reduces the pheromone concentration of the path when the actual time consumption exceeds the predicted value, prompting ants to find better paths. The resource conflict detection module is used to handle the case where multiple tasks compete for the same device. It dynamically adjusts the device allocation path based on the task priority index to ensure that high-priority tasks can obtain resources first. The objective functions of the multi-objective optimization model include transportation cost, scheduling response time, and resource idle rate, which reflect the key performance indicators in resource scheduling. The scheduled resources include construction materials, transportation vehicles, and labor teams, and the reasonable allocation of these resources is crucial for construction progress and cost control.

[0060] Preferably, the construction of the improved ant colony optimization algorithm can be achieved through the following steps: first, construct a multi-objective optimization model and define the objective functions including transportation cost, scheduling response time, and resource idle rate. Transportation cost can be determined by calculating the distance and cost of material transportation; scheduling response time refers to the time interval from task allocation to task start; resource idle rate refers to the proportion of time that resources are not fully utilized. Then, introduce the dynamic pheromone evaporation mechanism based on the standard ant colony algorithm, set the pheromone evaporation rate parameter, and reduce the pheromone concentration of the path according to the preset rules when the actual time consumption of the scheduling path exceeds the predicted value. Next, set the resource conflict detection module, and when multiple tasks compete for the same device, dynamically adjust the device allocation path based on the task priority index. The priority index can be calculated based on the urgency of the task and the importance of the work area. Finally, by simulating the behavior of ants, continuously update the pheromone concentration, and find the optimal resource scheduling scheme to ensure the rationality and efficiency of resource allocation.

[0061] In some embodiments, it also includes real-time visualization interaction of the digital twin platform: Mapping the scheduling scheme to the BIM model generates a three-dimensional dynamic deduction view; Embedding a bias heat map in the view interface, with the color of the heat map positively related to the progress bias index; When the user clicks on a high-risk work area, the twin resource consumption factor historical curve and prediction results of that work area are automatically popped up.

[0062] It should be noted that the present application further refines the construction process of the digital twin platform, especially the real-time visual interactive function of the digital twin platform. This function maps the scheduling scheme to the building information model (BIM) to generate a three-dimensional dynamic deduction view, enabling managers to visually observe the construction progress and resource allocation. In the view interface, a deviation heat map is embedded, and the color of the heat map is positively correlated with the progress deviation index, thereby intuitively displaying the progress deviation. In addition, when the user clicks on a high-risk work area, the historical curve and prediction results of the twin resource consumption factor of the work area are automatically popped up, providing detailed data support for managers and helping them make more accurate decisions.

[0063] Specifically, the real-time visual interactive function of the digital twin platform realizes a three-dimensional dynamic deduction view by mapping the scheduling scheme to the BIM model. The BIM model is a digital model that integrates information throughout the life cycle of a building project, containing information such as the geometry, spatial relationships, geographic information, and attributes of building components. The deviation heat map is a visualization tool that uses color depth to represent the size of the progress deviation, with deeper color indicating greater deviation. The progress deviation index is an indicator that measures the deviation between the actual progress and the planned progress, obtained by calculating the difference between the actual progress and the planned progress. The historical curve and prediction results of the twin resource consumption factor provide historical data and future trend predictions of resource consumption in the work area, helping managers understand resource usage and potential risks.

[0064] Preferably, the real-time visual interactive function of the digital twin platform can be realized by the following steps: first, integrate the data of the scheduling scheme with the BIM model to ensure the accuracy and real-time nature of the data. Then, develop a three-dimensional dynamic deduction view to dynamically display the scheduling scheme in the BIM model through visualization technology, enabling managers to visually see changes in construction progress and resource allocation. Next, embed a deviation heat map that dynamically adjusts the color of the heat map based on the size of the progress deviation index to enable managers to quickly identify areas with greater progress deviation. Finally, set an interactive function for high-risk work areas that automatically pops up the historical curve and prediction results of the twin resource consumption factor when the user clicks, providing detailed data analysis and prediction information. The implementation of these functions requires the combination of advanced visualization technology and data analysis algorithms to ensure the accuracy and usability of the information.

