A method for digital management and control of construction progress of an ultra-large-scale plant and operation regulation

Abstract

The application discloses a kind of super large scale factory construction progress digitization management and control and operation regulation method, belong to intelligent construction and digital twin construction management technical field, by constructing structure deformation, clean environment and electromechanical installation three-domain coupling digital twin, by the domain measured deviation is normalized after weighting summation, calculate comprehensive determination index to identify parallel operation start window, dynamic release clean room wallboard, elevated floor and electromechanical pipeline construction interface, and by augmented reality terminal to construction team present space-time operation boundary, guide multi-professional in main body structure middle section synchronous insertion construction, for progress deviation, based on historical conflict data inversion and forgetting weighted recursive least square principle adaptive update coupling weight coefficient, and introduce duration, cost and clean quality multi-objective collaborative optimization model solution optimal resource allocation scheme.The application breaks traditional serial construction mode, can shorten clean room construction cycle, guarantee high-precision clean environment construction quality.

Description

Technical Field

[0001] This invention belongs to the field of intelligent construction and digital twin construction management technology, specifically relating to a method for digital control and operation regulation of construction progress in ultra-large-scale factory buildings. Background Technology

[0002] Ultra-large-scale wafer fabs (typically referring to wafer fabrication plants with a single building area of ​​≥100,000㎡, a cleanliness level of ISO 14644-1 Class 3 or higher, and a heavy-duty floor load of ≥5t / ㎡) are the core infrastructure of the semiconductor industry. The construction of these fabs is characterized by massive investment (investment in a single 12-inch wafer fab can reach billions of US dollars), tight construction cycles (market windows typically require completion from groundbreaking to mass production within 18-24 months), and dense cross-disciplinary operations (involving more than ten disciplines such as civil engineering, cleanroom enclosure systems, raised floors, electromechanical pipelines, and process piping, operating simultaneously or alternately). Among these, the cleanroom area, as the core production environment for chip manufacturing, has extremely stringent physical control requirements regarding structural micro-deformation, temperature and humidity fluctuations, dust particle concentration, and micro-vibration amplitude.

[0003] Currently, the construction management of ultra-large-scale factory buildings mainly relies on the following two types of technical methods: Firstly, there are BIM (Building Information Modeling)-based schedule planning and visualization management systems. Mainstream BIM collaboration platforms, such as Autodesk BIM360 and Glodon BIM5D, can 4D correlate 3D building models with construction schedules (usually based on Critical Path Method (CPM) or P6 software), enabling visualized simulation and static comparison of construction progress. These systems allow construction managers to view the differences between the planned and actual completion status of various areas at project meetings using the 3D model, and to make manual coordination and decisions accordingly. In addition, some projects have begun to incorporate IoT monitoring data (such as foundation pit monitoring and concrete strain sensing) into construction management; however, this data is typically used only for independent safety monitoring and has not yet been integrated with the BIM schedule model to form a data fusion and closed-loop feedback mechanism.

[0004] Secondly, a construction process simulation and status display system based on digital twins. In recent years, digital twin technology has been gradually introduced into the field of semiconductor factory construction. By constructing a virtual factory model and mapping it with real-time data from the construction site, dynamic visualization of the construction status can be achieved. Some existing technologies (such as the PSIM digital twin factory solution from the China Electronics Research Institute) have attempted to apply digital twins to the full lifecycle management of semiconductor factories, covering design optimization, construction simulation, and operation and maintenance management. Patent documents (such as CN113110313A) also propose using digital twin models to predict the construction process in advance and optimize construction parameters.

[0005] However, the aforementioned existing technologies have the following significant shortcomings when it comes to controlling the construction progress of ultra-large-scale factory buildings: First, the construction organization model still primarily follows the traditional sequential logic of "structure → decoration → electromechanical". Current construction plans typically require the main structure to be fully capped and structural acceptance completed before the cleanroom enclosure system installation can begin; the pre-laying and installation of main electromechanical pipelines can only commence after the cleanroom enclosure system is completed. This sequential model results in a cleanroom construction cycle of 6-8 months and an overall production delay of 4-6 months, making it difficult to meet the semiconductor industry's urgent requirements for construction schedules. Although some studies have proposed the concepts of "parallel construction of structure and electromechanical systems" or "interleaved construction of decoration and cleanroom engineering", existing technologies have failed to solve the coupling interference problem between high-precision cleanroom area construction and main structure construction—the continuous construction of the main structure generates dynamic deformation, vibration, and dust, while cleanroom wall panel installation, raised floor laying, and other processes are sensitive to structural deformation, micro-vibration, and air cleanliness at the millimeter or even micrometer level. While the main structure is still under construction, factors such as daily structural deformation, concrete curing shrinkage, and impact from heavy equipment hoisting will continue to affect the precision and quality of cleanroom components that have been installed or are being installed. Existing technologies lack quantitative assessment and dynamic control methods for this coupled interference.

[0006] Secondly, the progress management functions of existing BIM and digital twin systems are limited to static comparisons and visualizations of "planned vs. actual" progress, lacking dynamic decision-making and automatic adjustment capabilities. When actual progress deviates from planned progress, existing systems can only indicate the deviation through color markings or reports, unable to automatically assess the scope and severity of the deviation's impact on subsequent processes (especially the handover between different disciplines), let alone automatically generate targeted resource scheduling or process adjustment instructions. The identification and handling of progress deviations heavily rely on manual judgment and coordination meetings by project management personnel, with response cycles typically measured in days, making it difficult to meet the demand for minute-level rapid response in the context of multi-disciplinary parallel construction of ultra-large-scale factories.

[0007] Third, existing technologies have failed to establish a coupled constraint model and dynamic early warning mechanism among the three elements of "structural construction—clean environment—mechanical and electrical installation." The construction of cleanrooms in ultra-large-scale factories involves multiple and interrelated constraints on structural deformation, temperature and humidity, dust particles, and micro-vibrations. For example, welding operations of mechanical and electrical pipelines may generate vibrations and particulate matter, affecting the micro-vibration and cleanliness indicators of adjacent cleanroom areas; the concrete pouring of the main structure may cause excessive settlement at the installed raised floor support points; and the installation window for cleanroom wall panels depends on the convergence state of structural deformation in that area. These three elements have complex interrelationships, but existing technologies isolate structural monitoring, environmental monitoring, and progress management, resulting in severe data silos and an inability to achieve dynamic coupled assessment and coordinated control of these three constraints.

