Construction method and system of full life cycle digital twin based on multi-source heterogeneous data

Abstract

This invention relates to the field of digital monitoring technology and discloses a method and system for constructing a full lifecycle digital twin based on multi-source heterogeneous data. The method involves acquiring the original design data of a target building, identifying key nodes in the construction progress of the target building and reproducing a digital model based on the original design data to obtain expected reference information for the construction progress, collecting multi-source monitoring data during actual construction, matching the multi-source monitoring data according to the expected reference information, scheduling a specified expected reference model as the digital simulation model for the current construction cycle, and performing data segmentation and feedback content parsing on the multi-source monitoring data based on the digital simulation model to deploy parameters for the digital simulation model and obtain a real-time digital twin.

Description

Technical Field

[0001] This invention relates to the technical field of digital surveillance, and in particular to a method and system for constructing a full lifecycle digital twin based on multi-source heterogeneous data. Background Technology

[0002] In traditional building construction, quality and safety management mainly relies on manual inspection and experience-based judgment, which has certain limitations and lags. Building projects are usually characterized by long cycles, large scale, and many participants, making it difficult to discover potential problems in a timely manner and take measures to deal with them, which can easily lead to quality hazards and safety accidents. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for constructing a digital twin throughout its entire lifecycle based on multi-source heterogeneous data, aiming to solve the problem in the prior art that it is difficult to detect potential problems in a timely manner and take measures to deal with them.

[0004] This invention is implemented as follows: Firstly, this invention provides a method for constructing a full lifecycle digital twin based on multi-source heterogeneous data, comprising: The original design data of the target building is obtained, and the key nodes of the construction progress of the target building are identified and digitally modeled based on the original design data to obtain the expected reference information of the construction progress. Collect multi-source monitoring data during actual construction, and match the multi-source monitoring data according to the expected reference information to schedule the specified expected reference model as the digital simulation model for the current construction cycle; Based on the digital simulation model, the data of the current construction cycle is segmented and the feedback content is analyzed to deploy parameters of the digital simulation model and obtain a real-time digital twin.

[0005] Secondly, the present invention provides a full lifecycle digital twin construction system based on multi-source heterogeneous data, used to implement the full lifecycle digital twin construction method based on multi-source heterogeneous data as described in any one of the first aspects, including: The expected reference module is used to acquire the original design data of the target building, and to identify key nodes of the construction progress of the target building and reproduce the digital model based on the original design data, so as to obtain expected reference information on the construction progress. The digital simulation module is used to collect multi-source monitoring data during actual construction and match the multi-source monitoring data according to the expected reference information to schedule the specified expected reference model as the digital simulation model for the current construction cycle. The content feedback module is used to perform data segmentation and feedback content parsing of multi-source monitoring data based on the digital simulation model for the current construction cycle, so as to deploy parameters of the digital simulation model and obtain a real-time digital twin.

[0006] This invention provides a method for constructing a digital twin throughout its entire lifecycle based on multi-source heterogeneous data, which has the following beneficial effects: In construction management, this invention can accurately grasp the construction progress by identifying key nodes and replicating models, facilitating timely adjustment. Matching multi-source monitoring data with expected reference information makes the scheduling digital simulation model more realistic, improving simulation accuracy. Based on the model, the data segmentation and interpretation and the deployment of parameters to construct a real-time digital twin can reflect the actual building status in real time. Combined with AI and the fusion of multi-source heterogeneous data, it can also assist in quality supervision, providing strong support for the quality and progress management of the entire building life cycle, and improving construction efficiency and quality. Attached Figure Description

[0007] Figure 1 This is a schematic diagram illustrating the steps of a method for constructing a full lifecycle digital twin based on multi-source heterogeneous data, provided by an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a full lifecycle digital twin construction system based on multi-source heterogeneous data provided in an embodiment of the present invention. Detailed Implementation

[0008] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0009] The implementation of the present invention will be described in detail below with reference to specific embodiments.

[0010] Reference Figure 1 , Figure 2 The diagram shows a preferred embodiment of the present invention.

[0011] In a first aspect, the present invention provides a method for constructing a full lifecycle digital twin based on multi-source heterogeneous data, comprising: S1: Obtain the original design data of the target building, and identify key nodes of the construction progress and reproduce the digital model based on the original design data to obtain expected reference information on the construction progress; S2: Collect multi-source monitoring data during actual construction, and match the multi-source monitoring data according to the expected reference information, so as to schedule the specified expected reference model as the digital simulation model of the current construction cycle; S3: Based on the digital simulation model, perform data segmentation and feedback content analysis on the multi-source monitoring data for the current construction cycle, so as to deploy parameters for the digital simulation model and obtain a real-time digital twin. Specifically, in step S1 of the embodiment provided by the present invention, the original design data of the target building is obtained. This data usually includes information such as the building's design drawings, technical specifications, and construction plans. Based on this data, digital models of the target building's initial state and completed state are constructed using professional Building Information Modeling (BIM) software or related modeling tools. These are the initial digital model and the completed digital model. The initial digital model and the completed digital model are the basis for subsequent model deduction and key node identification. By clarifying the initial and final states of the building, clear boundaries and objectives can be provided for the entire construction process, which is helpful for subsequent analysis of changes and developments during the construction process.