[0065] In some embodiments, after generating the twin scheduling instructions, a closed-loop verification is initiated: After the execution of the twin scheduling instructions, real-time collection of work area sensor data is performed to calculate the actual progress deviation index; The actual index and the predicted index are input into a residual analysis model, and if the absolute value of the residual continuously exceeds the tolerance threshold, the parameter recalibration of the digital twin platform is triggered; After recalibration, new twin scheduling instructions are generated to replace the original instructions until the absolute value of the residual is below the set threshold.

[0066] It should be noted that the present application further refines the construction process of the digital twin platform, especially the closed-loop verification mechanism after the execution of the twin scheduling instructions. Closed-loop verification is a key link to ensure the stable operation of the digital twin platform and the accuracy of the scheduling instructions. After the execution of the twin scheduling instructions, real-time collection of work area sensor data is performed to calculate the actual progress deviation index, and the actual index and the predicted index are input into the residual analysis model. If the absolute value of the residual continuously exceeds the tolerance threshold, the parameter recalibration of the digital twin platform is triggered. After recalibration, new twin scheduling instructions are generated to replace the original instructions until the absolute value of the residual is below the set threshold. This process ensures the continuous optimization of the digital twin platform and the accuracy of the scheduling instructions.

[0067] Specifically, the closed-loop verification mechanism includes real-time collection of work area sensor data, calculation of actual progress deviation index, residual analysis, and parameter recalibration. The work area sensor data refers to real-time data collected from the sensor array on the construction site, including tower crane inclination parameters, concrete temperature and humidity parameters, construction personnel positioning parameters, etc. The actual progress deviation index is calculated based on the real-time collected data, reflecting the actual deviation between the construction progress and the planned progress. The residual analysis model is used to compare the difference between the actual progress deviation index and the predicted progress deviation index, and the residual is the difference between the two. The tolerance threshold is a parameter set according to the project requirements, used to judge whether the residual is within an acceptable range. When the absolute value of the residual continuously exceeds the tolerance threshold, it indicates that the current digital twin platform parameters need to be adjusted, at which time parameter recalibration is triggered. After recalibration, new twin scheduling instructions are generated to ensure that subsequent scheduling instructions can more accurately reflect the actual situation.

[0068] Preferably, the implementation of the closed-loop verification mechanism can be realized through the following steps: First, set up a real-time data collection system to ensure that work area sensor data can be obtained in a timely manner. Then, calculate the actual progress deviation index based on the collected data. This process can use the same calculation method as the predicted progress deviation index to ensure comparability. Next, input the actual progress deviation index and the predicted progress deviation index into the residual analysis model to calculate the residual and determine whether it exceeds the tolerance threshold. The tolerance threshold can be set according to the accuracy requirements of the project and the fluctuation of historical data. In this embodiment, it is set to 5% or 10%. If the absolute value of the residual exceeds the tolerance threshold, it indicates that there is a large deviation between the current model's prediction and the actual situation, and parameter recalibration is needed. Recalibration can be achieved by adjusting model parameters, updating data models, or optimizing algorithms. Finally, new twin scheduling instructions are generated based on the recalibrated model to replace the original instructions, ensuring that subsequent scheduling instructions can more accurately reflect the actual situation until the absolute value of the residual is below the set threshold, completing the closed-loop verification.

[0069] Further, the storage medium of the embodiments of the present application stores program instructions capable of implementing all the methods described above, wherein the program instructions can be stored in the storage medium in the form of a software product, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor execute all or part of the steps of the methods described in various embodiments of the present application. And the aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media capable of storing program codes, or a computer, a server, a mobile phone, a tablet, and other terminal devices.

[0070] The above description is merely some preferred embodiments of the present application and a description of the technical principles applied. Those skilled in the art should understand that the inventive scope of the embodiments of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combinations of the above technical features or their equivalent features without departing from the inventive concept. In the embodiments, the above features are replaced with the technical features disclosed in the embodiments of the present application (but not limited to) having similar functions to form technical solutions.