[0008] Therefore, it is necessary to propose a digital management and control method for the construction progress of ultra-large-scale factory buildings to solve the above problems. Summary of the Invention

[0009] In view of this, the purpose of this invention is to provide a digital management and control method for the construction progress of ultra-large-scale factory buildings, which solves the problem in the prior art of how to safely and accurately identify the start time and available window of parallel operations by quantitatively assessing the dynamic coupling relationship between structural deformation, clean environment and electromechanical installation before the main structure is completed, and automatically identify progress deviations and generate resource scheduling instructions during multi-disciplinary parallel construction, thereby realizing real-time digital management and control of the construction progress of ultra-large-scale factory buildings and automatic control of parallel operations.

[0010] To achieve the above objectives, the present invention provides the following technical solution: This invention discloses a method for digital management and operation control of construction progress in ultra-large-scale factory buildings, comprising the following steps: S1. Acquire multi-source monitoring data during the construction phase. The multi-source monitoring data should include at least structural deformation data, clean area environmental data, and electromechanical installation disturbance data; upload the multi-source monitoring data to the digital twin platform. S2. Construct a coupled digital twin of the 3D building information model and the schedule model in the digital twin platform. The coupled digital twin integrates the structural construction status, cleanroom installation logic constraints, and electromechanical pipeline conflict detection information. Based on the coupled digital twin, establish a three-domain coupled constraint model of the structural domain, cleanroom domain, and electromechanical domain. The three-domain coupled constraint model is used to evaluate the degree of mutual influence between structural deformation, cleanroom environment fluctuations, and electromechanical installation disturbances. S3. Based on the evaluation results of the three-domain coupling constraint model, determine whether the current construction status meets the preset parallel operation start conditions; if it does, automatically identify the types of construction interfaces that can be released. The construction interface types include at least the cleanroom wall panel installation interface, the raised floor laying interface, and the electromechanical pipeline pre-embedded working surface; for the identified construction interface, calculate its available time window and spatial operation boundary, and generate a construction interface release command. S4. Push the construction interface release command to the terminal devices of each relevant construction team. The terminal devices display the spatial boundary and time window information of the construction interface on the construction site in an augmented reality manner, guiding each professional team to carry out parallel construction operations within the spatial boundary and time window. S5. During parallel operations, continuously collect actual progress data from each construction interface, compare the actual progress with the planned progress, and identify processes with progress deviations. For the identified deviation processes, call the resource pool status data, generate resource scheduling instructions according to the preset priority scheduling rules, and issue them to compensate for the deviation processes with resources.

[0011] Furthermore, the three-domain coupling constraint model constructs the judgment logic in the following way: based on the measured cumulative structural deformation, cleanroom environmental fluctuation, and electromechanical installation disturbance at the current moment, each is normalized according to its preset tolerance threshold, so that the three physical quantities with different dimensions are uniformly converted into dimensionless relative deviation values; then, the three normalized values ​​are weighted and summed according to the weight coefficients corresponding to the structural domain, cleanroom domain, and electromechanical domain, to calculate the comprehensive coupling judgment index at the current moment; when the comprehensive judgment index is not greater than the preset threshold, it is determined that the current parallel operation start condition is met; otherwise, it is determined that it is not met; wherein, the weight coefficients of the structural domain, cleanroom domain, and electromechanical domain are all positive values ​​and the sum of the three is always equal to 1. Each weight coefficient is dynamically adjusted according to the construction stage of the plant. In the early stage of the main structure construction, the weight value of the structural domain is relatively large, and in the middle and later stages of the main structure construction, the weight values ​​of the cleanroom domain and electromechanical domain gradually increase to adapt to the difference in the degree of dominance of each domain on the parallel operation constraint conditions at different construction stages.

[0012] Furthermore, the parallel operation initiation conditions include structural domain determination conditions, cleanroom domain determination conditions, and electromechanical domain determination conditions. The structural domain determination condition is that the structural deformation amount does not exceed a preset structural deformation threshold. The cleanroom domain determination condition is that the particle concentration in the clean area does not exceed a preset particle concentration threshold and the vibration amplitude in the clean area does not exceed a preset vibration threshold. The electromechanical domain determination condition is that the position deviation of the reserved hole does not exceed a preset deviation threshold and the vibration amplitude in the electromechanical operation area does not exceed a preset electromechanical vibration threshold. The parallel operation initiation condition is determined to be valid only when all three domain determination conditions are met. If any condition is not met, the system marks the domain type that is not met and generates a corresponding adjustment suggestion.

[0013] Furthermore, the method for generating the construction interface release instruction is as follows: based on the current construction progress and the completion status of the preceding procedures of each construction interface, select the currently releaseable interface type from the preset construction interface type library; for each releaseable interface, determine the available time window based on the completion time of its preceding procedures and the latest start time of its subsequent procedures; determine the spatial operation boundary based on the three-dimensional model coordinates of the interface; and encapsulate the time window and spatial boundary into a structured instruction, which at least includes an interface type identifier, a spatial boundary coordinate set, the start and end times of the time window, and the procedure handover requirements.

[0014] Furthermore, the process of displaying the construction interface in an augmented reality manner by the terminal device includes: acquiring the real-time spatial pose of the terminal device; converting the spatial boundary coordinates in the received construction interface release command from the model coordinate system to the construction site coordinate system, and projecting the spatial boundary onto the real-world image of the terminal device according to the spatial pose of the terminal device, displaying the work boundary outline, time window information, and safety distance warning in an overlay manner; the terminal device monitors the position of construction personnel in real time, and issues a boundary crossing alarm when it detects that construction personnel have exceeded the work boundary; the terminal device receives the work procedure confirmation operation from the construction personnel, and sends the confirmation information back to the digital twin platform to update the actual progress data.

