Road construction BIM-digital twinning integration progress quality linkage management and control method and device

Abstract

The invention discloses a road construction BIM-digital twinning integration progress quality linkage management and control method and device, and belongs to the technical field of intelligent construction and digital engineering management. The method is used for solving the problem that in the prior art, progress and quality lack bidirectional dynamic coupling and closed loop feedback under a BIM-digital twinning frame. A progress-quality combined state space model based on multi-dimensional space-time constraints is constructed, a cooperative control framework composed of five engines including progress driving, quality verification, deviation analysis, model correction and plan rearrangement is designed, and closed-loop regulation and control of progress execution, quality verification, deviation early warning, model correction and plan rearrangement are achieved through a unified data bus. According to the method, deep coupling, dynamic linkage and self-adaptive optimization of progress and quality can be realized, and the controllability and robustness of the highway construction process are improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent construction and digital twin technology, specifically relating to a method and device for integrated progress and quality control of highway construction using BIM-digital twin. Background Technology

[0002] The integration of Building Information Modeling (BIM) and digital twin technology is considered a key path to improve the level of refined management in highway engineering construction. BIM provides a structured information foundation for the entire lifecycle of a project, while digital twin technology, through real-time perception and dynamic simulation, enables physical entities to have calculable and interactive digital mapping capabilities. This lays the technological foundation for realizing the visualization, analysis, and intervention capabilities of the construction process.

[0003] Patent CN120851442A discloses a method for constructing a risk matrix by integrating multi-source sensor data and using an artificial intelligence model in a digital twin environment for construction safety simulation and anomaly visualization, primarily focusing on proactive early warning of structural safety risks. Patent CN120124929A discloses dynamic maintenance of component-level digital assets. By encoding and linking design and construction data, it predicts the component qualification probability based on real-time data streams to trigger model updates, improving the consistency between the digital model and the physical entity.

[0004] However, structural limitations remain in addressing the engineering challenges of coordinating schedule control and quality management. Traditional solutions follow a paradigm of unidirectional data-driven or static mapping, failing to establish a bidirectional, dynamic, and quantitative coupling relationship between schedule execution status and quality acceptance results. For example, while the aforementioned safety early warning scheme can identify risks, it fails to establish a causal link between risk events (such as process delays) and the quality formation conditions of subsequent processes (such as the time window for basic maintenance); and while the asset update scheme can synchronize the model, it lacks a mechanism to trigger and optimize subsequent schedule plans based on actual quality measurement data (such as insufficient compaction).

[0005] Current technologies lack a complete cybernetics loop that connects schedule execution to quality verification, deviation diagnosis, model correction, and schedule rescheduling. This results in a disconnect between schedule and quality management logic and data flow, with either remaining parallel or reactive. It hinders collaborative perception, causal tracing, and coordinated optimization in the early stages of deviations, making it difficult to address the dynamic challenges posed by multiple overlapping processes, strict resource constraints, and sensitive quality windows in complex highway construction projects. Therefore, constructing an integrated intelligent management and control mechanism that achieves deep coupling, real-time linkage, and automatic correction of schedule and quality within a BIM-digital twin framework has become a critical technological bottleneck that urgently needs to be overcome in this field. Summary of the Invention

[0006] This invention discloses a method and device for integrated BIM-digital twin-based progress and quality control in highway construction, which solves the problem that existing technologies lack bidirectional dynamic coupling and closed-loop feedback mechanisms between progress planning and project quality within the BIM-digital twin framework.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0008] This invention provides a method for integrated BIM-digital twin-based progress and quality control in highway construction, comprising the following steps:

[0009] Step 1: Construct a schedule-quality joint state space model using BIM components as basic units. Each BIM component is uniquely identified by the IfcGloballyUniqueId defined in the ISO 12006-2 standard and associated with a four-tuple state vector. ;

[0010] in: This is the theoretical completion timestamp for the component in the schedule. This is the actual completion timestamp. The set of quality indicator thresholds specified in the design specifications. The overall quality score is calculated based on on-site measured data;

[0011] Step 2: Deploy a five-element collaborative control architecture consisting of a progress-driven engine, a quality verification engine, a deviation analysis engine, a model correction engine, and a schedule rescheduling engine. Each engine achieves status synchronization and command interaction through a unified data bus.

[0012] Step 3: The progress-driven engine dynamically generates the process execution instruction flow based on the critical chain method, and updates the... Pushed to the digital twin platform;

[0013] Step 4: After receiving the process completion confirmation signal, the quality verification engine performs a process verification based on multi-source sensor data collected by the on-site sensor network. Perform the calculations and write the results into the joint state space model;

[0014] Step 5: The deviation analysis engine is based on schedule deviation. With quality deviation Calculations are performed, including the quality deviation. according to Calculations show that (when (When it is a Boolean value or a normalized pass rate) or (Calculated based on specific indicators), combined with the preset process dependency impact matrix, the comprehensive impact index of subsequent processes is calculated, and a high-risk warning is generated when the comprehensive impact index exceeds the threshold.

