BIM (Building Information Modeling) intelligent management platform and method for project construction full life cycle

Abstract

The invention provides a BIM intelligent management platform oriented to a whole life cycle of project construction. A building information model, Internet of Things sensing data and a block chain evidence storage mechanism are integrated through a multi-source data fusion technology, and a whole-process data chain of association planning, design, construction, operation and maintenance is associated. The platform adopts space optimization Huffman coding to realize model lightweight, combines a constraint genetic algorithm to optimize a construction path, and applies a bidirectional long-short-term memory network to analyze an equipment state. A three-chain block chain system is reconstructed on the architecture, intelligent association of engineering quantity and payment nodes is realized through cooperation of a main chain, a calculation quantity side chain and an auditing side chain, and mobile terminal offline interaction is supported based on a digital-analog separation technology. The platform covers an intelligent design management unit, a block chain investment management unit, a dynamic correction management unit, a quality safety responsibility tracing unit, an NLP risk management unit and a digital twin operation and maintenance unit. The units achieve cross-system cooperation through a unified data bus, and a closed-loop management architecture covering the whole life cycle of project construction is formed.

Description

Technical Field

[0001] The present invention belongs to the field of construction engineering information technology, and specifically relates to a digital management platform and method for the entire life cycle of construction projects that integrates building information modeling (BIM), the Internet of Things (IoT), artificial intelligence (AI) and blockchain. Background Art

[0002] The core challenges currently facing project management for construction contractors profoundly impact project quality, schedule, and cost control, primarily manifesting in the following key areas: Data fragmentation throughout the entire lifecycle. In typical engineering projects, BIM models, created during the design phase, often become static data silos. After the design institute completes model delivery, real-time data generated during the construction phase (such as concrete pouring temperature and steel structure installation deviations) and equipment operating data during the maintenance phase (such as energy consumption and vibration frequency) cannot be dynamically linked to the original BIM model. While platforms like Autodesk BIM 360 enable multi-disciplinary collaboration, their IoT integration capabilities have significant limitations. Dynamic control faces a lag in response: Traditional deviation management relies on a linear, manual inspection-reporting-decision-making process, resulting in delayed responses to critical issues. Audit and traceability mechanisms are inefficient, with project payment audits still heavily reliant on paper-based processes. Key technical support bottlenecks exist: existing technologies suffer from multiple capabilities, including lightweight model limitations, limited AR positioning accuracy, a lack of real-time deviation correction mechanisms, the need for rescanning when deviations exceed limits, insufficiently optimized deviation correction paths, and unclear accountability.

[0003] Existing BIM management technologies in the construction engineering sector face several challenges: Regarding model lightweighting and interactivity, WebGL engines generally use general compression algorithms (such as octree spatial partitioning), resulting in compression rates of only 40-50%. Loading a 1GB Revit model on a mobile device takes 8-12 minutes, requiring full download of non-geometric data and lacking offline functionality. This inefficient loading prevents real-time operation of large models on mobile devices, delaying on-site decision-making. Geometric and non-geometric data are coupled and stored separately, making them difficult to access on demand. Regarding project audits, blockchains often employ a dual-chain structure (main chain + business chain), with audit reports generated manually. The correlation between BIM project quantities and payment nodes suffers from large errors, resulting in long average audit cycles. Payment verification requires manual comparison of drawings and contracts, and audit logs are stored unencrypted, exposing them to post-event tampering. Regarding quality traceability and maintenance, RFID is used to identify components but is not linked to contractual liability clauses. Energy consumption forecasts rely solely on historical data, ignoring real-time equipment status. Consequently, determining responsibility for quality incidents takes an average of 24 hours. Existing technology systems lack an integrated platform that links design models, construction data, and maintenance information. Summary of the Invention

[0004] This invention provides a BIM intelligent management platform for the entire lifecycle of engineering construction. Through three key technical approaches: multi-source data fusion (BIM+IoT+blockchain), algorithmic innovation (constrained genetic algorithm / Huffman space coding), and architectural reconstruction (triple-chain blockchain / digital-analog separation), it systematically addresses core issues in the engineering construction sector, such as data fragmentation, control lag, and inefficient traceability. This multi-technology fusion integrates cutting-edge technologies such as a WebGL lightweight engine, augmented reality (AR), blockchain, the Internet of Things (IoT), and LSTM neural networks. This platform enables efficient and lightweight processing of BIM models, mobile-end interaction, and deep fusion and visualization of multi-source data, providing powerful technical support for the management of engineering construction throughout its lifecycle. Utilizing intelligent algorithms such as Monte Carlo simulation, genetic algorithms, and fuzzy comprehensive evaluation, it accurately predicts and optimizes construction deviations and risks, enhancing the scientific nature of engineering management and the accuracy of decision-making. Comprehensively covering all stages of engineering construction projects, including planning, design, construction, and operation and maintenance, it enables digital management throughout the entire lifecycle, breaking down information silos between stages and improving overall project management efficiency. The various functional units are organically integrated and work closely together. For example, the intelligent design management unit cooperates with the blockchain investment management unit and the dynamic correction management unit to ensure seamless connection and collaborative work in all aspects of the project, including design, investment, and construction, forming a complete management closed loop.

[0005] The present invention is realized through the following technologies: a BIM intelligent management platform for the entire life cycle of engineering construction, consisting of a small front-end module and a large back-end module.

[0006] The small front-end module uses a WebGL lightweight engine and a specific compression algorithm (such as Draco compression + entropy coding) to achieve lossless conversion of multi-source BIM models (Revit / CATIA) and mobile terminal interaction, improve loading efficiency, and integrate an augmented reality (AR) data console to overlay construction progress and quality inspection data onto the 3D real-life model in real time.

[0007] The large backend module includes: an intelligent design management unit, which realizes multi-disciplinary collaborative design based on the BIM-GIS fusion model, identifies code conflicts through an automatic collision detection algorithm (recognition rate ≥ 95%), and the collision detection algorithm supports BIM models with millions of triangles, and uses VR virtual reality technology to verify the feasibility of construction plans.

[0008] The blockchain investment management unit links BIM project quantities with contract payment nodes through smart contracts, with an error rate of ≤1%. It generates structured audit reports and records fund flows, shortening the audit cycle.

[0009] The dynamic deviation correction management unit combines IoT sensors to collect construction deviation data in real time, predicts the impact of deviations through Monte Carlo simulation, combines genetic algorithms with the movement path of construction machinery as the optimization variable, and visualizes the impact range of construction deviation data in the BIM-GIS model.

[0010] The quality and safety responsibility traceability unit uses RFID tags to bind the construction process ID and responsible person information, builds a quality and safety responsibility map, and supports scanning codes to retrieve full-process construction records.

[0011] The NLP risk management unit extracts keywords from contract texts through natural language processing, builds an evaluation model based on multi-dimensional risk indicators (such as weather data and construction period data), adopts a fuzzy comprehensive evaluation indicator system (such as risk probability, impact level, and response cost), quantifies the risk level through a comprehensive evaluation method, and generates an emergency plan library. The risk assessment response time for a single contract is ≤1 minute.

[0012] The digital twin operation and maintenance unit creates a virtual replica corresponding to the actual construction by digitally modeling the construction and its equipment and facilities. It not only contains the static geometric information of the building, but also includes dynamic data such as the operating parameters and real-time status of the equipment dynamically corrected by the virtual model based on the real-time posture feedback of the construction machinery; it provides a complete workflow from data collection, transmission, processing to final visualization and decision support.

[0013] The WebGL lightweight engine adopts digital-analog separation technology, and entropy coding uses Huffman coding to optimize geometric data flow. Geometric data is triangulated and compressed for storage. Non-geometric data is dynamically called through a cloud API, supporting offline browsing and annotation functions on mobile devices.