[0015] Furthermore, the identification of schedule deviation processes and the generation of resource scheduling instructions are as follows: Using construction components as basic units, the actual completion percentage of each component is compared with the planned completion percentage to calculate the deviation value; components with deviation values ​​exceeding a preset deviation threshold are marked as deviation components, and these deviation components are clustered according to construction area and professional type to obtain a set of deviation processes; the critical path influence coefficient is calculated for each deviation process, the critical path influence coefficient being determined based on the float time of the process in the construction network plan, with a smaller float time resulting in a larger influence coefficient; the deviation processes are sorted from largest to smallest influence coefficient, and an initial resource scheduling scheme is generated based on available data from the resource pool; the initial resource scheduling scheme is optimized according to preset priority scheduling rules, where clean area construction resources have the highest scheduling priority; the optimized resource scheduling scheme is converted into structured resource scheduling instructions and issued.

[0016] Furthermore, after the resource scheduling instruction is issued, the digital twin platform recalculates the estimated completion time of each deviation process based on the current actual progress data, the scheduled resource increments, and the remaining unfinished work, and compares the estimated completion time with the planned completion time to obtain the corrected progress deviation. If the corrected deviation is within a preset acceptable range, the process is marked as compensated. If the corrected deviation exceeds the acceptable range but does not exceed a preset warning threshold, it is marked as continuously monitored. If the corrected deviation exceeds the warning threshold, an early warning message is generated and pushed to the project control center. For processes marked as compensated, the consistency between their actual progress and predicted progress is continuously verified in multiple subsequent sampling periods, and the compensation status is turned off after verification.

[0017] Furthermore, during parallel operations, the particle concentration and vibration amplitude within the clean area are continuously monitored. When the particle concentration or vibration amplitude exceeds the corresponding preset threshold but does not exceed the preset emergency threshold, the system automatically triggers local environmental control measures, which include at least increasing the air supply volume or adjusting the timing of surrounding vibration operations. When the particle concentration or vibration amplitude exceeds the emergency threshold, the system automatically suspends the relevant construction operations in that area and issues an environmental warning. The suspension is automatically lifted after the environmental parameters return to within the threshold and stabilize for a preset time. All abnormal environmental events and system response actions are recorded in the environmental event log.

[0018] Furthermore, the weight coefficients of each domain in the three-domain coupled constraint model are adjusted in real time through an adaptive update step. This adaptive update step includes: based on actual construction conflict record data from historical sampling times within a preset time period, using a preset conflict severity mapping function to comprehensively invert discrete conflict events such as the number of work stoppages, rework times, and accuracy deviation repair times into measured conflict judgment values ​​for each historical time; simultaneously, based on the current candidate weight coefficients, recalculating the predicted judgment value for the monitoring data of the same historical time, and using the squared difference between the predicted judgment value and the measured conflict judgment value as the fitting error for that time; then, using the weight coefficients closer to the current... Based on the principle of assigning higher influence weights to historical data at each moment, an exponentially decaying forgetting factor is used to weight and sum the fitting errors at each moment. The goal is to minimize the sum of squared weighted cumulative errors. The optimal weight coefficients for the current moment are obtained by solving the problem using the recursive least squares optimization principle. During the solution process, the constraint that each weight coefficient is positive and the sum is always equal to 1 is used to ensure that the updated weight coefficients always maintain physical rationality. The updated weight coefficients replace the weight coefficients of the previous moment and are used to determine the parallel operation start conditions in subsequent moments, so as to realize the dynamic adaptive evolution of the weight coefficients of the three-domain coupled constraint model with the construction process.

[0019] Furthermore, in the step of generating resource scheduling instructions based on preset priority scheduling rules, a multi-objective collaborative optimization model is introduced. This model simultaneously optimizes three objectives under resource conflict conditions: minimizing project delay, minimizing scheduling costs, and minimizing cleanroom quality risk, and outputs a Pareto-optimal resource allocation scheme. The optimization logic of the multi-objective collaborative optimization model is as follows: based on the resource allocation decision vectors of each deviation process, the three objective values ​​corresponding to the allocation scheme are calculated: the total project delay value predicted based on the resource increment and remaining work volume of each deviation process; the additional total scheduling cost obtained by accumulating the unit scheduling costs of various resources; and the cleanroom quality risk index converted based on the intensity of resource disturbance allocated to processes surrounding the cleanroom. The calculated values ​​are then calculated using the planned allowable total project delay upper limit, the preset scheduling cost, and the... Based on the budget ceiling and the maximum tolerance value for cleanroom quality risk, the three target values ​​with different dimensions are normalized into dimensionless relative values. Then, the three normalized target values ​​are weighted and summed according to work option weight, cost weight, and quality risk weight to obtain the comprehensive cost index of the resource allocation scheme. With the feasible region constraint that the total allocation of all types of resources does not exceed the current available inventory and the amount of resources allocated to any process does not exceed the maximum acceptable capacity of that process, the optimal resource allocation decision vector is obtained by minimizing the comprehensive cost index. The work option weight, cost weight, and quality risk weight are all positive values, and their sum is always equal to 1. Each weight is preset by the user or automatically recommended by the system based on the current construction stage of the project. The optimal resource allocation scheme obtained is then converted into a structured resource scheduling instruction and issued for execution.

[0020] The beneficial effects of this invention are as follows: 1. Based on the constructed "structure-cleanroom-electromechanical" three-domain coupled digital twin and dynamic weight adaptive normalization judgment model, this invention can automatically and safely identify the parallel operation start windows and available time boundaries of cleanroom wall panel installation, raised floor laying and electromechanical main pipeline pre-embedding when the main structure is under construction to the middle stage. It can also guide multiple professional teams to safely and accurately insert construction synchronously before the main structure is capped by releasing instructions through the structured construction interface and superimposing the spatiotemporal boundaries of the augmented reality terminal.

[0021] 2. Based on the inversion of historical construction conflict data and the weight coefficient adaptive update mechanism of the forgetting weighted recursive least squares principle, the three-domain coupled judgment model can dynamically learn and self-correct the contribution weight of each domain as the construction progresses, avoiding the static nature and subjective bias of relying on engineers' experience to preset weights. At the same time, by introducing a three-objective collaborative optimization model of construction delay, scheduling cost and clean area quality risk, the Pareto optimal resource allocation scheme is solved through normalized weighted synthesis and constraint optimization under resource conflict conditions. This makes resource scheduling no longer limited to the single objective of shortest construction period, but also takes into account cost controllability and minimization of clean environment risk.