[0012] More specifically, guided by the original design data, the initial digital model is gradually evolved according to the normal construction sequence, moving towards the completed digital model. During this process, the changes in the model at each step are recorded, forming a forward deduction information flow. Similarly, based on the original design data, starting from the completed digital model, the reverse process of building demolition or construction is simulated, moving towards the initial digital model. The model evolution process is recorded, resulting in a reverse deduction information flow. The forward deduction simulates the actual construction process of the building, which can intuitively show the development stages of the building from scratch. The reverse deduction examines the construction process from another perspective, which helps to discover potential problems and risks. By combining forward and reverse deduction, we can more comprehensively and accurately grasp each stage of building construction and provide richer information for the identification of key nodes.

[0013] More specifically, by combining forward and backward information flow analysis, the construction process of the target building is analyzed. Through comparison and screening, the time points or events that have a significant impact on the construction progress are identified. These points or events are the key construction nodes, such as the completion of foundation construction, the topping out of the main structure, and the commencement of decoration work. Key nodes are important milestones in the building construction process, marking the phased achievements of the construction progress. Identifying key nodes helps to effectively monitor and manage the construction process, and can promptly detect whether the construction progress meets expectations, so as to take corresponding measures for adjustment.

[0014] More specifically, for each key construction node, a digital model of the corresponding key node is reconstructed based on the original design data and information from the simulation process, resulting in a corresponding expected reference model for each key construction node. These models accurately reflect the status and characteristics of the target building at each key node. The expected reference models provide important reference for subsequent construction progress monitoring and the construction of digital twins. By comparing the actual construction situation with the expected reference models, deviations and problems in the construction process can be identified in a timely manner, providing support for real-time adjustments and optimizations. At the same time, these models also help to assess and predict the construction progress at different stages.

[0015] Specifically, in step S2 of the embodiment provided by the present invention, multi-source data of the actual construction process is collected by a multi-source sensing system pre-deployed at the construction site. The multi-source sensing system includes, but is not limited to, cameras, laser scanners, and sensors (such as temperature sensors, humidity sensors, stress sensors, etc.). Cameras can be used to capture images and videos of the construction site and record the construction scene and progress; laser scanners can acquire three-dimensional spatial information of the building structure; and various sensors can monitor environmental parameters and the physical state of the building structure in real time. These sensors transmit the collected data to the data processing center to form multi-source monitoring data.

[0016] More specifically, the construction process is a complex system involving multiple aspects of information. A single type of data cannot fully reflect the actual construction situation. By collecting different types of data through a multi-source sensing system, information about the construction site can be obtained from multiple dimensions, providing a richer and more accurate data foundation for subsequent analysis and decision-making.

[0017] More specifically, semantic fusion of heterogeneous data is performed on the collected multi-source monitoring data. Since the data collected by different types of sensors have different formats and meanings, data preprocessing and semantic analysis techniques are needed to transform these heterogeneous data into unified information with clear semantics to obtain fused identification information.

[0018] More specifically, the previously obtained expected reference information is input into a pre-trained construction progress recognition algorithm. This algorithm uses machine learning or deep learning techniques to match the fused recognition information and determines the real-time progress characteristics of the actual construction by comparing the features of the fused recognition information with the expected reference information.

[0019] More specifically, heterogeneous data semantic fusion can eliminate the differences between data, enabling data from different sources to be correlated and integrated, facilitating subsequent analysis and processing. The construction progress recognition algorithm can use expected reference information as a benchmark to quickly and accurately determine the actual construction progress, providing a basis for scheduling appropriate expected reference models.

[0020] More specifically, based on the identified real-time progress characteristics of the actual construction, the corresponding expected reference model is determined from the expected reference information. For example, if the real-time progress characteristics indicate that the current construction is in the stage of foundation construction completion, the expected reference model corresponding to the key node of foundation construction completion is found from the expected reference information, and the determined expected reference model is scheduled out as the digital simulation model of the current construction cycle.

[0021] More specifically, the expected reference model is constructed based on the key nodes of building construction and can accurately reflect the state of the building at different stages. By scheduling an appropriate expected reference model as a digital simulation model according to the actual construction progress, the digital simulation model can be matched with the actual construction situation, providing an accurate foundation for subsequent data segmentation, feedback content analysis and the construction of digital twins. This allows for more accurate simulation and analysis of the current construction cycle and timely detection of problems and deviations in the construction process.

[0022] Specifically, in step S3 of the embodiment provided by the present invention, based on the scheduled digital simulation model, combined with the construction stage characteristics represented by the model and time information, the construction cycle of the target building at the current moment is determined. For example, if the digital simulation model corresponds to the construction stage of a wall in the main structure construction, and with reference to the construction plan schedule, the specific time period of the current wall construction can be clearly identified, thereby defining the current construction cycle.

[0023] More specifically, clearly defining the current construction cycle is the foundation for subsequent data processing and analysis. Different construction cycles have different construction tasks and key focus areas. For example, the data to be monitored and the indicators to be focused on differ greatly between the foundation construction stage and the decoration stage. Only by accurately determining the current construction cycle can we process and analyze multi-source monitoring data in a targeted manner.

[0024] More specifically, the multi-source monitoring data is analyzed to determine the target of the data feedback content, that is, to determine whether each data point reflects the situation of the past construction cycle or the current construction cycle. Based on the analysis results, the multi-source monitoring data is divided into non-key data corresponding to the past construction cycle and key data corresponding to the current construction cycle. This can be automated by setting rules such as data timestamp range and data-related construction task nodes, or it can be supplemented by manual review and confirmation.