[0014] The AR data console uses a lidar point cloud matching algorithm to achieve millimeter-level precision (≤3mm) alignment between the BIM model and the construction site, uses an IMU (inertial measurement unit) to compensate for point cloud matching delays, and triggers an audible and visual alarm when the deviation exceeds a threshold.

[0015] The blockchain investment management unit adopts a three-chain architecture (main chain + calculation side chain + audit side chain). The main chain stores contract payment vouchers, the side chain records BIM calculation data, and the audit side chain uses Merkle Patricia Tree to store audit report hash values, enhancing data immutability and achieving privacy protection through zero-knowledge proof technology.

[0016] The digital twin operation and maintenance unit analyzes historical energy consumption data through an LSTM neural network. The LSTM neural network adopts a BiLSTM (bidirectional long short-term memory network) model. The BiLSTM model is input to predict the peak energy consumption curve for the next 7 days. The device operating status data (such as voltage and temperature) and historical energy consumption data are integrated through the spatiotemporal alignment layer, and the device operation strategy is dynamically adjusted.

[0017] A BIM intelligent management method for the entire life cycle of engineering construction includes the following steps: S100, Data Integration and Lightweight Processing: Import multi-source BIM models (Revit / CATIA), optimize the code table generation algorithm based on Huffman coding for the spatial locality of BIM models, use the WebGL engine to triangulate and compress geometric data and store non-geometric data in the cloud, and improve model loading efficiency by ≥60% for BIM models with less than 100,000 triangles; generate lightweight models that are interactive on mobile devices based on digital-analog separation technology.

[0018] S200, real-time data acquisition and fusion: Construction progress and quality inspection data are collected through IoT sensors and matched with the BIM model through lidar point cloud to achieve AR augmented reality real-scene overlay with an alignment accuracy of ≤3mm; GNSS-RTK positioning data is corrected using a six-axis attitude sensor to generate construction dynamic coordinates in the ECEF coordinate system.

[0019] S300, Intelligent Decision-making and Execution: Collect historical construction project data and real-time construction data, use data mining and machine learning algorithms to analyze and model the data, train LSTM neural networks based on historical project data, predict investment risks (material price fluctuations) and construction period deviations (critical path optimization), and thus predict project progress, cost and quality risks, providing data support for decision-making. The response time for BIM models with less than 100,000 triangles is ≤2 seconds; combine BIM models with intelligent decision-making systems, and use the visualization and information advantages of BIM models to provide intuitive project information and 3D model support for decision-making; use data in BIM models to interact with intelligent decision-making algorithms, and automatically associate BIM project quantities with contract payment nodes through smart contracts, with an error rate of ≤1%, and shortening the time to generate audit reports by 50%.

[0020] S400, dynamic correction and visual feedback: Based on real-time monitoring data, Monte Carlo simulation is used to analyze deviations, and corresponding measures are taken to adjust and correct them to ensure that the system output is consistent with the preset target or minimize the deviation process, generate the optimal correction path and visualize it through the BIM-GIS model, and present the operation status and data information of the construction process to users or operators in an intuitive and easy-to-understand visual form, so that they can quickly and accurately understand the construction situation, deviation status and correction effect, etc., thereby providing a basis for decision-making and operation. When the deviation exceeds the threshold, the AR sound and light alarm is triggered and the correction instruction is pushed to the responsible person's terminal.

[0021] S500, quality and safety responsibility definition and traceability: During the construction process, RFID tags are used to bind the construction process ID and the responsible person information, build a quality and safety responsibility map, and collect and record information on construction processes, materials, equipment, personnel, etc. in real time. Each RFID tag uniquely corresponds to a construction process or a material batch, recording its related responsible person, construction time, construction process parameters and other detailed information; the collected information is stored in the database to form a complete construction process data chain. The data includes construction logs, inspection reports, hidden project acceptance records, material incoming inspection records, safety inspection records, safety accident reports, etc., to ensure that every construction link and material use is traceable; the management platform is used to bind each process and element in the construction process with the corresponding responsible parties, including construction units, construction teams, specific operators, material suppliers, supervision units, etc., to clarify the quality and safety responsibilities of each link. When a quality problem or safety incident occurs, the database can be searched and relevant data analyzed to quickly locate the specific link and time point where the problem occurred. Based on the pre-set responsibility binding relationship and the responsibility weight distribution based on FMEA (Failure Mode Analysis), the relevant responsible persons and responsible units can be traced back. At the same time, relevant construction parameters, environmental factors and other information before and after the quality problem or safety incident can be viewed to provide a basis for problem cause analysis. Quality issues and traceability results are promptly fed back to the relevant responsible persons and management, urging the responsible parties to take corrective measures, repair or rework quality defects, properly handle safety incidents, and impose corresponding penalties or education on relevant responsible persons to prevent similar problems from recurring. The system supports scanning codes to retrieve full-process construction records, including quality and safety-related records, for rapid traceability and verification.

[0022] S600, full life cycle operation and maintenance management: Based on digital twin technology simulation, data from the entire life cycle of construction projects is collected, analyzed and utilized to achieve comprehensive monitoring, intelligent diagnosis and optimized decision-making of the construction process, so as to achieve efficient operation, cost control, quality assurance and sustainable development of the project; advanced algorithms are used to analyze and process data (such as predictive analysis of energy consumption data through machine learning algorithms, anticipating possible equipment failures in advance, and providing support for operation and maintenance decisions); the analysis results are displayed to operation and maintenance personnel through a visual interface, allowing them to intuitively understand the operating status of the building, and through the remote control function of the digital twin model, construction operations are optimized to improve work efficiency and management level.

[0023] S700, Engineering management data classification and storage: Project management materials should be categorized according to different project stages, such as pre-planning, design, construction, completion acceptance, and operation and maintenance. Materials should also be categorized based on the different disciplines involved, such as construction, structural engineering, plumbing, electrical engineering, and HVAC. Materials should be categorized based on their content and purpose, such as administrative documents, technical documents, quality control documents, schedule control documents, cost control documents, and contract documents. Materials should also be categorized by information format, such as electronic files, images, and audio and video materials. A professional file management system should be used to centrally store and manage project management materials. Structured data, such as bills of quantities, quality inspection data, and schedule data, should be stored in a database system. Using cloud storage platforms, electronic files, images, audio and video materials should be uploaded to cloud storage and access rights should be set. Data encryption technology should be implemented to encrypt sensitive data stored in file management systems and cloud storage services. Important paper-based project materials, such as completion records and original contracts, should be stored in dedicated archives. A detailed index should be established for project management materials, including file directories, keyword indexes, and category indexes. Data sharing and collaboration: Share data through file management systems and cloud storage platforms, view and edit files in real time, and improve work efficiency. Set access control policies to limit access rights to data by different users, ensuring that only authorized personnel can view, edit, and use specific files. Perform data analysis and mining on stored engineering management data to extract valuable information and knowledge, providing decision support for project management.

[0024] In step S100, the digital-analog separation technology includes: compressing geometric data using the Draco algorithm, combining data flow optimization based on spatial locality Huffman coding, dynamically calling non-geometric data through a RESTful API interface, and supporting offline annotation and version rollback.

[0025] In step S200, the lidar point cloud matching includes: aligning the BIM model and the construction site point cloud data through the ICP iterative closest point algorithm, and removing noise points using the RANSAC algorithm, with a matching error of ≤3mm.

[0026] In step S300, the smart contract adopts a three-chain structure (main chain + calculation side chain + audit side chain). The main chain stores fund payment vouchers, the side chain records calculation data, and the audit side chain uses Merkle Patricia Tree to store the audit report hash value; data consistency is verified through zero-knowledge proof technology to ensure audit transparency.