[0022] 3. Based on the comparison of progress deviations for each component and the ranking of critical path influence coefficients, scheduling instructions for prioritizing cleanroom resources are automatically generated. After resource scheduling, the completion time of the deviation process is re-predicted using a Bayesian sequential update method driven by resource efficiency coefficients. After three-level state judgment and continuous multi-cycle verification, the compensation state is officially closed to ensure that each scheduling is effectively tracked and evaluated. During the entire parallel operation, a graded response is further implemented for the particle concentration and micro-vibration amplitude in the cleanroom. When the level is slightly exceeded, the air volume is automatically increased or the vibration operation sequence is adjusted. When the level is seriously exceeded, the operation is automatically suspended and an early warning is issued. The warning is automatically lifted after the environment returns to stability, so that the cleanroom environment quality is always under control. Attached Figure Description

[0023] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is a flowchart of an embodiment of the present invention. Detailed Implementation

[0024] like Figure 1 As shown, this invention discloses a method for digital management and operation control of construction progress in ultra-large-scale factory buildings, including the following steps: S1. During the main structure construction phase, a multi-source heterogeneous sensor network deployed on the construction site continuously collects structural layer monitoring data, environmental layer monitoring data, and operational layer monitoring data at a sampling period of no more than 5 minutes. The structural layer monitoring data includes steel column stress values, concrete strain values, cumulative structural settlement values, and daily structural deformation. The environmental layer monitoring data includes temperature and humidity values ​​in the clean area, 5... The monitoring data includes particle size dust concentration, pressure difference, and micro-vibration amplitude; the operational layer monitoring data includes GPS positioning information of each professional team, equipment operating status parameters, and material entry barcode scanning records; all monitoring data are connected to edge computing nodes via 5G communication links, and after data cleaning and timestamp alignment by the edge computing nodes, they are uploaded to the digital twin platform in real time. S2. Based on the uploaded monitoring data, a coupled digital twin of the 3D BIM model and the 4D progress model is constructed in the digital twin platform. The coupled digital twin integrates the structural construction progress status, cleanroom installation logic constraints, and electromechanical pipeline path conflict detection results. In the coupled digital twin, a three-domain coupling constraint equation is established for the structural domain, cleanroom domain, and electromechanical domain. The three-domain coupling constraint equation uses the weighted combination value of the cumulative structural deformation, the temperature and humidity fluctuation of the cleanroom, and the electromechanical installation disturbance as the judgment index to dynamically evaluate whether the degree of mutual influence between the three domains under the current construction state meets the conditions for starting parallel operations. The expression for the three-domain coupling constraint equation is as follows: In the formula, For the current moment, As of The cumulative structural deformation at any given time is calculated by spatial interpolation of measurements taken from an array of fiber optic strain sensors deployed at key nodes of the main structure. These key nodes include at least the mid-span positions of each floor slab, the corner positions of the core tube, and the mid-span positions of large-span steel beams. The structural deformation tolerance threshold. As of The fluctuation of the cleanroom environment at any given time (unit: dimensionless, expressed as temperature deviation from set value) The absolute difference and humidity deviation from the set value The absolute difference is represented by a weighted normalized value. The temperature and humidity fluctuation is calculated from the measurements of an array of temperature and humidity sensors deployed in the pre-installation area of ​​the cleanroom. This is the tolerance threshold for environmental fluctuations. As of The electromechanical installation disturbance at any given time includes, at least, the weighted normalized result of the cumulative vibration energy generated by welding operations of electromechanical pipelines, the cumulative impact load generated by hoisting operations, and the cumulative deviation of reserved hole positions. This is the electromechanical disturbance tolerance threshold. , , These are the weighting coefficients for the structural domain, cleanroom domain, and electromechanical domain, respectively, satisfying... ;when When, it is determined that the conditions for starting a parallel job are met; when If the conditions for starting parallel operations are not met, the system will pause the release of new construction interfaces and issue an early warning signal. S3. When the calculation result of the three-domain coupling constraint equation meets the preset parallel operation window trigger threshold condition, the digital twin platform automatically identifies the construction interface types that can be released in the current construction stage. The construction interface types include at least the clean room wall panel installation surface, the raised floor laying surface, and the electromechanical main pipeline pre-embedded working surface. For each type of construction interface identified, the system calculates the available time window length and spatial operation boundary coordinates of the interface at the current time and spatial location based on real-time monitoring data, and generates a dynamic construction interface release command. S4. The dynamic construction interface release command is pushed to the augmented reality terminal devices of each relevant professional team. The augmented reality terminal devices display the spatial boundary of the specified construction interface, the allowed operation time window, and the process handover conditions in the real scene of the construction site in an overlay manner. Each professional team carries out parallel construction operations within the spatial boundary and time window to realize the synchronous insertion installation of the clean room enclosure system, raised floor, and electromechanical main pipeline when the main structure is constructed to the middle stage. S5. During parallel operations, the digital twin platform continuously collects actual progress data for each construction interface at a sampling period of no more than 5 minutes. The actual progress data is compared with the planned progress data component by component to identify progress deviation processes. When a progress deviation process is identified, the system calls the resource pool status data (the resource pool status data includes at least the remaining available manpower for each type of work, the number of available shifts for major construction equipment, and the inventory of prefabricated components), and generates resource scheduling instructions according to preset priority scheduling rules. In the priority scheduling rules, clean area construction resources have the highest scheduling priority. The system automatically sends the resource scheduling instructions to the corresponding resource management department and construction team, and updates the progress prediction status in the digital twin platform, forming a closed-loop control process of "monitoring-identification-scheduling-verification".