[0025] More specifically, during the entire construction process, multi-source monitoring data will continuously accumulate. While data from past construction cycles are of reference value for the overall construction, their importance is relatively low when analyzing the current construction cycle. By segmenting the data and focusing on the key data of the current construction cycle, we can reduce the interference of irrelevant data, improve the efficiency and accuracy of data processing, and more accurately analyze the current construction status.

[0026] More specifically, the interpretation of feedback content based on the structure and characteristics of the digital simulation model is carried out. For example, if the stress distribution of the wall is a key parameter in the digital simulation model, and the key data includes the monitoring values ​​of the wall stress sensor, the actual stress situation of the wall can be analyzed by comparing the theoretical stress value in the model with the actual monitoring value. Then, the actual building appearance reflected by the key data can be obtained. Based on the interpretation results, the parameters of the digital simulation model are deployed, that is, the actual monitored values ​​and characteristics are updated into the digital simulation model, such as adjusting the stress parameters and location parameters of the wall in the model, to obtain a real-time digital twin.

[0027] More specifically, the focus is on data that directly reflects the actual construction situation. By interpreting this data, we can understand the actual state of the building in the current construction cycle. Deploying parameters to the digital simulation model enables the model to more accurately simulate the real situation of the building. The constructed real-time digital twin can intuitively reflect the actual state of the target building in the real world, providing strong support for the real-time monitoring, decision analysis, and optimization of the building.

[0028] More specifically, the construction progress of the target building at different time periods is analyzed based on the digital twin information flow of past records, the patterns and trends are summarized, and the construction progress monitoring information of the target building is generated. Based on the construction progress monitoring information, time series analysis, machine learning prediction algorithms and other technologies are used to predict the construction progress at future time nodes, and the expected feedback of the key data collected at future time nodes is obtained.

[0029] More specifically, the expected feedback is fed back into the digital simulation model. By comparing the actual feedback with the expected feedback based on the key data, a deviation analysis is conducted. If the actual feedback deviates significantly from the expected feedback, it indicates that there may be problems in the construction process. The building appearance represented by the digital simulation model needs to be corrected, such as by adjusting parameters like size, shape, and material properties in the model, to ultimately obtain a more accurate reflection of the actual building appearance based on the key data.

[0030] More specifically, digital twin information flow records the status information at various moments during the construction process, containing rich construction patterns and trends. By analyzing this information, the current construction status can be more accurately assessed and predicted. By using expected feedback scenarios for deviation analysis, abnormal situations in the construction process can be detected in a timely manner, and the digital simulation model can be corrected, further improving the accuracy and reliability of the digital twin and providing a more effective decision-making basis for the quality control and schedule management of building construction.

[0031] This invention provides a method for constructing a digital twin throughout its entire lifecycle based on multi-source heterogeneous data, which has the following beneficial effects: In construction management, this invention can accurately grasp the construction progress by identifying key nodes and replicating models, facilitating timely adjustment. Matching multi-source monitoring data with expected reference information makes the scheduling digital simulation model more realistic, improving simulation accuracy. Based on the model, the data segmentation and interpretation and the deployment of parameters to construct a real-time digital twin can reflect the actual building status in real time. Combined with AI and the fusion of multi-source heterogeneous data, it can also assist in quality supervision, providing strong support for the quality and progress management of the entire building life cycle, and improving construction efficiency and quality.

[0032] Preferably, the steps of acquiring the original design data of the target building and identifying key nodes in the construction progress and reproducing the digital model based on the original design data to obtain expected reference information on the construction progress include: S11: Obtain the original design data of the target building, and construct digital models of the target building in the initial and completed states based on the original design data to obtain the initial digital model and the completed digital model. S12: Guided by the original design data, perform model deduction from the initial digital model toward the completed digital model, and record the model evolution process to obtain a forward deduction information flow; S13: Guided by the original design data, simulate and extrapolate the completed digital model toward the initial digital model, and record the model evolution process to obtain the reverse extrapolation information flow; S14: Combining the reverse deduction information flow and the forward deduction information flow, identify key nodes in the construction process of the target building to obtain several key construction nodes; S15: Perform digital model reproduction on each of the key construction nodes to obtain the expected reference model for each key construction node, which will be used together as expected reference information for construction progress.

[0033] Specifically, the original design data of the target building is obtained from the architectural design firm, construction company, or relevant document management system. This data typically includes architectural design drawings (such as floor plans, sections, and elevations), structural design documents, equipment installation drawings, construction schedules, and material lists. Using professional Building Information Modeling (BIM) software, such as Revit and Archicad, an initial digital model and a completed digital model of the target building are constructed based on the original design data. The initial digital model reflects the building's state at the start of construction, such as the terrain after site leveling and the initial state of the foundation. The completed digital model presents the building's state after all construction is completed and it is ready for use, including all building structures, interior decorations, equipment, and facilities. The initial and completed digital models provide clear start and end states for subsequent model derivation and key node identification, forming the basic framework for the entire digital twin construction process. The digital model constructed through BIM software can integrate various design information, forming a unified data standard, which facilitates subsequent data processing and analysis.