[0027] In step S500, the construction of the quality responsibility map includes: associating failure modes with responsible units based on the construction process log collected by RFID, quantifying responsibility weights, associating the responsibility weights of the design unit, construction unit, and supervision unit, extracting responsibility clauses in the contract text through NLP, and generating a responsibility traceability path.

[0028] The beneficial effects of the present invention are as follows: with the help of the powerful computing power and big data processing technology of cloud computing, massive amounts of engineering data can be quickly stored, managed and analyzed, providing accurate data support for real-time monitoring and intelligent decision-making of projects. Various data from the construction site are collected in real time through IoT sensors and fed back to the system for processing and analysis in a timely manner, achieving real-time control and rapid response to project progress, quality, safety, etc., ensuring that the project proceeds smoothly as planned. Based on historical project data, models such as LSTM neural networks are trained to predict and analyze investment risks, construction period deviations, etc., providing scientific and accurate decision-making basis for project management and reducing decision-making risks. NLP technology is used to extract contract text keywords and build a risk assessment model, quantify risk levels and generate an emergency plan library, provide strong support for risk prevention and control during project implementation, and enhance the project's ability to cope with risks. RFID tags and other technologies are used to bind construction process IDs and responsible person information, build a quality responsibility map, achieve rapid tracing of quality problems and definition of responsibilities, ensure project quality and promote the improvement of quality awareness of all participants. Leveraging blockchain technology's smart contracts and distributed ledgers, the project data's authenticity and immutability are ensured. Zero-knowledge proof technology also protects data privacy, providing a trusted and reliable data foundation for project investment management and auditing. Deviations during construction are monitored and corrected in real time, reducing costs associated with construction errors. Smart contracts link BIM project quantities with contract payment nodes, improving investment control accuracy and reducing project costs. The digital twin operation and maintenance unit integrates IoT data from building equipment with BIM models to predict equipment aging trends and optimize energy consumption, improving energy efficiency and achieving energy conservation and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a system architecture diagram of the present invention; Figure 2 Design management flow charts for intelligence; Figure 3 This is a diagram of the three-chain architecture of the blockchain investment management unit; Figure 4 A sequence diagram of the blockchain engineering payment verification process; Figure 5 This is the architecture diagram of the dynamic correction management unit; Figure 6 This is the structure diagram of the quality and safety responsibility traceability unit; Figure 7 This is the NLP risk management unit architecture diagram; Figure 8 This is the architecture diagram of the digital twin operation and maintenance unit. DETAILED DESCRIPTION

[0030] In order for those skilled in the art to better understand the present invention, Figures 1 to 8 To further clarify this application, terms such as "architecture," "process," and "algorithm" in this description have standard definitions commonly recognized by those skilled in the art. These terms are used to simplify the description of the present invention and do not imply that the architectures, processes, or algorithms involved must be arranged, constructed, or operated in a specific order. The embodiments described herein are merely illustrative and do not constitute any limitation on the present invention.

[0031] The BIM intelligent management platform for the entire life cycle of engineering construction of the present invention is composed of a small front-end module and a large back-end module. The system architecture diagram is shown in Figure 1In the technical architecture of an engineering management system, the small front-end module refers to a lightweight, focused front-end submodule. Its core goal is to enable smooth interaction and real-time data integration of complex engineering models on mobile or web devices through efficient data processing and rendering technologies. The small front-end module utilizes a lightweight WebGL engine and specific compression algorithms (such as Draco compression + entropy coding) to reduce the amount of geometric data. Server-side conversion generates web-friendly formats (such as glTF), enabling lossless conversion and mobile interaction of multi-source BIM models (Revit / CATIA), improving loading efficiency. The lightweight WebGL engine utilizes digital-analog separation technology and entropy coding using Huffman coding to optimize the geometric data flow. Geometric data is triangulated and compressed for storage, while non-geometric data is dynamically accessed via a cloud API, supporting offline browsing and annotation on mobile devices. The small front-end module integrates an augmented reality (AR) data console, overlaying construction progress and quality inspection data onto 3D real-world models in real time. The AR data console integrates augmented reality (AR) technology for centralized management and display of AR-related data. In smart building management, the AR data console can overlay construction progress, quality inspection, and other data onto 3D models in real time, improving management efficiency and decision-making accuracy. Augmented reality technology overlays virtual information onto the real world, while the AR data console is used for centralized data management, analysis, and presentation. The AR data console uses a LiDAR point cloud matching algorithm to align the BIM model with the construction site with millimeter-level accuracy (≤3mm). It also employs an IMU (Inertial Measurement Unit) to compensate for point cloud matching delays, triggering audible and visual alarms when deviations exceed a threshold.

[0032] The large backend module includes: intelligent design management unit, blockchain investment management unit, dynamic correction management unit, quality and safety responsibility traceability unit, NLP risk management unit, and digital twin operation and maintenance unit.

[0033] The intelligent design management unit, based on a BIM-GIS fusion model, enables multi-disciplinary collaborative design and is a core module of modern digital engineering management. It utilizes a closed-loop technology framework: a BIM-GIS fusion engine, layered clash detection, and VR dynamic verification. Based on a data fusion mechanism, and using a unified spatial coordinate system (e.g., WGS84 to engineering Cartesian coordinate conversion), it links the building structural geometric attributes (dimensions and materials) of BIM models (such as Revit / CATIA) with GIS geographic information (topography, geology, and pipeline networks). This enables multi-disciplinary collaborative design and construction plan optimization, achieving "zero-clash, high-coordination, and rehearsal" engineering design. An automatic clash detection algorithm (with a recognition rate of ≥95%) identifies regulatory conflicts, supporting BIM models with millions of triangles, and utilizes VR technology to verify the feasibility of construction plans. Its value lies not only in improving single-point efficiency (e.g., a clash detection rate of >95%), but also in reshaping the collaborative model across the entire engineering supply chain, providing a core support platform for intelligent construction.

[0034] The intelligent design management process is as follows: VR verification scheme starts the process, simulates and verifies the construction scheme, and outputs a feasibility report; the blockchain unit receives the feasibility report and ensures the safe storage and verification of the report; the VR verification scheme also outputs the optimized model to the BIM-GIS fusion model module; the BIM-GIS fusion model generates fusion data and outputs it to the data lake in IFC format for storage; the BIM-GIS fusion model data is used to generate a collision detection report, which is completed by the collision detection report module; the collision detection report transmits the conflict location information to the dynamic correction unit for real-time adjustment and optimization. The intelligent design management process diagram is shown in Figure 2 .

[0035] The blockchain investment management unit is an intelligent investment management platform built on blockchain distributed ledgers, smart contracts, and encryption technology. It reconstructs the traditional investment process through decentralization, transparency, and automation. The blockchain investment management unit of the present invention adopts a three-chain architecture (main chain + calculation side chain + audit side chain). The main chain stores contract payment vouchers, the side chain records BIM calculation data, and the audit side chain uses Merkle Patricia Tree to store audit report hash values, enhancing data immutability and achieving privacy protection through zero-knowledge proof technology. The three-chain architecture diagram of the blockchain investment management unit is shown in Figure 3The investment decision-making main chain is used for investment decision-making and fund management. Block 1 is project approval, which records project approval information and ensures project legality and compliance. Block 2 is contract payment, which manages the contract payment process and ensures the rational allocation and use of funds. Block 3 is progress payment settlement, which handles the calculation and payment of progress payments and ensures the transparency of fund flows. Block 4 is post-investment evaluation, which conducts post-evaluation of investment projects and provides reference for future investment decisions. The engineering quantity verification chain is used for engineering quantity verification and storage of related data. Block A is the BIM quantity calculation model, which stores engineering quantity calculation data related to the BIM model. Block B is progress review, which is used to review project progress and ensure that the project is progressing as planned. Block C is change approval, which records engineering change and approval information and ensures data transparency and traceability. The audit evidence chain is used for auditing and evidence storage, ensuring the authenticity and compliance of project data. Block α is the hidden project evidence storage, which records data and vouchers for hidden projects and ensures project transparency. Block β is for payment compliance proof: it provides proof of payment compliance and ensures the legal use of funds. Block γ is for decision audit: it records the decision-making process and audit results to ensure transparency and traceability. Cross-chain bridges are used to connect different blockchains and are responsible for transmitting information and instructions between different chains.