[0025] In one embodiment of the present invention, the parallel operation window trigger threshold condition includes independent and joint determinations of three dimensions: structural domain sub-condition, cleanroom domain sub-condition, and electromechanical domain sub-condition; the structural domain sub-condition is the daily structural deformation. And the cumulative settlement value of the structure ,in, It is obtained by differential calculation of structural deformation monitoring data from two adjacent sampling periods (each sampling period not exceeding 5 minutes). The clean zone sub-condition is calculated by spatially weighted averaging the measurements from an array of static levels at the support points of the raised floor; the clean zone sub-condition is clean zone 5. Particle size dust concentration And the amplitude of micro-vibration in the clean area ,in, The maximum value was obtained by taking the measurements from laser dust particle counters deployed at the boundary and inside the cleanroom's planned installation area. The root mean square value is obtained by taking the measurements from vibration sensors deployed at the same location; the electromechanical domain sub-condition is the positional deviation of the reserved holes in the electromechanical main pipe gallery. Furthermore, the vibration amplitude in the electromechanical pipeline welding work area ,in, The deviation was calculated after point cloud registration between the actual location of the reserved hole and the design location in the BIM model using a 3D laser scanner. The peak values ​​are obtained by measuring the vibration values ​​deployed around the welding work area. The joint determination is as follows: if and only if all the above-mentioned structural domain sub-conditions, clean domain sub-conditions and electromechanical domain sub-conditions are met, the system determines that the parallel operation window trigger condition is met and releases the corresponding construction interface. If any sub-condition is not met, the system marks the domain type and specific excessive parameters that do not meet the conditions in the digital twin platform, and generates parameter adjustment suggestions according to the preset priority order (clean domain first, structural domain first, electromechanical domain first). The parameter adjustment suggestions include adjusting the steel structure lifting rhythm to reduce the structural deformation rate, starting the temporary clean area positive pressure control system to reduce the dust particle concentration, or adjusting the electromechanical welding operation sequence to reduce the vibration superposition effect.

[0026] In one embodiment of the present invention, the generation process of the dynamic construction interface release command includes the following sub-steps: S3.1, The system is based on the current time. Calculation results of the three-domain coupling constraint equations The system retrieves the set of currently releaseable interface types from the preset construction interface type library, which includes at least the cleanroom metal wall panel installation interface, the cleanroom raised floor (including waffle module) laying interface, the electromechanical main air duct installation interface, the electromechanical main water pipe installation interface, and the electrical main cable tray installation interface. Each interface type is associated with the completion conditions of the preceding process, the required resource list, and the standard time quota. S3.2, For each releasable interface type retrieved The system calls upon the three-dimensional spatial coordinate data and construction logic constraint data of this interface, combined with the current structural construction progress (based on the number of floors of the main structure already completed). With the total number of floors (represented by the ratio), calculate the start time of the available time window for this interface. and the end time ,in Take the current time Compared with the expected completion time of the preceding process The larger value in It was calculated based on the constraints of the latest allowed start time of subsequent construction processes and the structural construction progress; S3.3, The system will use the available time window The spatial operation boundary coordinates of the interface (represented by the bounding box coordinates of the components in the BIM model) are associated and encapsulated to generate a structured construction interface release instruction. The instruction includes at least an interface type identifier, a set of spatial operation boundary coordinates (represented in three-dimensional point cloud coordinate format), allowed start and end timestamps of the operation, minimum safe distance requirements between adjacent interfaces, and quality acceptance standards for process handover. S3.4 The system pushes the structured construction interface release command to the augmented reality terminal equipment of the corresponding professional team and the digital twin platform large screen display system of the project's overall control center through the 5G communication link.

[0027] In one embodiment of the present invention, after the dynamic construction interface release command is pushed to the augmented reality terminal device, the augmented reality terminal device performs the following display and interaction operations: S4.1 The terminal device acquires real-time images of the construction site through its built-in camera, and obtains the spatial pose data (including three-dimensional position coordinates and three-axis attitude angles) of the current device by combining the built-in inertial measurement unit and ultra-wideband positioning module. The positioning accuracy of the spatial pose data is better than ±5cm, and the attitude angle accuracy is better than ±1°. S4.2 The terminal device converts the spatial operation boundary coordinate set in the received structured construction interface release command from the BIM model coordinate system to the construction site geodetic coordinate system. Then, based on the current spatial pose data of the device, it projects the operation boundary in the geodetic coordinate system onto the real scene image on the device display screen, and displays the operation boundary outline, the countdown of the allowed operation time window, and the warning line of the minimum safe distance between adjacent interfaces in the form of a color semi-transparent overlay. S4.3 The terminal device displays the standard construction procedure flow, key quality acceptance control points and preset safety operation procedure prompts in the form of arrows and text labels in the area inside the work boundary. The standard construction procedure flow is automatically generated based on the standard time quota and process specifications associated with the interface type. S4.4 The terminal equipment monitors in real time whether construction personnel enter or exceed the designated work boundary. When construction personnel are detected to have exceeded the work boundary, the terminal equipment will issue a boundary crossing alarm through vibration prompts and a red flashing border on the display screen, and will send the time, location and personnel information of the boundary crossing event back to the digital twin platform for recording. S4.5 Construction personnel confirm the start and completion of a work process through the touch interface of the terminal device. The terminal device sends the confirmation information along with a timestamp back to the digital twin platform to update the actual progress data of the construction interface.