Claims

1. A method for intelligent management of construction site projects based on digital twins, characterized in that, Includes the following steps: Sensor arrays are deployed on the construction site to collect construction status data, and the data is spatiotemporally aligned. The aligned data is then used to generate twin input data through an adaptive weighted fusion algorithm, and data re-sampling and calibration are triggered when fluctuations exceed the threshold. A digital twin platform is built based on fused data. The schedule deviation index is calculated to assess the construction status. A sliding window filter is used to smooth short-term fluctuations. The future schedule trend is predicted through an LSTM model. An early warning is triggered when the schedule deviates from the planned baseline. The twin risk level is divided according to the schedule deviation index and the twin resource consumption factor. A twin scheduling instruction is generated when the risk is high. Based on schedule assessment and resource forecasting, scheduling priorities are set, urgency is calculated by combining schedule deviation index and twin resource risk, and the importance of work area is determined by work area contribution quantification method. Multi-objective optimization algorithm is applied to evaluate scheduling resource demand and material inventory status is dynamically monitored. The resource scheduling scheme is dynamically optimized using an improved ant colony optimization algorithm. The process of setting scheduling priorities based on schedule assessment and resource forecasting, calculating urgency by combining schedule deviation index and twin resource risk, and determining the importance of work areas through a work area contribution quantification method includes the following steps: Based on the progress assessment and resource forecast results, the urgency of scheduling needs for each work area is quantified, and the urgency index is used to represent the intensity of resource demand for the work area at the current moment. The relative importance of each work area is determined by the work area contribution metric method, a work area importance judgment matrix is ​​constructed, and values ​​are assigned according to the spatial relationship between the work area and the critical path; The weight distribution of the judgment matrix is ​​calculated by the coefficient of variation method, and the calculation results are standardized to obtain the weight of each work area as the work area importance index. After the calculation is completed, the key scheduling work areas are determined based on the obtained work area importance index value; After obtaining the urgency of the task and the importance of the work area, the comprehensive scheduling priority index is calculated, expressed by the following formula: ； in, For the work area The comprehensive scheduling priority index; To adjust the parameters; This represents the urgency index of work zone i. This represents the cumulative amount of twin resource consumption factors in work area i at historical time t; This is the attenuation coefficient, which reflects the degree to which the impact of historical data on the current scheduling priority decays over time. is the importance index of work area i; b is the base of the exponential function, used to adjust the calculation scale of the logarithmic term; T is the length of the time window; The scheduling tasks of each work area are sorted according to the comprehensive scheduling priority index to generate a scheduling task queue. The task sorting follows the principle of priority from high to low, and the work area at the top of the queue is given priority in resource allocation.

2. The method according to claim 1, characterized in that, A sensor array is deployed at the construction site to collect construction status data, and the data undergoes spatiotemporal alignment processing. The aligned data is then used to generate twin input data through an adaptive weighted fusion algorithm, and data re-acquisition and calibration are triggered when fluctuations exceed a threshold. The specific steps are as follows: The sensor array collects data including tower crane tilt angle parameters, concrete temperature and humidity parameters, and construction personnel positioning parameters; The collected data is standardized to ensure that the data ranges of different physical dimensions are consistent. Anomaly detection was performed on each standardized data set using the Isolation Forest algorithm to remove noisy data, and the collected data was spatiotemporally aligned to align data from different work areas to the same timestamp. After completing the spatiotemporal alignment, an adaptive weighted fusion algorithm based on dynamic weight allocation is used to fuse the collected data. The weight allocation is dynamically adjusted according to the historical correlation between each parameter and the construction progress. The fused data is used as the twin input data at the current moment, and a feedback mechanism is set up. When the difference between the fused value and the value at the previous time point exceeds a set threshold, the sensor array data acquisition is restarted and the measurement equipment is calibrated.

3. The method according to claim 2, characterized in that, A digital twin platform is built based on the fused data. The schedule deviation index is calculated to assess the construction status, and a sliding window filter is used to smooth short-term fluctuations. The specific steps include: The fused data is used as input data for progress assessment. A progress deviation model is established on the fused data, and a construction progress deviation index is generated through multivariate time series analysis and covariance matrix decomposition. The delay status of the work area is determined based on the progress deviation index. Sliding window filtering is used to dynamically smooth the short-term disturbances affecting the construction progress deviation index. For work areas where delays consistently exceed limits, the deviation of the feature values ​​from the fused data from the planned thresholds is used to determine the twin resource consumption factor, which is then used to analyze twin cost risks.

4. The method according to claim 3, characterized in that, The LSTM model is used to predict future progress trends, and an early warning is triggered when the progress deviates from the planned baseline. The specific steps are as follows: A schedule prediction model is built based on historical data, and an LSTM model is applied to predict the schedule deviation index at future moments. The inputs to the schedule prediction model are the current and historical schedule deviation indices and twin resource consumption factors; If the predicted schedule deviation index continues to deviate from the planned baseline threshold, an early warning signal will be triggered, indicating a risk of project delay.