[0034] More specifically, guided by the construction schedule in the original design data, the initial digital model is gradually evolved according to the normal construction sequence. For example, starting from the foundation construction, the model sequentially simulates the main structure construction, roof construction, decoration and finishing construction, and other stages. During the model evolution, the process recording function of BIM software or the development of specialized data recording tools are used to record the changes in the model at each step, including newly added components, modified parameters, and completed construction tasks, forming a forward simulation information flow. Forward simulation can intuitively demonstrate the construction process of a building from scratch, helping construction and management personnel to better understand the construction process and schedule. During the simulation, potential problems in the construction process can be identified in advance, such as unreasonable construction sequence or component collisions, so that the construction plan can be adjusted in a timely manner.

[0035] More specifically, based on the original design data, starting from the completed digital model, the reverse process of building demolition or construction is simulated. For example, the decoration and finishing parts are removed first, then the main structure is demolished, and finally the foundation is removed. Using the same data recording method as the forward simulation, the changes in the model during the reverse simulation process are recorded to obtain the reverse simulation information flow. The reverse simulation can examine the rationality of the construction plan from another perspective, analyze the reversibility and operability of each construction step, and predict the response of the building structure in emergency situations (such as building demolition, renovation, etc.) through reverse simulation, providing a reference for emergency response.

[0036] More specifically, the forward and backward information flows are integrated to form a comprehensive set of construction process information. The integrated information is analyzed, and combined with the characteristics and experience of building construction, the time points or events that have a significant impact on the construction progress are identified as key construction nodes, such as the completion of foundation construction, the topping out of the main structure, and the completion of equipment installation and commissioning.

[0037] More specifically, key construction milestones are important milestones in the construction process. Identifying these milestones can help construction managers control the construction progress more accurately, detect progress deviations in a timely manner, and take corresponding measures to adjust them. By clearly defining key milestones, human, material, and financial resources can be allocated rationally, thereby improving resource utilization efficiency.

[0038] More specifically, based on the identified key construction nodes, the model status information of the corresponding nodes is extracted from the forward and reverse derivation information flow. Using BIM software or other modeling tools, the digital models of each key construction node are reconstructed based on the extracted information to ensure the accuracy and completeness of the model. The expected reference model provides a clear reference standard for the actual construction process. Construction personnel can compare the actual construction situation with the expected reference model to identify and correct differences in a timely manner. During the construction process, managers can evaluate and analyze the construction progress, quality, and cost based on the expected reference model to support decision-making.

[0039] Preferably, the steps of collecting multi-source monitoring data during actual construction and matching the multi-source monitoring data according to the expected reference information to schedule a specified expected reference model as the digital simulation model for the current construction cycle include: S21: Multi-source data is collected from the actual construction process through a multi-source sensing system pre-deployed at the construction site to obtain multi-source monitoring data; S22: Perform semantic fusion of heterogeneous data on the multi-source monitoring data to obtain fused identification information, and input the expected reference information into a pre-trained building progress identification algorithm to match the fused identification information and obtain the real-time progress features of the actual construction. S23: Determine the corresponding expected reference model from the expected reference information based on the real-time progress characteristics, and schedule the expected reference model as the digital simulation model for the current construction cycle.

[0040] Specifically, various types of sensors are pre-installed at the construction site to construct a multi-source sensing system. For example, cameras are installed to acquire images and video information of the construction site, monitoring the activities of construction workers, the stacking and use of materials; laser scanners are deployed to acquire three-dimensional spatial data of the building structure, allowing real-time monitoring of changes in the building's shape and size; stress sensors, displacement sensors, etc., are set up to monitor the mechanical properties and deformation of the building structure; and temperature and humidity sensors, light sensors, etc., are used to monitor environmental parameters at the construction site. Each sensor automatically collects data at set time intervals and transmits the data to a data acquisition center for storage and preliminary processing via wired or wireless communication networks. For example, cameras take images every few minutes, and stress sensors collect data every second.

[0041] More specifically, the construction process involves information from multiple aspects. A single type of sensor can only acquire limited information. Multi-source sensing systems can monitor the construction process from different angles and dimensions, thereby comprehensively and accurately reflecting the actual construction situation. Abundant multi-source monitoring data is the foundation for subsequent data fusion, progress identification, and model matching. Only by acquiring enough data can we more accurately analyze and judge the construction progress.

[0042] More specifically, since the data collected by different types of sensors have different formats, meanings, and expressions, semantic fusion of heterogeneous data is required. First, the multi-source monitoring data is cleaned to remove noise and erroneous data. Then, natural language processing, machine learning, and other technologies are used to perform semantic analysis on the data, converting different data into information with unified semantics to obtain fused recognition information. For example, the image information captured by the camera is converted into textual descriptions of the construction scene and then associated and integrated with data from other sensors.

[0043] More specifically, the previously obtained expected reference information is input into a pre-trained building progress recognition algorithm. This algorithm can be a deep learning-based model such as a convolutional neural network (CNN) or a recurrent neural network (RNN), or a traditional machine learning algorithm such as a support vector machine (SVM). The algorithm analyzes and matches the fused recognition information. By comparing the features of the fused recognition information with the expected reference information, it determines the real-time progress features of the actual construction. For example, based on information such as the number of floors of the building in the image, the completion status of the structure, and the mechanical properties monitored by the sensors, it determines which construction stage the current construction is in.