[0028] In one embodiment of the present invention, the process of identifying the schedule deviation process and generating resource scheduling instructions includes the following sub-steps: S5.1 The digital twin platform uses each component (which includes at least steel columns, steel beams, concrete floor slabs, cleanroom wall panels, raised floor panels, and electromechanical pipeline sections) as a basic unit, and compares the actual progress data with the planned progress data component by component to calculate the progress deviation value for each component. ,in, For components Deadline The cumulative percentage of the plan completed; For components Deadline The actual cumulative percentage of completion; when At that time, the component is marked as a deviation component; S5.2 The system clusters all deviation components according to their construction area and professional type, and identifies the set of deviation processes. And for each deviation process Calculate its critical path impact coefficient The critical path influence coefficient Based on the floating time of this process in the overall construction network plan calculate: ,in, The preset time constant (value is 5 working days). The smaller The larger the value, the more severe the impact of the deviation in that process on the overall project duration; S5.3, System Basis The deviation processes are prioritized from largest to smallest, and the resource pool status data (obtained through the real-time data interface of each resource management subsystem, including the remaining available man-hours of human resources for each type of work, the number of available shifts of major construction equipment, and the inventory of prefabricated components) is called up to match the available resource types and quantities for each deviation process, and an initial resource scheduling plan is generated. S5.4 The system optimizes and adjusts the initial resource scheduling scheme according to preset priority scheduling rules, which include: (a) Cleanroom construction resources (including cleanroom wall panel installers, waffle module and raised floor panels) have the highest scheduling priority, and no non-cleanroom process may occupy the cleanroom resources already allocated; (b) Critical path impact coefficient The deviation process has the second highest scheduling priority; (c) Resource scheduling should aim to minimize overall project time loss, and when resource conflicts occur, priority should be given to ensuring the resource needs of high-priority processes; S5.5 The system converts the optimized resource scheduling scheme into structured resource scheduling instructions. The instructions include at least the target process identifier, the type and quantity of scheduled resources, the resource arrival time requirement, and the scheduling priority level. The instructions are automatically sent to the terminal devices of the corresponding resource management departments and construction teams through the 5G communication link.

[0029] In one embodiment of the present invention, the progress prediction status update in the closed-loop control process is achieved through the following method: After each resource scheduling instruction is issued, the digital twin platform, based on the current time... Actual progress data and scheduled resource increments and the remaining unfinished work The expected completion time of each deviation process is recalculated using the Bayesian sequential update method. : In the formula, For process The amount of resources currently allocated (expressed in equivalent working hours). For process The resource efficiency coefficient (obtained based on historical output data of similar processes). The system will update the estimated completion time. Completion time of the plan Compare and calculate the corrected schedule deviation value. ;when When the process has been effectively compensated through resource scheduling, it is determined that the process has been effectively compensated, and the process is marked as "compensated - normal" in the digital twin platform; when If the process is deemed to require continuous monitoring, it is marked as "Compensation in Progress - Observation," and the system continues to track the actual progress of the process at a sampling period of no more than 5 minutes; when If the current resource scheduling plan is deemed insufficient to compensate for the schedule deviation, the system will automatically upgrade to a yellow alert and push the alert information to the project's central control center. For processes marked as "compensated-normal", the system continuously verifies whether the actual progress is consistent with the updated predicted progress in the next three consecutive sampling cycles (each sampling cycle is no more than 5 minutes). If the verification is successful, the process is removed from the deviation process set and the compensation status is turned off; if the verification fails, the deviation identification process is re-entered.

[0030] In one embodiment of the present invention, the method further includes a step of dynamically maintaining and restoring the cleanroom environment quality during parallel operations: After the cleanroom wall panel installation interface and raised floor laying interface are released, the system continuously monitors the 5 clean areas at a sampling period of no more than 1 minute. Particle size dust concentration and micro-vibration amplitude The monitored values ​​will be compared with the clean area sub-condition thresholds in real time. Perform a comparison; When detected However, if the increase does not exceed 10%, the system will automatically trigger a local clean area positive pressure control enhancement command to increase the air supply volume of the area to 120% of the rated value for a duration of not less than 30 minutes. When detected However, when the increase exceeds 10%, the system will automatically suspend all parallel construction operations in the area and issue a red alert for cleanliness. The system will automatically lift the suspension and notify the relevant work teams to resume operations after the dust particle concentration returns to below the threshold and remains stable for no less than 15 minutes. When detected However, when the increase does not exceed 20%, the system will automatically adjust the construction sequence of electromechanical installation work in the surrounding area, and transfer welding and cutting operations that generate high-frequency vibrations to a time period or location far away from the clean area. When detected However, when the increase exceeds 20%, the system will automatically suspend all vibration-generating operations in the area (including welding of electromechanical pipelines and hoisting of large equipment) and issue a red warning for micro-vibration. The system will automatically lift the suspension state after the vibration amplitude returns to below the threshold and remains stable for no less than 10 minutes. All environmental anomalies (including cleanliness exceeding standards, micro-vibration exceeding standards, and their corresponding system response actions) are recorded in the environmental event log database of the digital twin platform. The logs include at least the event occurrence time, the type and value of the exceeding parameters, the type of system response action, the response execution time, and the recovery confirmation time, which are used for subsequent construction environmental quality analysis and process optimization.

[0031] In one embodiment of the present invention, the weight coefficients of each domain in the three-domain coupling constraint model are adjusted in real time through an adaptive update step, the adaptive update step including: Obtain historical judgment datasets within a preset sliding time window. The historical judgment data includes at least the actual monitored cumulative structural deformation, clean area environmental fluctuations, and electromechanical installation disturbances at each historical moment, as well as actual construction conflict record data at each historical moment. The actual construction conflict record data includes at least the number of work stoppages caused by exceeding cleanliness standards, the number of reworks caused by exceeding structural deformation limits, and the number of accuracy deviation repairs caused by electromechanical installation disturbances. Based on the historical judgment dataset, a recursive least squares algorithm with a forgetting factor is used to update the weight coefficient vector at the current time online. The update process is defined by the following optimization objective function: In the formula, This is the index number of the historical sampling time. The number of historical sampling points included in the preset sliding time window. It is a forgetting factor and satisfies Used to control the decay rate of historical data, when The closer the value is to 1, the higher the degree of retention of historical data. The smaller the value, the greater the weight of recent data; The first candidate weight coefficient is calculated based on the current candidate weight coefficient. The predicted value at any given time; ; For the first The actual construction conflict inversion judgment value at any given moment. According to the The actual construction conflict records at any given time are calculated using a preset conflict severity mapping function; The update process also needs to meet constraints. and , , Updated weighting coefficients The weight coefficients of the previous time step are used to determine the conditions for starting parallel operations in subsequent time steps, thereby realizing the dynamic adaptive evolution of the weight coefficients of the three-domain coupled constraint model as the construction process progresses.