5. The method according to claim 4, characterized in that, Based on the schedule deviation index and twin resource consumption factor, twin risk levels are classified, and twin scheduling instructions are generated when the risk is high, including the following steps: Based on the deterioration rate of the schedule deviation index and the cumulative amount of the twin resource consumption factor, the twin risk level of each work area is determined, including low risk level, warning risk level and high risk level; When the twin risk level is low, no twin scheduling instruction is generated; When the twin risk level is the warning risk level, the work area is marked as a warning work area, and an inspection enhancement plan is formulated. When the twin risk level is high, a twin scheduling instruction signal is generated to trigger the automated material scheduling system to allocate resources. The conditions for the low-risk level are as follows: The deterioration rate is lower than the first deterioration rate threshold, and the cumulative amount of the twin resource consumption factor is lower than the first resource consumption cumulative amount threshold. The conditions for the aforementioned alert risk level are as follows: The deterioration rate reaches or exceeds a first deterioration rate threshold but is lower than a second deterioration rate threshold, and the cumulative amount of the twin resource consumption factor is lower than a first resource consumption accumulation threshold; or, the cumulative amount of the twin resource consumption factor reaches or exceeds a first resource consumption accumulation threshold but is lower than a second resource consumption accumulation threshold, and the deterioration rate is lower than a first deterioration rate threshold. The conditions for a high-risk level are: the deterioration rate reaches or exceeds a second deterioration rate threshold; or, the cumulative amount of the twin resource consumption factor reaches or exceeds a second resource consumption accumulation threshold; or, the deterioration rate reaches or exceeds a first deterioration rate threshold and the cumulative amount of the twin resource consumption factor reaches or exceeds a first resource consumption accumulation threshold.

6. The method according to claim 5, characterized in that, The application of multi-objective optimization algorithms to evaluate scheduling resource requirements and dynamically monitor material inventory status includes the following steps: Based on the priority and resource gap of each scheduled task, assess the required volume of concrete and the number of tons of steel reinforcement. Based on the scheduling task type and construction process requirements, determine the material allocation scheme required for each task, and calculate the material transportation plan that meets the needs of all tasks by taking into account warehouse inventory constraints and supply chain capabilities. For each scheduling task, analyze the equipment requirements, including the number of pump trucks and the duration of tower crane occupancy; Based on the task priority index and material, equipment, and manpower requirements, and using a multi-objective particle swarm optimization algorithm for dynamic resource scheduling, the resource scheduling cost is determined.

7. The method according to claim 6, characterized in that, The improved ant colony optimization algorithm is used to dynamically optimize resource scheduling schemes, including the following steps: A multi-objective optimization model is constructed based on resource constraints and task priorities. The objective function of the model includes transportation cost, scheduling response time and resource idle rate. The resources to be scheduled include construction materials, transportation vehicles and labor teams. A dynamic pheromone evaporation mechanism is introduced into the standard ant colony algorithm, which automatically reduces the pheromone concentration of a scheduled path when the actual time taken for the scheduled path exceeds the predicted value. A resource conflict detection module is set up to dynamically adjust the device allocation path based on the priority index when multiple tasks compete for the same device.

8. The method according to claim 7, characterized in that, It also includes real-time visual interaction on the digital twin platform: Map the scheduling scheme to the BIM model to generate a three-dimensional dynamic simulation view; Embed a deviation heatmap in the view interface; the color of the heatmap is positively correlated with the progress deviation index. When a user clicks on a high-risk work area, the historical curve and prediction results of the twin resource consumption factor for that work area will automatically pop up.

9. The method according to any one of claims 1-8, characterized in that, Initiate closed-loop verification after generating twin scheduling instructions: After the twin scheduling command is executed, real-time sensor data from the work area is collected to calculate the actual progress deviation index; Input the actual index and the predicted index into the residual analysis model. If the absolute value of the residual continues to exceed the tolerance threshold, the parameters of the digital twin platform will be recalibrated. After recalibration, a new twin scheduling instruction is generated to overwrite the original instruction until the absolute value of the residual is lower than a set threshold.

Images