[0044] More specifically, heterogeneous data semantic fusion can eliminate the differences between data from different sensors, enabling data to be correlated and integrated, facilitating subsequent analysis and processing. The pre-trained construction progress recognition algorithm uses expected reference information as a benchmark, which can quickly and accurately identify the actual construction progress characteristics, providing a basis for scheduling appropriate expected reference models.

[0045] More specifically, based on the identified real-time progress characteristics of the actual construction, the corresponding expected reference model is searched from the expected reference information. For example, if the real-time progress characteristics indicate that the current construction is in the second stage of the main structure construction, the expected reference model corresponding to the completion of the second stage of the main structure construction is found from the expected reference information. The determined expected reference model is retrieved from the storage system as the digital simulation model of the current construction cycle. At the same time, the multi-source monitoring data is associated with the digital simulation model for subsequent data comparison and analysis.

[0046] More specifically, the expected reference model is constructed based on the key nodes of building construction and can accurately reflect the state of the building at different stages. By scheduling the appropriate expected reference model as the digital simulation model according to the actual construction progress, the digital simulation model can be matched with the actual construction situation, and more accurately simulate the building state in the current construction cycle. The real-time digital simulation model can provide strong support for real-time monitoring, decision analysis and optimization of building construction, and help construction managers to discover problems in a timely manner and take corresponding measures.

[0047] Preferably, the step of performing data segmentation and feedback content parsing of multi-source monitoring data based on the digital simulation model to deploy parameters of the digital simulation model and obtain a real-time digital twin includes: S31: Determine the current construction cycle of the target building based on the digital simulation model; S32: Analyze the target of the data feedback content of the multi-source monitoring data, and perform data segmentation on the multi-source monitoring data according to the data feedback target to obtain non-key data corresponding to the past construction cycle and key data corresponding to the current construction cycle; S33: Based on the digital simulation model, the feedback content of the key data is interpreted to obtain the actual building appearance reflected by the key data, so as to deploy parameters of the digital simulation model and obtain a real-time digital twin.

[0048] Specifically, the selected digital simulation model is analyzed in depth. This model contains detailed characteristics and status information of the building at different construction stages, such as the completion level of the building structure and the construction progress of each part. By sorting out and comparing this information, combined with the current actual time point, the construction cycle of the target building at the current moment can be determined. For example, if the digital simulation model shows that the main structure construction is half completed at a certain time point, and the current time corresponds to this, then it can be determined that the current stage is the middle stage of the main structure construction.

[0049] More specifically, the digital simulation model is linked to a pre-defined construction plan. The construction plan typically specifies the start and end times and key milestones for each construction phase. By comparing the status of the digital simulation model with the timeline of the construction plan, the current construction cycle can be more accurately determined. For example, if the construction plan stipulates that the foundation construction should be completed within 1-10 days, and the current digital simulation model shows that the foundation has been completed and is currently on day 8, then it can be determined that we are currently in the final stage of foundation construction.

[0050] More specifically, accurately determining the current construction cycle is a prerequisite for subsequent data processing and analysis. Different construction cycles have different construction content and focus points. Clarifying the current cycle can focus the analysis scope on the specific construction tasks currently underway, avoid interference from irrelevant data and information, improve the accuracy and efficiency of the analysis, and provide clear guidance for subsequent operations such as data segmentation and feedback content analysis. Only by determining the current construction cycle can we know which data is closely related to the current stage and which data can be temporarily ignored, thus enabling more targeted data processing.

[0051] More specifically, a detailed analysis of multi-source monitoring data is conducted to determine which construction cycle each data point relates to. For example, by analyzing the data's timestamp information, the associated construction location, and the data's trend, it can be determined whether the data reflects completed construction work from past cycles or ongoing construction in the current cycle. Data mining and machine learning algorithms can be used to train models to automatically identify the target of the data feedback. The target of the data feedback refers to which construction cycle the feedback content depicted by the multi-source monitoring data pertains to, i.e., whether it belongs to the current or previous construction cycle. It is evident that data whose target is a previous construction cycle does not require further detailed analysis. Based on the analysis results of the data feedback, the multi-source monitoring data is divided into non-key data corresponding to past construction cycles and key data corresponding to the current construction cycle. Two different data storage areas can be established to store the non-key and key data separately for convenient subsequent processing.

[0052] More specifically, throughout the construction process, multi-source monitoring data will continuously accumulate. If all data is processed indiscriminately, it will increase the complexity and time cost of data processing. By segmenting the data and focusing on the key data of concern in the current construction cycle, the amount of irrelevant data can be reduced, and the speed and efficiency of data processing can be improved. Separating the key data can more clearly highlight the critical information closely related to the current construction cycle, making it easier to conduct in-depth analysis and interpretation of this information and promptly identify problems and potential risks in the current construction process.

[0053] More specifically, based on the structure and parameters of the digital simulation model, in-depth analysis and interpretation of key data are conducted. For example, if the digital simulation model contains the strength parameters of the wall, and the key data contains the monitoring values ​​of the wall stress sensor, the stress situation of the wall and potential safety hazards can be analyzed by comparing the theoretical strength values ​​in the model with the actual monitoring values. Data analysis and visualization technologies are used to correlate and display the key data with the digital simulation model, intuitively presenting the actual situation of the building.