[0036] The blockchain investment management unit's workflow is as follows: When construction reaches a certain stage or meets specific payment conditions, the owner initiates a payment request, initiating the entire payment verification process. Upon receiving the owner's payment request, the smart contract retrieves the construction quantity data from the quantity calculation sidechain. The smart contract then requests the current hash value (HO) from the audit sidechain. After auditing the construction quantity data, the audit sidechain adds the relevant information to the MPT tree and generates a new hash value (HO'), which represents the current state of the audited construction quantity data. The audit sidechain compares the retrieved HO' with the original hash value (HO) stored in the main chain to verify whether the construction quantity data has been tampered with since the last verification. The audit sidechain then feeds the verification results back to the smart contract. If HO' matches HO, the construction quantity data has not been tampered with and is authentic. If they do not, there may be issues with the data, requiring further verification. Based on the verification results, the smart contract decides whether to proceed with the payment. If the verification is successful, the smart contract will automatically execute the payment according to the pre-set terms; if it fails, the payment process will be suspended and the relevant parties will be notified to handle it. Figure 4 .

[0037] The blockchain investment management unit's main chain is the investment decision-making chain (the capital flow hub), with a core function of full-lifecycle capital management, encompassing project approval, financing agreement, progress payment, and final settlement. The quantity calculation sidechain is a quantity verification chain (physical-digital mapping). Its core function is BIM-integrated intelligent quantity calculation, automatically extracting Revit model parameters (such as concrete volume and rebar specifications), generating on-chain quantity orders, dynamically adjusting construction costs, connecting to material price oracles (such as steel futures data), and updating comprehensive unit prices in real time. The audit sidechain is an audit evidence chain (a compliance barrier). Its core function is zero-knowledge proofs (ZK-SNARKs), verifying the authenticity of construction quantities (e.g., proving that hidden construction images have not been tampered with), automating compliance reviews, and using pre-set rules to check compliance with payment conditions. This invention links BIM construction quantities with contract payment nodes through smart contracts, with an error rate of ≤1%. It generates structured audit reports and records fund flows, shortening the audit cycle.

[0038] The dynamic deviation correction management unit, incorporating IoT sensors, collects construction deviation data in real time, predicts the impact of deviations through Monte Carlo simulation, and employs a genetic algorithm to optimize the movement paths of construction machinery. The impact of these deviations is then visualized within the BIM-GIS model. The dynamic deviation correction management unit for construction projects serves as the intelligent hub of modern project management. Through real-time monitoring, intelligent analysis, and precise control, it ensures that the construction process remains efficient and consistent with design objectives (schedule, quality, safety, and cost). The core principle of dynamic deviation correction management is dynamic closed-loop control, essentially a real-time feedback loop of "perception-decision-execution." Its key feature is the simultaneous control of multiple objectives: progress (e.g., node completion rate), quality (e.g., concrete strength deviation), safety (e.g., support deformation), and cost. Control methods include feedforward control, which anticipates risks and proactively intervenes (e.g., preemptively reinforcing foundation pits due to heavy rain warnings), and feedback control, which adjusts subsequent plans based on historical deviations (e.g., optimizing subsequent processes when progress lags).

[0039] The dynamic deviation correction management unit of the present invention adopts a three-layer intelligent system architecture, 1. Perception layer (data acquisition and fusion), 2. Decision layer (intelligent analysis and strategy generation), 3. Execution layer (precise control and feedback). Perception layer: responsible for data acquisition and fusion, including GNSS positioning, laser SLAM, BIM model, etc., and output to the decision layer through data preprocessing (Kalman filtering). Decision layer: performs intelligent analysis and strategy generation, including deviation analysis, root cause diagnosis, strategy library, algorithm engine (genetic algorithm). Execution layer: responsible for precise control and feedback, including automated equipment control, manual intervention mechanism, and feedback of execution effects to the perception layer to form a closed loop. Relationship between layers: perception layer → decision layer (real-time monitoring data), decision layer → execution layer (control instructions), execution layer → perception layer (effect feedback). See the architecture diagram of the dynamic deviation correction management unit. Figure 5 .

[0040] A quality and safety accountability traceability unit architecture digitizes responsible entities, creating blockchain digital identities for each participant (design institute, construction company, supervisor, supplier, etc.). RFID tags are used to bind construction process IDs and responsible personnel information, creating a quality and safety responsibility map. Full-process construction records can be retrieved by scanning a QR code. Key quality and safety operations (such as material delivery, hidden work acceptance, and inspection reports) are automatically collected and uploaded to the blockchain through IoT devices and manually confirmed through a dual-track system. Smart contracts, with pre-set quality and safety standards, automatically trigger accountability decisions (e.g., if concrete strength fails to meet standards, trace back to the supplier and supervisor). Three-dimensional spatial positioning, combined with BIM models, enables spatial localization of quality issues (including those responsible for construction processes).

[0041] The quality and safety responsibility traceability unit architecture is broken down into the following: 1. Basic Data Layer: Identity chain, participant registration (construction unit / designer / construction / supervisor / supplier, etc.); Personnel registration (project manager, supervisory engineer, special trades, etc. (linked to professional qualification certificates); Material and Equipment Chain (materials such as steel bars and concrete, factory QR code → on-site acceptance → re-examination report); Construction equipment installation acceptance → regular inspection records; Process Chain (key process connections, hidden acceptance reports, acceptance records, video data + inspection data + signatures). 2. Smart Contract Layer: Quality standard contracts, codified national and industry standards, automatic judgment; Safety standard contracts, management and control of hazardous projects; Responsibility identification rules, problem location, and traceability paths. 3. Traceability Application Layer: Quality issue traceability (input → component number / QR code, output → designer, material, construction, and supervisor); Safety incident traceability (input → incident location / time, output → direct responsible party, management, and technical responsible party); Credit rating system with automatic deductions and blacklisting. 4. Technical Implementation Layer: Blockchain, Hyperledger Fabric consortium chain (construction industry chain); Internet of Things, sensors + edge computing gateway (automatic data collection and uploading); BIM, IFC standard component unique coding; big data, quality problem pattern mining (quality fluctuation rate > 5% → early warning). See the quality and safety responsibility traceability unit architecture diagram. Figure 6 .