[0032] In one embodiment of the present invention, in the step of generating resource scheduling instructions according to preset priority scheduling rules, a multi-objective collaborative optimization model is introduced. This model simultaneously optimizes three objectives—minimizing project delay, minimizing scheduling costs, and minimizing cleanroom quality risks—under resource conflict conditions, and outputs a Pareto-optimal resource allocation scheme. The objective function expression of the multi-objective collaborative optimization model is: In the formula, For a moment The resource allocation decision vector, , Indicates assignment to the first The first deviation process The number of class resources, This represents the total number of processes with the current deviation. Total number of resource types; This is the total project delay value (in days) predicted based on the current resource allocation plan. This value is obtained by summing the differences between the expected completion time and the planned completion time of each deviation process. This is the maximum allowable total project delay (in days), used to normalize project delay items; The total additional scheduling cost (in yuan) incurred by the current resource allocation plan, which includes at least overtime labor costs, equipment rental premium costs, and emergency replenishment costs of prefabricated components; This is a preset upper limit for the scheduling cost budget (in yuan), used to normalize cost items; The risk index (dimensionless, ranging from 0 to 1) for the environmental quality of the clean area under the current resource allocation scheme is calculated based on the expected disturbance intensity generated by the resources allocated to the construction process near the clean area. The greater the disturbance intensity, the higher the risk index. The preset maximum tolerance value for cleanroom quality risk (dimensionless, value is 1) is used to normalize risk items; , , These are the target weighting coefficients for construction delays, scheduling costs, and cleanroom quality risks, respectively. The resource allocation decision vector The feasible domain is subject to the following constraint: the total allocation of all types of resources does not exceed the current available inventory of the corresponding resource type, i.e. ,in, For the first Class resources at any time Available inventory levels; the quantity of each type of resource allocated to any process shall not exceed the maximum acceptable resource quantity for that process, i.e. To avoid diminishing marginal efficiency or construction site congestion caused by excessive resource concentration, the multi-objective collaborative optimization model outputs the corresponding optimal resource allocation decision vector by solving for the minimum value of the above objective function within the feasible region. The optimal allocation scheme is then converted into structured resource scheduling instructions and issued for execution.

[0033] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. A method for digital management and operation control of construction progress in ultra-large-scale factory buildings, characterized in that, Includes the following steps: S1. Obtain multi-source monitoring data during the construction phase. The multi-source monitoring data should include at least structural deformation data, clean area environmental data, and electromechanical installation disturbance data. Upload multi-source monitoring data to the digital twin platform; S2. Construct a coupled digital twin of the 3D building information model and the schedule model in the digital twin platform. The coupled digital twin integrates the structural construction status, cleanroom installation logic constraints, and electromechanical pipeline conflict detection information. Based on the coupled digital twin, establish a three-domain coupled constraint model of the structural domain, cleanroom domain, and electromechanical domain. The three-domain coupled constraint model is used to evaluate the degree of mutual influence between structural deformation, cleanroom environment fluctuations, and electromechanical installation disturbances. S3. Based on the evaluation results of the three-domain coupling constraint model, determine whether the current construction status meets the preset parallel operation start conditions; if it does, automatically identify the types of construction interfaces that can be released. The construction interface types include at least the cleanroom wall panel installation interface, the raised floor laying interface, and the electromechanical pipeline pre-embedded working surface; for the identified construction interface, calculate its available time window and spatial operation boundary, and generate a construction interface release command. S4. Push the construction interface release command to the terminal devices of each relevant construction team. The terminal devices display the spatial boundary and time window information of the construction interface on the construction site in an augmented reality manner, guiding each professional team to carry out parallel construction operations within the spatial boundary and time window. S5. During parallel operations, continuously collect actual progress data from each construction interface, compare the actual progress with the planned progress, and identify processes with progress deviations. For the identified deviation processes, call the resource pool status data, generate resource scheduling instructions according to the preset priority scheduling rules, and issue them to compensate for the deviation processes with resources.

2. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that, The three-domain coupled constraint model constructs the judgment logic in the following way: based on the measured cumulative structural deformation, clean area environmental fluctuation, and electromechanical installation disturbance at the current moment, normalization is performed on each of the three physical quantities with different dimensions as a benchmark, so that the three physical quantities with different dimensions are uniformly converted into dimensionless relative deviation values. The three normalized values ​​are then weighted and summed according to the weight coefficients corresponding to the structural domain, cleanroom domain, and electromechanical domain, respectively, to calculate the comprehensive coupling judgment index at the current moment. If the comprehensive coupling judgment index at the current moment is not greater than a preset threshold, it is determined that the current parallel operation start condition is met; otherwise, it is determined that it is not met. The weight coefficients of the structural domain, cleanroom domain, and electromechanical domain are all positive values, and the sum of the three is always equal to 1. Each weight coefficient is dynamically adjusted according to the construction stage of the plant. In the early stage of the main structure construction, the weight value of the structural domain is relatively large, and in the middle and later stages of the main structure construction, the weight values ​​of the cleanroom domain and electromechanical domain gradually increase to adapt to the different degrees of dominance of each domain on the constraints of parallel operation at different construction stages.

3. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: The parallel operation initiation conditions include structural domain determination conditions, cleanroom domain determination conditions, and electromechanical domain determination conditions. The structural domain determination condition is that the structural deformation amount does not exceed a preset structural deformation threshold. The cleanroom domain determination condition is that the particle concentration in the clean area does not exceed a preset particle concentration threshold and the vibration amplitude in the clean area does not exceed a preset vibration threshold. The electromechanical domain determination condition is that the position deviation of the reserved hole does not exceed a preset deviation threshold and the vibration amplitude in the electromechanical operation area does not exceed a preset electromechanical vibration threshold. The parallel operation initiation condition is determined to be valid only when all three domain determination conditions are met. If any condition is not met, the system marks the domain type that is not met and generates a corresponding adjustment suggestion.

4. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: The method for generating the construction interface release instruction is as follows: based on the current construction progress and the completion status of the preceding procedures of each construction interface, select the currently releaseable interface type from the preset construction interface type library; for each releaseable interface, determine the available time window based on the completion time of its preceding procedures and the latest start time of its subsequent procedures; determine the spatial operation boundary based on the three-dimensional model coordinates of the interface; and encapsulate the time window and spatial boundary into a structured instruction, which at least includes an interface type identifier, a set of spatial boundary coordinates, the start and end times of the time window, and the procedure handover requirements.

5. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: The process of displaying the construction interface in augmented reality by the terminal device includes: acquiring the real-time spatial pose of the terminal device; converting the spatial boundary coordinates in the received construction interface release command from the model coordinate system to the construction site coordinate system, and projecting the spatial boundary onto the real-world image of the terminal device according to the spatial pose of the terminal device, displaying the work boundary outline, time window information, and safety distance warning in an overlay manner; the terminal device monitors the position of construction personnel in real time, and issues a boundary crossing alarm when it detects that construction personnel have exceeded the work boundary; the terminal device receives the work procedure confirmation operation from the construction personnel and sends the confirmation information back to the digital twin platform to update the actual progress data.

6. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: The method for identifying schedule deviation processes and generating resource scheduling instructions is as follows: Using construction components as basic units, the actual completion percentage of each component is compared with the planned completion percentage to calculate the deviation value; components with deviation values ​​exceeding a preset deviation threshold are marked as deviation components, and these deviation components are clustered according to construction area and professional type to obtain a set of deviation processes; the critical path influence coefficient is calculated for each deviation process, the critical path influence coefficient being determined based on the float time of the process in the construction network plan, with a smaller float time resulting in a larger influence coefficient; deviation processes are sorted from largest to smallest influence coefficient, and an initial resource scheduling scheme is generated based on available data from the resource pool; the initial resource scheduling scheme is optimized according to preset priority scheduling rules, where clean area construction resources have the highest scheduling priority; the optimized resource scheduling scheme is converted into structured resource scheduling instructions and issued.

7. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: After the resource scheduling instruction is issued, the digital twin platform recalculates the estimated completion time of each deviation process based on the current actual progress data, the scheduled resource increments, and the remaining unfinished work, and compares the estimated completion time with the planned completion time to obtain the corrected progress deviation. If the corrected deviation is within a preset acceptable range, the process is marked as compensated. If the corrected deviation exceeds the acceptable range but does not exceed a preset warning threshold, it is marked as continuously monitored. If the corrected deviation exceeds the warning threshold, an early warning message is generated and pushed to the project control center. For processes marked as compensated, the consistency between their actual progress and predicted progress is continuously verified in multiple subsequent sampling periods. Once the verification is successful, the compensation status is turned off.

8. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: During parallel operations, the particle concentration and vibration amplitude within the clean area are continuously monitored. When the particle concentration or vibration amplitude exceeds the corresponding preset threshold but does not exceed the preset emergency threshold, the system automatically triggers local environmental control measures, which include at least increasing the air supply volume or adjusting the timing of surrounding vibration operations. When the particle concentration or vibration amplitude exceeds the emergency threshold, the system automatically suspends the relevant construction operations and issues an environmental warning. The suspension is automatically lifted after the environmental parameters return to within the threshold and stabilize for a preset time. All abnormal environmental events and system response actions are recorded in the environmental event log.

9. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: In the three-domain coupled constraint model, the weight coefficients of each domain are adjusted in real time through an adaptive update step. This adaptive update step includes: based on actual construction conflict record data from historical sampling times within a preset time period, using a preset conflict severity mapping function to comprehensively invert the number of work stoppages, rework times, and accuracy deviation repair times into measured conflict judgment values ​​for each historical time; simultaneously, based on the current candidate weight coefficients, recalculating the predicted judgment value for the monitoring data of the same historical time, and using the squared difference between the predicted judgment value and the measured conflict judgment value as the fitting error for that time; then assigning weights to historical data closer to the current time. Based on the principle of high influence weight, an exponentially decaying forgetting factor is used to weight and sum the fitting errors at each time step to obtain the weighted cumulative error sum of squares. With the goal of minimizing the weighted cumulative error sum of squares, the optimal weight coefficients at the current time step are obtained by solving the recursive least squares optimization principle. During the solution process, the constraint condition is that all weight coefficients are positive and the sum is always equal to 1, to ensure that the updated weight coefficients always maintain physical rationality. The updated weight coefficients replace the weight coefficients of the previous time step and are used to determine the parallel operation start conditions in subsequent time steps, so as to realize the dynamic adaptive evolution of the weight coefficients of the three-domain coupling constraint model with the construction process.

10. The method for digital management and operation control of construction progress of ultra-large-scale factory buildings according to claim 1, characterized in that: In the step of generating resource scheduling instructions based on preset priority scheduling rules, a multi-objective collaborative optimization model is introduced. This model simultaneously optimizes three objectives—minimizing project delay, minimizing scheduling costs, and minimizing cleanroom quality risk—under resource conflict conditions, and outputs a Pareto-optimal resource allocation scheme. The optimization logic of the multi-objective collaborative optimization model is as follows: based on the resource allocation decision vector of each deviation process, the three objective values ​​corresponding to the allocation scheme are calculated respectively: the total project delay value predicted based on the resource increment and remaining work volume of each deviation process, the additional total scheduling cost obtained by accumulating the unit scheduling costs of various resources, and the cleanroom quality risk index converted based on the resource disturbance intensity allocated to the processes surrounding the cleanroom. Based on the planned maximum allowable total project delay, the preset scheduling cost budget limit, and the maximum tolerable cleanroom quality risk, the three objective values ​​with different dimensions—total project delay, total additional scheduling cost, and cleanroom quality risk index—are normalized into dimensionless relative values. Then, the three normalized objective values ​​are weighted and summed according to work option weight, cost weight, and quality risk weight to obtain the comprehensive cost index of the resource allocation scheme. With the feasible region constraint that the total allocation of all types of resources does not exceed the current available inventory and the resource allocation to any process does not exceed the maximum acceptable capacity of that process, the optimal resource allocation decision vector is obtained by minimizing the comprehensive cost index. The work option weight, cost weight, and quality risk weight are all positive values, and their sum is always equal to 1. Each weight is preset by the user or automatically recommended by the system based on the current construction stage of the project. The obtained optimal resource allocation scheme is then converted into a structured resource scheduling instruction and issued for execution.

Images