[0054] More specifically, based on the interpretation of key data, the relevant parameters of the digital simulation model are adjusted and deployed. For example, if the actual wall deformation exceeds the model's expected value, the elastic modulus, geometric dimensions, and other parameters of the wall in the model need to be adjusted so that the digital simulation model can more accurately reflect the actual state of the building. By updating the parameters of the digital simulation model, a real-time digital twin is obtained, which can reflect the real situation of the target building in the real world in real time.

[0055] More specifically, by interpreting key data and deploying model parameters, the digital twin can maintain a real-time mapping relationship with the target building in reality. The digital twin can accurately reflect the actual state of the building, including changes in structural deformation, mechanical performance, and environmental parameters, providing strong support for real-time monitoring and management of the building. The real-time digital twin provides accurate basis for decision-making during the building construction process. Construction managers can adjust construction plans and optimize resource allocation in a timely manner based on the actual situation reflected by the digital twin, thereby improving the quality and efficiency of building construction and reducing construction costs and risks.

[0056] Preferably, the digital twins are recorded in real time, and the digital twins recorded at each time point are arranged according to the temporal relationship to obtain a digital twin information flow. Based on the digital twin information flow, the digital simulation model is assisted in interpreting the feedback content of the key data.

[0057] Preferably, the step of using the digital twin information flow to assist the digital simulation model in interpreting the feedback content of the key data includes: S51: Based on the digital twin information flow, analyze the construction progress of the target building in each time period to generate construction progress monitoring information of the target building; S52: Based on the construction progress monitoring information, predict the construction progress at future time nodes to obtain the expected feedback of the key data collected at future time nodes. S53: Feed the expected feedback scenario back to the digital simulation model, and perform deviation analysis on the key data based on the expected feedback scenario to correct the building appearance represented by the digital simulation model, thereby obtaining the actual building appearance reflected by the key data.

[0058] Specifically, construction information corresponding to the digital twin at each recorded moment is extracted from the digital twin information flow, including the completion status of the building structure, the progress of each construction process, and resource usage. Data analysis tools and algorithms are used to sort and analyze this information to determine the specific construction progress of the target building in each time period. For example, how many more floors of the wall were built in a certain time period, what is the completion rate of equipment installation, etc. The construction progress information obtained from each time period is integrated and presented in the form of charts, reports, etc., to form the construction progress monitoring information of the target building. Line graphs can be used to show the trend of the construction progress of each part of the building over time, and tables can be used to list the actual start time, end time, and completion status of key construction processes.

[0059] More specifically, by analyzing the information flow of digital twins to generate construction progress monitoring information, it is possible to comprehensively and systematically understand the entire construction process of the target building from the start of construction to the present moment, clearly grasp the construction progress of each time period, and provide basic data for subsequent progress prediction and decision-making. Construction progress monitoring information can help managers to promptly identify potential problems in the construction process, such as delays in certain construction procedures or unreasonable resource allocation. By promptly identifying and addressing these problems, it is possible to ensure that the construction project proceeds smoothly according to plan.

[0060] More specifically, based on the generated construction progress monitoring information, appropriate prediction methods and models are selected, such as time series analysis models (e.g., ARIMA model, exponential smoothing model), machine learning models (e.g., neural network model, decision tree model), etc. The historical data in the construction progress monitoring information is used as the training set to train and optimize the model so that it can learn the changing patterns and trends of construction progress.

[0061] More specifically, by using a trained prediction model and combining it with the current construction status and construction plan, the construction progress at future time points can be predicted. For example, it can predict which floor the main structure will be completed within the next week, whether a certain equipment installation process can be completed on time, etc. Based on the predicted construction progress, the expected feedback of key data collected at future time points can be inferred, such as the stress value of the wall and the reading of the temperature sensor at a certain time in the future.

[0062] More specifically, forecasting future construction progress and data feedback can provide forward-looking guidance for construction management. Managers can plan resource allocation, arrange construction personnel and equipment in advance based on the forecast results, and avoid delays caused by insufficient resources or improper arrangements. By forecasting the expected feedback of future data, potential risks and problems can be identified in advance. For example, if it is predicted that the stress value of the wall will exceed the safe range at a certain point in the future, measures can be taken in advance to reinforce or adjust the construction plan to reduce safety risks.

[0063] More specifically, the expected feedback scenarios of the key data at predicted future time points are fed into the digital simulation model, enabling the model to simulate the building's state and appearance under the expected conditions. This provides a basis for comparison in subsequent deviation analysis. The actual collected key data is compared with the results shown in the digital simulation model based on the expected feedback scenarios, and the deviation between the two is calculated. For example, the actual monitored wall stress values ​​are compared with the expected stress values ​​in the model to analyze the magnitude and causes of the deviation. Statistical analysis methods, such as calculating the deviation rate and standard deviation, can be used to quantify the degree of deviation.

[0064] More specifically, based on the results of the deviation analysis, the building appearance represented by the digital simulation model is corrected. If the actual stress value is found to be larger than the expected value, the material properties and structural parameters of the walls in the model need to be adjusted so that the model can more accurately reflect the actual situation of the building and obtain the actual building appearance reflected by the key data.

[0065] More specifically, through deviation analysis and model correction, the digital simulation model can be continuously adjusted to better match the actual construction situation, thereby improving the accuracy and reliability of the digital simulation model. An accurate digital simulation model can provide more effective support for monitoring, analysis, and decision-making during the construction process. Timely detection and correction of deviations between the digital simulation model and the actual situation can ensure more accurate assessment and judgment of the building's condition. This helps to promptly identify quality problems and safety hazards during the construction process, allowing for appropriate measures to be taken to address them and ensuring the construction quality and safety of the building.