[0042] The NLP risk management unit, which manages NLP risk throughout the entire lifecycle of construction projects, primarily identifies and manages risks by analyzing text data. Natural language processing (NLP) extracts contract text keywords, analyzes and mines unstructured text data, and constructs an evaluation model based on multi-dimensional risk indicators (such as weather data and construction period data). A fuzzy comprehensive evaluation indicator system (such as risk probability, impact level, and response cost) is used to identify potential risk factors, providing support for risk assessment, monitoring, and decision-making. A comprehensive evaluation method quantifies risk levels, enabling early identification and response to risks and minimizing losses. The NLP risk management unit of this invention includes data acquisition, data sources, a contract management system, and a project supervision system. The contract management system, as one of the data sources, collects various contract-related information, including textual content such as contract terms, responsibilities and obligations of each party, and timelines, providing the foundation for subsequent risk identification and analysis. The project supervision system collects data generated during the project supervision process, such as supervision logs and notices. This data contains supervisory records and potential risk clues regarding project quality, progress, and safety, and is transmitted to subsequent processing stages via the data acquisition layer. The data collection layer is responsible for collecting and aggregating data from the above two systems, playing the role of a hub for data aggregation and ensuring that all types of text data can flow smoothly to the NLP processing engine. The NLP processing engine receives text data from the data collection layer, performs basic natural language processing operations on these texts, and converts the original text into a structured or semi-structured data form that is convenient for subsequent in-depth analysis, laying the foundation for subsequent risk information mining. The NLP processing engine includes text preprocessing and risk knowledge graphs. The text preprocessing module includes entity recognition, sentiment analysis, and relationship extraction. The risk knowledge graph module includes risk nodes, correlation relationships, risk warnings, and disposal. See the NLP risk management unit architecture diagram. Figure 7 .

[0043] The NLP risk management unit automatically extracts key contract terms, such as construction period, quality standards, and payment methods, through contract risk identification. It identifies risky clauses, such as unreasonable allocation of responsibilities and ambiguous wording, and provides early warnings of clauses that may lead to disputes. Safety risk monitoring analyzes text data such as construction logs and accident reports to extract information on safety hazards, such as failure to wear a hard hat and illegal operations, to promptly identify on-site safety hazards. Project risk assessment: This unit studies project reports, emails, meeting minutes, and other documents to extract risk-related information, such as schedule delays and cost overruns, and assess the overall risk level of the project. New text data is monitored in real time. When potential risks are discovered, early warning signals are issued promptly, and a contingency plan library is generated. The risk assessment response time for a single contract is ≤1 minute.

[0044] The digital twin operation and maintenance unit is an intelligent management center for the entire life cycle of buildings based on the full-time mapping of "physical-virtual". It integrates BIM+IoT+AI to realize real-time perception, simulation deduction, and autonomous decision-making of the building operation and maintenance process. Digital twin operation and maintenance of construction project management is a key part of intelligent operation and maintenance management of construction projects throughout the life cycle of the building, with digital twin technology as the core, integrating technologies such as the Internet of Things, big data, cloud computing, and artificial intelligence. By digitally modeling the construction and its equipment and facilities, a virtual copy corresponding to the actual construction is created, which not only contains the static geometric information of the building, but also includes the dynamic data such as the operating parameters and real-time status of the equipment dynamically corrected by the virtual model based on the real-time posture feedback of the construction machinery; a complete workflow from data collection, transmission, processing to final visualization and decision support.

[0045] The digital twin operation and maintenance unit consists of the following parts: 1. Three-dimensional model creation: Based on BIM technology, a detailed three-dimensional model of the building is created to accurately reflect the building's structure, layout, and the spatial positional relationships of various equipment, pipelines, etc. This is the basic framework of the digital twin operation and maintenance unit. 2. Data acquisition system: By widely deploying various sensors in the building, the building's operating data is collected in real time, including equipment operating status, environmental parameters, energy consumption and other information, and this data is transmitted to the digital twin platform. 3. Data processing and analysis platform: Cleans, integrates, stores and analyzes the massive amount of multi-source data collected. Big data technology is used to achieve efficient data storage and management, and valuable information is extracted from the data through tools such as data mining algorithms and machine learning models to provide support for operation and maintenance decisions. 4. Visual interface: Displays the digital twin model and related data in an intuitive and easy-to-understand visual way. For example, through three-dimensional modeling, geographic information system (GIS) and other technologies, the building's appearance, internal structure, equipment distribution, etc. are presented in a graphical way, and various operating data and analysis results are displayed in real time, making it convenient for managers to quickly understand the overall operating status of the building. See the digital twin operation and maintenance unit architecture diagram. Figure 8 The digital twin operation and maintenance unit of the present invention analyzes historical energy consumption data through an LSTM neural network. The LSTM neural network adopts a BiLSTM (bidirectional long short-term memory) model. The BiLSTM model is input to predict the peak energy consumption curve for the next 7 days. The device operation status data (such as voltage and temperature) and historical energy consumption data are integrated through the spatiotemporal alignment layer to dynamically adjust the device operation strategy.

[0046] The BIM intelligent management method for the entire life cycle of engineering construction of the present invention comprises the following steps: S100, Data Integration and Lightweight Processing: When importing multi-source BIM models (Revit / CATIA), the import process requires first checking the model's format compatibility and performing necessary format conversions (such as unifying the coordinate system and unit system) for incompatible parts to ensure smooth subsequent processing. A code table generation algorithm based on Huffman coding optimizes the spatial locality of BIM models. This algorithm first analyzes the distribution and associations of the BIM model's spatial elements (such as rooms and building components), identifies groups of elements with spatial locality, and then generates a more efficient code table based on the characteristics and frequency of occurrence of these groups to effectively compress the model data. Geometric data is triangulated and compressed using the WebGL engine. This operation decomposes complex geometric shapes into multiple triangular facets. Based on the model's accuracy requirements and display quality, the number of facets is reduced while maintaining a largely unchanged model's appearance. Non-geometric data is stored in the cloud using a distributed storage architecture (such as cloud object storage services, combined with data redundancy and backup strategies) to ensure data security and high availability. For BIM models with fewer than 100,000 triangles, this processing improves model loading efficiency by 60% or more. A lightweight, mobile-interactive model is generated based on digital-analog separation technology (including: geometric data compression using the Draco algorithm, data flow optimization based on spatial locality Huffman coding, dynamic access to non-geometric data via a RESTful API, and support for offline annotation and version rollback). During this separation process, the model's appearance data (such as texture, color, and shape) is separated from business logic data (such as attribute information and relationships). When generating a mobile-interactive model, appearance data is loaded first for fast display, followed by on-demand loading of business data to support interactive operations.

[0047] S200, real-time data collection and fusion: IoT sensors deployed at key locations on the construction site, such as cranes and concrete pouring points, collect construction progress and quality inspection data (including key indicators such as concrete strength and rebar spacing) and transmit this data in real time to a data center. Simultaneously, lidar (LiDAR) is used to scan the construction site, generating high-precision point cloud data. LiDAR is widely used on construction sites for topographic mapping and 3D building modeling. LiDAR's key features include high-precision measurement, real-time data acquisition, and non-contact measurement. The raw data collected by LiDAR consists of a large amount of point cloud data, requiring specialized software for processing and analysis. Common methods include point cloud filtering, feature extraction, and 3D modeling. These methods can extract useful information from the raw data, such as building outlines, ground elevation, and object shape, and generate 3D models to visually display the construction site. Data collection can also be achieved through wearable equipment, such as AR glasses, smart helmets, and smart clothing. These integrate sensors, communication modules, and data processing units to monitor personnel's position, posture, physiological data, and environmental information in real time. This data is then transmitted to a management platform via wireless networks, enabling interoperability between personnel and the platform. The smart helmet is also equipped with a camera, lighting system and voice recognition module, which supports on-site image acquisition, lighting and voice command control, improving the convenience and safety of personnel operation. The collected point cloud data is precisely aligned with the BIM model using a point cloud matching algorithm. A feature-based point cloud-model matching algorithm is used to extract key feature points from the point cloud and model. The ICP iterative closest point algorithm is then used to align the BIM model with the construction site point cloud data. The RANSAC algorithm is used to remove noise points, enabling augmented reality (AR) overlays. The alignment accuracy is ≤3mm. For positioning data acquisition, GNSS-RTK technology is first used to obtain initial positioning data for construction equipment or personnel. Because construction sites may be subject to occlusion or signal interference, positioning accuracy is affected. Therefore, a six-axis attitude sensor is introduced and installed on construction equipment or helmets to monitor the three-axis acceleration and three-axis angular velocity of the equipment or personnel in real time. The GNSS-RTK positioning data is corrected based on the attitude information measured by the sensor (including yaw, pitch, and roll angles). Fusion algorithms (such as the Kalman filter) leverage the complementary strengths of the two data sources to ultimately generate dynamic construction coordinates in the ECEF coordinate system, providing reliable spatial positioning information for precise control of the construction process.