[0066] Preferably, the quality performance of the target building is inferred based on the original design data to obtain the theoretical quality performance characteristics of the target building, and the quality performance of the multi-source monitoring data is monitored based on the theoretical quality performance characteristics to obtain the quality monitoring information of the target building.

[0067] Specifically, information on various materials used in the building is extracted from the original design data, including the type, specifications, and performance indicators of the materials. For example, for concrete, its design strength grade and durability requirements are analyzed; for steel, parameters such as yield strength and tensile strength are examined. Based on relevant material standards and specifications, and in conjunction with design requirements, the ideal quality characteristics that these materials should possess are determined. Using the architectural structural drawings and mechanical models in the design data, structural mechanics analysis is conducted. Through professional structural analysis software, the stress conditions of the building under different load conditions are simulated, and the mechanical responses such as internal forces and deformations of structural components are calculated. For example, the stress and displacement of beams, columns, and other components under vertical loads and horizontal seismic actions are calculated to determine the structural bearing capacity and stability, among other quality indicators.

[0068] More specifically, referring to the construction techniques and methods specified in the design documents and comparing them with the corresponding construction quality acceptance specifications, the quality standards that each construction procedure should meet are clarified. For example, in waterproofing projects, the thickness of the waterproofing layer, construction process requirements, and inspection standards for the waterproofing effect are determined based on the waterproofing grade required by the design. These standards are then integrated to form the theoretical quality performance characteristics of the target building.

[0069] More specifically, theoretical quality performance characteristics provide a benchmark for the quality assessment of a target building. They clarify the quality level that the building should achieve at the design level and serve as an important basis for subsequent quality supervision and judgment of the actual quality status. Based on these theoretical quality characteristics, the construction team can formulate specific construction plans and quality control measures. At the same time, managers can conduct targeted management and supervision of the construction process based on these characteristics to ensure that construction activities meet design requirements.

[0070] More specifically, multi-source monitoring data is correlated with theoretical quality performance characteristics. For each monitoring data point, a corresponding theoretical quality index is found. For example, temperature monitoring data during concrete pouring is compared with the temperature range specified in concrete construction specifications; deformation monitoring data of structural components is compared with the allowable deformation value calculated by structural mechanics.

[0071] More specifically, by setting reasonable thresholds and judgment rules, multi-source monitoring data are analyzed to identify abnormal data that deviate from the theoretical quality performance characteristics. For example, if the monitored concrete strength is lower than a certain proportion of the design strength grade, or the deformation of structural components exceeds the allowable value, it is marked as abnormal data.

[0072] More specifically, for the identified abnormal data, a comprehensive evaluation and in-depth analysis should be conducted, taking into account the influence of multiple factors, such as construction technology, environmental conditions, and material quality, to determine whether the abnormal data is caused by accidental factors or reflects potential quality problems. Data analysis methods, such as correlation analysis and trend analysis, can be used to help determine the cause and degree of impact of the abnormal data.

[0073] More specifically, by monitoring multi-source monitoring data, we can understand the quality status of the target building in real time during the construction process, promptly identify quality deviations and potential problems, and take timely and effective measures to correct and deal with them, so as to avoid further deterioration of quality problems. Quality supervision is an important means to ensure the quality and safety of buildings. By supervising the monitoring data based on the theoretical quality performance characteristics, we can ensure that the building always moves towards the quality goals required by the design during the construction process, thereby improving the reliability and safety of the building.

[0074] More specifically, the monitoring results from multi-source monitoring data are organized and summarized, including statistical information on normal data, detailed information on abnormal data (such as the time, location, and degree of deviation), and analytical conclusions on the abnormal data. The organized information is then used to generate a quality monitoring report in a clear and intuitive manner. The report content may include an overview of the quality status, analysis of abnormal situations, assessment of potential quality risks, and corresponding recommendations and measures. Various formats, such as charts, tables, and text descriptions, can be used to make the report easy to understand and use.

[0075] More specifically, quality supervision information should be promptly fed back to relevant personnel, such as construction teams, supervision units, and construction units. By establishing an effective information sharing mechanism, it can be ensured that all parties can understand the quality status of the building in a timely manner and jointly participate in quality control and management.

[0076] More specifically, quality supervision information provides crucial information for decision-making during the construction process. Construction teams can adjust construction plans and quality control measures based on this information; supervision units can strengthen oversight of key stages; and construction companies can assess and make decisions regarding the overall project quality. Through feedback and sharing of quality supervision information, all parties can jointly analyze the causes of quality problems and take targeted measures for improvement. Continuously summarizing experiences and lessons learned will enhance the quality level and management capabilities of building construction.

[0077] Reference Figure 2 As shown, in a second aspect, the present invention provides a full lifecycle digital twin construction system based on multi-source heterogeneous data, used to implement the full lifecycle digital twin construction method based on multi-source heterogeneous data as described in any one of the first aspects, comprising: The expected reference module is used to acquire the original design data of the target building, and to identify key nodes of the construction progress of the target building and reproduce the digital model based on the original design data, so as to obtain expected reference information on the construction progress. The digital simulation module is used to collect multi-source monitoring data during actual construction and match the multi-source monitoring data according to the expected reference information to schedule the specified expected reference model as the digital simulation model for the current construction cycle. The content feedback module is used to perform data segmentation and feedback content parsing of multi-source monitoring data based on the digital simulation model for the current construction cycle, so as to deploy parameters of the digital simulation model and obtain a real-time digital twin.