[0048] S300, Intelligent Decision-making and Execution: Collect historical construction project data (including multi-dimensional data such as cost, progress, and quality of previous projects, and remove duplicate and erroneous information through data cleaning) and real-time construction data (such as the work status of on-site personnel, equipment operating parameters, material inventory, etc., which are collected and transmitted in real time using wireless sensor networks). Utilize data mining (using association rule mining to find the relationship between different factors, such as the relationship between material quality and construction progress) and machine learning algorithms (such as decision tree algorithms for classifying different risk types) to analyze and model the data. Train LSTM neural networks based on historical project data (using a large amount of historical sample data to iterate the model for multiple rounds and adjust the weights and bias parameters of the neural network). Predict investment risks (material price fluctuations, considering a comprehensive analysis of multiple factors such as market trends and supplier reputation) and construction period deviations (critical path optimization, identifying key tasks and critical paths in the project, analyzing their influencing factors and optimizing adjustments), thereby predicting the project's progress and cost. and quality risks, provide data support for decision-making, and the response time for BIM models with less than 100,000 triangles is ≤2 seconds; combine BIM models with intelligent decision-making systems, and provide intuitive project information and 3D model support for decision-making through the advantages of BIM model visualization (using intuitive 3D graphics to display building structure, component location and other information for easy user understanding) and informatization (storing rich component properties, engineering quantities and other detailed information); use the data in the BIM model (such as engineering quantity bills, material bills, etc.) to interact with intelligent decision-making algorithms, and automatically associate BIM engineering quantities with contract payment nodes through smart contracts (setting contract terms and payment conditions, and automatically triggering the payment process when the conditions are met), with an error rate of ≤1% (using precise data matching algorithms to strictly check engineering quantities and payment nodes). The smart contract adopts a three-chain structure (main chain + calculation side chain + audit side chain). The main chain stores fund payment vouchers, the side chain records calculation data, and the audit side chain adopts Merkle Patricia Tree stores the hash value of the audit report; it verifies data consistency through zero-knowledge proof technology, ensures audit transparency, and shortens the time it takes to generate audit reports by 50% (reducing manual operations through automated processes and improving report generation efficiency).

[0049] S400, dynamic correction and visual feedback: Based on real-time monitoring data (including multi-dimensional information such as construction progress, quality, and cost, obtained through sensor networks and data acquisition systems), Monte Carlo simulation (using a large number of random sampling to simulate the possible results of the construction process under different scenarios) is used to analyze the deviation, compare the differences between actual construction data and planned data, calculate the size and trend of the deviation, and take corresponding measures (such as adjusting construction resource allocation, optimizing construction sequence, etc.) to adjust and correct it to ensure that the system output is consistent with the preset target or minimize the deviation process, generate the optimal correction path (taking into account factors such as construction cost, time, and quality, and using optimization algorithms to calculate the best correction plan) and use BIM-GIS model (combining building information model and geographic information system to display the construction site and The surrounding environment) is visualized, and the operating status of the construction process (such as equipment operating status, personnel work status, etc.), data information (such as material consumption, project quantity, etc.) are presented to users or operators in an intuitive and easy-to-understand visual form (such as three-dimensional graphics, charts, animations, etc.), so that they can quickly and accurately understand the construction situation, deviation status and correction effect, etc., thereby providing a basis for decision-making and operation. When the deviation exceeds the threshold (a reasonable deviation threshold is set, such as progress deviation exceeds 5%, cost deviation exceeds 3%, etc.), the AR sound and light alarm is triggered (using augmented reality technology to display the alarm information on the construction personnel's AR device and issue a sound prompt) and the correction instruction is pushed to the responsible person's terminal. Specific correction measures and requirements are sent to the relevant responsible person through mobile applications or text messages.

[0050] S500, quality and safety responsibility definition and traceability: During the construction process, RFID tags (with unique identification codes and a certain storage capacity) are used to bind the construction process ID and the responsible person information. Based on the construction process logs collected by RFID, the failure mode and the responsible unit are associated, the responsibility weight is quantified, and the responsibility weights of the design unit, construction unit, and supervision unit are associated. The responsibility clauses in the contract text are extracted through NLP, and the responsibility traceability path is generated. The quality and safety responsibility map is constructed, and the information of construction process, materials, equipment, personnel, etc. is collected and recorded in real time. Each RFID tag uniquely corresponds to a construction process or material batch, recording detailed information such as the responsible person (including the construction unit leader, construction team leader, and specific operators), construction time (accurate to year, month, day, hour, minute, and second), and construction process parameters (such as concrete pouring temperature, welding current and voltage). The collected information is stored in a database, forming a complete construction process data chain. The data includes construction logs (detailed records of daily construction content and progress), inspection reports (material quality and construction quality inspection results), concealed project acceptance records (concealed project acceptance time, participants, acceptance results, etc.), material incoming inspection records (material incoming time, inspection results, inspectors, etc.), safety inspection records (safety inspection time, inspection content, hidden danger discovery and handling, etc.), and safety accident reports (accident time, location, process, casualties, preliminary cause analysis, etc.), ensuring that every construction link and material use is documented. The management platform links each construction process and element to the corresponding responsible party, including the construction unit, construction team, specific operators, material suppliers, and supervision units, clarifying the quality and safety responsibilities of each link. When a quality issue or safety incident occurs, the system quickly locates the specific link and time point where the problem occurred by querying the database and analyzing relevant data. Based on pre-defined responsibility binding relationships and FMEA (Failure Mode Analysis) weighting (weights are assigned based on the role and impact of different responsible parties in the construction process, such as 40% for the construction unit, 30% for the construction team, 20% for the material supplier, and 10% for the supervision unit), the system traces the cause back to the relevant individuals and units. Furthermore, the system can review relevant construction parameters (such as equipment operating parameters, ambient temperature and humidity, etc.) and environmental factors before and after the quality issue or safety incident, providing a basis for cause analysis. Quality issues and traceability results are promptly fed back to the relevant responsible parties and management, urging them to take corrective measures, repair or rework quality defects, and properly handle safety incidents. Appropriate penalties or education (such as fines, public criticism, safety training, etc.) are also implemented to prevent similar issues from recurring. The system also supports scanning QR codes to retrieve full-process construction records, including quality and safety-related records, for rapid traceability and verification.

[0051] S600, full life cycle operation and maintenance management: Based on digital twin technology simulation, data from the entire life cycle of construction projects is collected, analyzed, and utilized to achieve comprehensive monitoring, intelligent diagnosis, and optimized decision-making of the construction process, so as to achieve efficient operation, cost control, quality assurance, and sustainable development of the project; advanced algorithms are used to analyze and process data (for example, energy consumption data is predicted and analyzed through machine learning algorithms to predict possible equipment failures in advance and provide support for operation and maintenance decisions). Specifically, time series analysis algorithms can be used to monitor equipment operation data in real time, and cluster analysis can be used to classify and evaluate construction quality data; the analysis results can be analyzed and processed through The visual interface (using three-dimensional visualization technology to display analysis results in the form of intuitive graphics and charts, such as energy consumption trend charts, equipment status indicator lights, etc.) is displayed to operation and maintenance personnel, allowing them to intuitively understand the building's operating status (including equipment operating status, energy consumption levels, construction progress, etc.) and respond to abnormal situations in a timely manner; through the remote control function of the digital twin model (using Internet of Things technology to achieve remote operation and adjustment of on-site equipment and construction processes), construction operations are optimized (such as adjusting equipment operating parameters, optimizing construction sequence, etc.), improving work efficiency and management level, while ensuring the safety and quality control of the construction process.

[0052] S700, Engineering management data classification and storage: Classify engineering management materials according to different stages of the project, such as project pre-planning, design stage, construction stage, completion acceptance stage, and operation and maintenance stage; classify materials according to the different disciplines involved in the project, such as construction engineering, structural engineering, water supply and drainage engineering, electrical engineering, and HVAC engineering; according to the content and purpose of the materials, they can be divided into management documents (including project plans, meeting minutes, decision-making documents, etc.), technical documents (such as design drawings, technical specifications, construction plans, etc.), quality control documents (such as quality inspection reports, material certificates, hidden project acceptance records, etc.), progress control documents (such as progress plans, progress reports, construction period change records, etc.), cost control documents (such as budget documents, settlement documents, cost analysis reports, etc.), contract documents (contract texts, change agreements, claim documents, etc.), etc.; classify by information format, and classify electronic files (such as Word documents, Excel spreadsheets, PDF files, etc.), image materials (such as photos, scanned drawings, etc.), and audio and video materials (such as construction site videos, meeting recordings, etc.). Utilize a professional file management system to centrally store and manage project management information. For structured data such as bills of quantities, quality inspection data, and progress data, database systems (such as SQL Server and Oracle) are used to store them, ensuring data integrity and efficient querying. Leverage cloud storage platforms (such as Alibaba Cloud and Tencent Cloud) to upload electronic documents, images, audio, and video materials to cloud storage spaces, and set access permissions (e.g., read-only, read-write, etc.) to ensure that team members can access the required data from different locations. Data encryption technologies (such as the AES encryption algorithm) are used to encrypt sensitive data (such as financial information and trade secrets) stored in file management systems and cloud storage services to prevent data leakage. For important paper-based project materials, such as completion records and original contracts, a dedicated archive room is set up for storage, equipped with fireproof, moisture-proof, and insect-proof facilities to ensure the physical security of the materials. Detailed indexing is established for project management materials, including file directories (multi-level directories by project phase and discipline), keyword indexing (extracting keywords from files for quick retrieval), and classification indexing (classified by material type and purpose). In terms of data sharing and collaboration, document management systems and cloud storage platforms are used for data sharing, allowing for real-time file viewing and editing, improving work efficiency while ensuring version control (recording file modification history and different versions). Access control policies are set (based on user role permissions, such as those for project managers, construction personnel, supervisors, etc.), limiting different users' access rights to data and ensuring that only authorized personnel can view, edit, and use specific files. Data analysis and mining are performed on stored engineering management data to extract valuable information and knowledge (such as identifying trends and causes of cost overruns through data analysis) to provide decision support for project management.

[0053] This management platform achieves closed-loop digital twin management in the construction engineering sector. Through algorithmic innovations (such as spatially localized Huffman coding and a three-chain blockchain architecture), it significantly elevates the industry's technical and management benchmarks. It transforms traditional challenges such as response latency, audit risk, and quality traceability into quantifiable and controllable technical indicators, significantly driving a substantial shift in engineering management from "experience-based management" to "data-based management" and "algorithm-based management."

[0054] The description and drawings of this application are only a specific implementation method and are not restrictive. Under the guidance of this application, those skilled in the art can make many forms without departing from the scope of protection of the purpose of this application, all of which are within the scope of protection of this application.

Claims

1. A BIM intelligent management platform for the entire life cycle of engineering construction, characterized by: The system consists of a small front-end module and a large back-end module. The small front-end module uses a lightweight WebGL engine and a specific compression algorithm (such as Draco compression + entropy coding) to achieve lossless conversion of multi-source BIM models (Revit / CATIA) and mobile terminal interaction, improving loading efficiency. It also integrates an augmented reality (AR) data console to overlay construction progress and quality inspection data onto the 3D real-life model in real time. Large backend module, including: The intelligent design management unit enables multi-disciplinary collaborative design based on a BIM-GIS fusion model. It identifies regulatory conflicts through an automatic collision detection algorithm (with a recognition rate of ≥95%). The collision detection algorithm supports BIM models with millions of triangles and uses VR technology to verify the feasibility of construction plans. The blockchain investment management unit links BIM project quantities with contract payment nodes through smart contracts, with an error rate of ≤1%. It generates structured audit reports and records fund flows, shortening the audit cycle. The dynamic deviation correction management unit uses IoT sensors to collect construction deviation data in real time, predicts the impact of deviations through Monte Carlo simulation, and uses genetic algorithms to optimize the movement paths of construction machinery. It also visualizes the impact range of construction deviation data in the BIM-GIS model. The quality and safety responsibility traceability unit uses RFID tags to bind construction process IDs and responsible person information to build a quality and safety responsibility map. It adopts a distributed storage architecture, uses blockchain to store key audit data, and uses a relational database to store frequently accessed construction logs. It realizes dual-database data association through process IDs and supports scanning codes to retrieve full-process construction records. The NLP risk management unit extracts contract text keywords through natural language processing, builds an evaluation model based on multi-dimensional risk indicators (such as weather data and construction period data), and uses a fuzzy comprehensive evaluation indicator system (such as risk probability, impact level, and response cost) to quantify the risk level through a comprehensive evaluation method. This unit then generates a contingency plan library, with a risk assessment response time of ≤1 minute for a single contract. The digital twin operation and maintenance unit creates a virtual replica corresponding to the actual construction by digitally modeling the construction and its equipment and facilities. It not only contains the static geometric information of the building, but also includes dynamic data such as the operating parameters and real-time status of the equipment dynamically corrected by the virtual model based on the real-time posture feedback of the construction machinery; it provides a complete workflow from data collection, transmission, processing to final visualization and decision support.

2. The BIM intelligent management platform for the entire life cycle of engineering construction according to claim 1 is characterized by: The WebGL lightweight engine adopts digital-analog separation technology, and entropy coding uses Huffman coding to optimize geometric data flow. Geometric data is triangulated and compressed for storage. Non-geometric data is dynamically called through a cloud API, supporting offline browsing and annotation functions on mobile devices.

3. The BIM intelligent management platform for the entire life cycle of engineering construction according to claim 1 is characterized by: The AR data console uses a lidar point cloud matching algorithm to achieve millimeter-level precision (≤3mm) alignment between the BIM model and the construction site, uses an IMU (inertial measurement unit) to compensate for point cloud matching delays, and triggers an audible and visual alarm when the deviation exceeds a threshold.

4. The BIM intelligent management platform for the entire life cycle of engineering construction according to claim 1 is characterized by: The blockchain investment management unit adopts a three-chain architecture (main chain + calculation side chain + audit side chain). The main chain stores contract payment vouchers, the side chain records BIM calculation data, and the audit side chain uses Merkle Patricia Tree to store audit report hash values, enhancing data immutability and achieving privacy protection through zero-knowledge proof technology.