[0078] In this embodiment, the specific implementation of each module in the above system embodiment is described in the above method embodiment, and will not be repeated here.

[0079] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for constructing a full lifecycle digital twin based on multi-source heterogeneous data, characterized in that, include: The original design data of the target building is obtained, and the key nodes of the construction progress of the target building are identified and digitally modeled based on the original design data to obtain the expected reference information of the construction progress. Collect multi-source monitoring data during actual construction, and match the multi-source monitoring data according to the expected reference information to schedule the specified expected reference model as the digital simulation model for the current construction cycle; Based on the digital simulation model, the data of the current construction cycle is segmented and the feedback content is analyzed to deploy parameters of the digital simulation model and obtain a real-time digital twin.

2. The method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in claim 1, characterized in that, The steps of obtaining the original design data of the target building, and identifying key nodes in the construction progress and reproducing the digital model based on the original design data to obtain expected reference information on the construction progress include: Obtain the original design data of the target building, and construct digital models of the target building in its initial and completed states based on the original design data to obtain the initial digital model and the completed digital model. Guided by the original design data, the initial digital model is deduced towards the completed digital model, and the model evolution process is recorded to obtain a forward deduction information flow; Guided by the original design data, the completed digital model is simulated and deduced to resemble the initial digital model, and the model evolution process is recorded to obtain the reverse deduction information flow; By combining the reverse deduction information flow and the forward deduction information flow, key nodes in the construction process of the target building are identified, resulting in several key construction nodes; Each of the key construction nodes is reproduced using a digital model to obtain a corresponding expected reference model for each key construction node, which together serve as expected reference information for the construction progress.

3. The method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in claim 1, characterized in that, The steps of collecting multi-source monitoring data during actual construction and matching the multi-source monitoring data according to the expected reference information to schedule a specified expected reference model as the digital simulation model for the current construction cycle include: Multi-source monitoring data is obtained by pre-deploying multi-source sensing systems at the construction site to collect multi-source data on the actual construction process. The heterogeneous data of the multi-source monitoring data are semantically fused to obtain fused identification information, and the expected reference information is input into a pre-trained building progress identification algorithm to match the fused identification information and obtain the real-time progress features of the actual construction. Based on the real-time progress characteristics, the corresponding expected reference model is determined from the expected reference information, and the expected reference model is scheduled as the digital simulation model for the current construction cycle.

4. The method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in claim 1, characterized in that, The steps for performing data segmentation and feedback content analysis on multi-source monitoring data based on the digital simulation model to deploy parameters of the digital simulation model and obtain a real-time digital twin include: The construction cycle of the target building at the current moment is determined based on the digital simulation model; The multi-source monitoring data is parsed to determine the target of the data feedback content. Based on the parsing results, the multi-source monitoring data is segmented to obtain non-key data corresponding to past construction cycles and key data corresponding to the current construction cycle. Based on the interpretation of the feedback content of the key data based on the digital simulation model, the actual building appearance reflected by the key data is obtained, and parameters are deployed on the digital simulation model to obtain a real-time digital twin.

5. The method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in claim 4, characterized in that, Real-time digital twins are recorded, and the digital twins recorded at each time point are arranged according to the temporal relationship to obtain a digital twin information flow. Based on the digital twin information flow, the digital simulation model is assisted in interpreting the feedback content of the key data.

6. The method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in claim 5, characterized in that, The steps for interpreting the key data based on the digital twin information flow to assist the digital simulation model include: Based on the digital twin information flow, the construction progress of the target building in each time period is analyzed to generate construction progress monitoring information of the target building; Based on the construction progress monitoring information, the construction progress at future time nodes is predicted in order to obtain the expected feedback of the key data collected at future time nodes. The expected feedback scenario is fed back to the digital simulation model, and a deviation analysis is performed on the key data based on the expected feedback scenario to correct the building appearance represented by the digital simulation model, thereby obtaining the actual building appearance reflected by the key data.

7. The method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in claim 1, characterized in that, Based on the original design data, the quality performance of the target building is inferred to obtain the theoretical quality performance characteristics of the target building. Based on the theoretical quality performance characteristics, the quality performance of the multi-source monitoring data is monitored to obtain the quality monitoring information of the target building.

8. A full lifecycle digital twin construction system based on multi-source heterogeneous data, characterized in that, A method for constructing a full lifecycle digital twin based on multi-source heterogeneous data as described in any one of claims 1-7 includes: The expected reference module is used to acquire the original design data of the target building, and to identify key nodes of the construction progress of the target building and reproduce the digital model based on the original design data, so as to obtain expected reference information on the construction progress. The digital simulation module is used to collect multi-source monitoring data during actual construction and match the multi-source monitoring data according to the expected reference information to schedule the specified expected reference model as the digital simulation model for the current construction cycle. The content feedback module is used to perform data segmentation and feedback content parsing of multi-source monitoring data based on the digital simulation model for the current construction cycle, so as to deploy parameters of the digital simulation model and obtain a real-time digital twin.

Images