5. The BIM intelligent management platform for the entire life cycle of engineering construction according to claim 1 is characterized by: The digital twin operation and maintenance unit analyzes historical energy consumption data through an LSTM neural network. The LSTM neural network adopts a BiLSTM (bidirectional long short-term memory network) model. The BiLSTM model is input to predict the peak energy consumption curve for the next 7 days. The device operating status data (such as voltage and temperature) and historical energy consumption data are integrated through the spatiotemporal alignment layer, and the device operation strategy is dynamically adjusted.

6. A BIM intelligent management method for the entire life cycle of engineering construction according to claim 1, characterized in that: The following steps are involved: S100, Data Integration and Lightweight Processing: Import multi-source BIM models (Revit / CATIA), optimize the code table generation algorithm based on Huffman coding for the spatial locality of BIM models, use the WebGL engine to triangulate and compress geometric data and store non-geometric data in the cloud, and improve model loading efficiency by ≥60% for BIM models with less than 100,000 triangles. Generate lightweight models that are interactive on mobile devices based on digital-analog separation technology. S200, real-time data acquisition and fusion: IoT sensors are used to collect construction progress and quality inspection data, which are then matched with the BIM model through LiDAR point cloud to achieve AR augmented reality overlay with an alignment accuracy of ≤3mm. GNSS RTK positioning data is corrected using a six-axis attitude sensor to generate dynamic construction coordinates in the ECEF coordinate system. S300, Intelligent Decision-making and Execution: Collect historical and real-time construction project data, analyze and model the data using data mining and machine learning algorithms, and train LSTM neural networks based on historical project data to predict investment risks (material price fluctuations) and construction period deviations (critical path optimization). This allows us to predict project progress, cost, and quality risks, providing data support for decision-making. For BIM models with fewer than 100,000 triangles, the response time is ≤ 2 seconds. Integrate BIM models with intelligent decision-making systems to provide intuitive project information and 3D model support for decision-making through the visualization and information advantages of BIM models; By leveraging data from the BIM model to interact with intelligent decision-making algorithms, smart contracts can automatically link BIM project quantities with contract payment nodes, with an error rate of ≤1% and a 50% reduction in audit report generation time. S400, dynamic correction and visual feedback: Based on real-time monitoring data, Monte Carlo simulation is used to analyze deviations, and corresponding measures are taken to adjust and correct them to ensure that the system output is consistent with the preset target or minimize the deviation process. The optimal correction path is generated and visualized through the BIM-GIS model. The operation status and data information of the construction process are presented to users or operators in an intuitive and easy-to-understand visual form, enabling them to quickly and accurately understand the construction status, deviation status and correction effect, etc., thereby providing a basis for decision-making and operation. When the deviation exceeds the threshold, the AR sound and light alarm is triggered and the correction instruction is pushed to the responsible person's terminal; S500, quality and safety responsibility definition and traceability: During the construction process, RFID tags are used to bind the construction process ID and the responsible person information, build a quality and safety responsibility map, and collect and record information on construction processes, materials, equipment, personnel, etc. in real time. Each RFID tag uniquely corresponds to a construction process or a material batch, recording its related responsible person, construction time, construction process parameters and other detailed information; the collected information is stored in the database to form a complete construction process data chain, including construction logs, inspection reports, hidden project acceptance records, material incoming inspection records, safety inspection records, safety accident reports, etc., to ensure that every construction link and material use is traceable; use the management platform to bind each process and element in the construction process with the corresponding responsible party, including the construction unit, construction team, specific operators, material suppliers, supervision units, etc., to clarify the quality and safety responsibilities of each link; When a quality problem or safety incident occurs, the system can quickly locate the specific link and time point where the problem occurred by querying the database and analyzing relevant data. Based on the pre-set responsibility binding relationship and the responsibility weight allocation based on FMEA (Failure Mode Analysis), the system can trace the relevant responsible persons and responsible units. At the same time, it can also view relevant construction parameters, environmental factors and other information before and after the quality problem or safety incident occurred, providing a basis for problem cause analysis. Provide timely feedback of quality issues and traceability results to relevant responsible persons and management, urging them to take corrective measures, repair or rework quality defects, and properly handle safety incidents. At the same time, impose corresponding penalties or education on relevant responsible persons to prevent similar problems from recurring. Support scanning QR codes to retrieve full-process construction records, including quality and safety-related records, for rapid traceability and verification. S600, Full lifecycle operation and maintenance management: Based on digital twin technology simulation, data from the entire life cycle of construction projects is collected, analyzed, and utilized to achieve comprehensive monitoring, intelligent diagnosis, and optimized decision-making of the construction process, thereby achieving efficient operation, cost control, quality assurance, and sustainable development of the project. Advanced algorithms are used to analyze and process data (for example, machine learning algorithms are used to predict and analyze energy consumption data, predict possible equipment failures in advance, and provide support for operation and maintenance decision-making). The analysis results are displayed to operation and maintenance personnel through a visual interface, allowing them to intuitively understand the operating status of the building. Through the remote control function of the digital twin model, construction operations can be optimized to improve work efficiency and management level. S700, Engineering management data classification and storage: Classify project management documents according to different stages of the project, such as pre-project planning, design, construction, completion acceptance, and operation and maintenance. Categorize documents based on the different disciplines involved in the project, such as construction engineering, structural engineering, water supply and drainage engineering, electrical engineering, and HVAC engineering. Based on the content and purpose of the documents, they can be divided into management documents, technical documents, quality control documents, progress control documents, cost control documents, contract documents, etc. Categorize by information format, such as electronic documents, image materials, and audio and video materials. Utilize a professional document management system to centrally store and manage project management data. A database system is used to store structured data, such as bills of quantities, quality inspection data, and progress data. Leverage a cloud storage platform to upload electronic documents, images, audio, and video materials to cloud storage, and set access permissions. Data encryption technology is used to encrypt sensitive data stored in document management systems and cloud storage services. Dedicated archives are set up to store important paper-based project materials, such as completion records and original contracts. Establish detailed indexes for project management data, including file directories, keyword indexes, and classification indexes; Data sharing and collaboration: Share data through file management systems, cloud storage platforms, etc., view and edit files in real time, and improve work efficiency; set access control policies to limit different users' access rights to data, ensuring that only authorized personnel can view, edit, and use specific files; Conduct data analysis and mining on stored engineering management data, extract valuable information and knowledge, and provide decision support for project management.

7. The BIM intelligent management method for the entire life cycle of engineering construction according to claim 6 is characterized by: In step S100, the digital-analog separation technology includes: Geometric data is compressed using the Draco algorithm, combined with Huffman coding based on spatial locality to optimize data flow. Non-geometric data is dynamically called through the RESTful API interface, supporting offline annotation and version rollback.

8. The BIM intelligent management method for the entire life cycle of engineering construction according to claim 6 is characterized by: In step S200, the laser radar point cloud matching includes: The BIM model and the construction site point cloud data were aligned using the ICP iterative closest point algorithm, and the RANSAC algorithm was used to remove noise points, with a matching error of ≤3mm.

9. The BIM intelligent management method for the entire life cycle of engineering construction according to claim 6 is characterized by: In step S300, the smart contract adopts a three-chain structure (main chain + calculation side chain + audit side chain). The main chain stores fund payment vouchers, the side chain records calculation data, and the audit side chain uses Merkle Patricia Tree to store the audit report hash value. Data consistency is verified through zero-knowledge proof technology to ensure audit transparency.

10. The BIM intelligent management method for the entire life cycle of engineering construction according to claim 6 is characterized by: In step S500, the construction of the quality responsibility map includes: associating failure modes with responsible units based on the construction process logs collected by RFID, quantifying responsibility weights, associating the responsibility weights of the design unit, construction unit, and supervision unit, extracting responsibility clauses in the contract text through NLP, and generating a responsibility traceability path.

Images