[0015] in: For the quantity of quality indicators, For the first The measured values ​​of each quality indicator For the first Standard thresholds for each quality indicator;

[0016] Step Six: The model correction engine performs geometric or attribute corrections on the BIM model based on the deviation type and generates a version snapshot;

[0017] Step 7: The schedule rescheduling engine solves a mixed-integer linear programming problem based on the modified model, resource availability, and environmental constraints (including meteorological data for a future period obtained from the meteorological API) to generate an optimized schedule, which is then fed back to the schedule-driven engine to form a closed-loop control process.

[0018] Furthermore, the schedule-quality joint state-space model uses BIM components as basic units, with each component associated with a four-tuple state vector:

[0019] ;

[0020] in: This is the theoretical completion timestamp for the component in the schedule. This is the actual completion timestamp. The set of quality indicator thresholds specified in the design specifications. The comprehensive quality score is calculated based on on-site measured data; the basic set of measured data is... ( This refers to the quantity of quality indicators, such as compaction degree and flatness.

[0021] The overall quality score From the set of measured values The following calculation method was used:

[0022] For each quality indicator (e.g., compaction, flatness, strength, etc.), define the single-point quality scoring function for this index: ;

[0023] in: For the first Single-point quality score for each quality indicator. For the first The standard pass threshold for each quality indicator, For the first The minimum allowable threshold for each quality indicator;

[0024] Overall quality score of components Take the minimum score of each indicator: .

[0025] The state vector is bound to the geometric and semantic information in the BIM model through a unique component code (using IfcGloballyUniqueId as defined by the ISO12006-2 standard) and stored in a distributed time-series database to ensure the traceability and timeliness of the state data.

[0026] In a preferred embodiment of the present invention, the schedule-driven engine is configured on a central control server. Its input is the project master schedule (a logical network diagram parsed from an XER file exported in Primavera P6 format), and its output is a daily-granular decomposition of process execution instructions. The schedule-driven engine internally includes a scheduler optimized based on the Critical Chain Method (CCM), whose core algorithm is:

[0027] ;

[0028] in: Number the current process; For process The set of all preceding processes; For process After completion, proceed to the next process The minimum process interval required to begin is determined according to the construction specifications. When resource constraints are not considered, the process The earliest possible start time is calculated by forward traversing the process network diagram; To implement a critical chain buffer management strategy for processes The allocated floating time; For process The project completion timestamp;

[0029] The scheduler recalculates all incomplete processes every 24 hours or upon receiving an external trigger signal (such as a weather warning or equipment failure). The updated plan will then be pushed to the progress visualization module of the digital twin platform.

[0030] Furthermore, the quality verification engine is deployed on edge computing nodes and directly interfaces with the field sensor network. This sensor network includes, but is not limited to: an intelligent road roller with a built-in accelerometer and GNSS positioning module for roadbed compaction detection; an infrared thermal imaging array for asphalt paving temperature monitoring; embedded strain gauges and temperature and humidity probes for concrete strength assessment; and a mobile 3D laser scanner for geometric dimension verification. All sensors are connected to the edge computing nodes via industrial Ethernet or a 5G private network, and the sampling frequency and accuracy meet the following technical specifications: compaction sensor sampling rate not less than 10Hz, positioning error less than ±2cm; infrared thermal imager spatial resolution of 640×480 pixels, temperature measurement range of 0–300℃, accuracy of ±1.5℃; strain gauge range of ±5000με, nonlinearity error less than 0.1%FS; laser scanner point cloud density not less than 1000 points / square meter, distance accuracy better than ±1mm.

[0031] The quality verification engine receives the list of processes to be verified and their corresponding data from the progress-driven engine. and and in Once confirmed, the verification process is initiated; when the actual completion time of a certain roadbed filling section... After being confirmed by the on-site monitoring terminal, the system immediately extracts the location of the road segment from the edge computing node. Compaction trajectory data within the time window, including: For a preset process duration (e.g., for subgrade filling in Zone 96), (hours); subsequently, the engine executes the following quality assessment algorithm:

[0032] Reconstructing the roller pass distribution map based on GNSS trajectory;

[0033] Calculate the average number of compaction passes for each 1m×1m grid cell. ;

[0034] like And coefficient of variation If so, the compaction quality of the unit is deemed qualified;

[0035] The pass rate of the entire road section ;

[0036] in: The average number of compaction passes per grid cell. To meet the minimum number of compaction passes required by the design specifications, for standard deviation for The mean, The coefficient of variation is 1. The maximum permissible coefficient of variation, For the road section qualification rate, The pass rate threshold For overall quality scoring;

[0037] like (For example, 95%), then If assigned a value of 1, otherwise assigned a value of 0;

[0038] The determination result, along with the original point cloud and sensor data, is written into the progress-quality joint state space model and triggers the deviation analysis engine.

[0039] As a key innovation of this invention, the deviation analysis engine adopts a two-dimensional deviation measurement mechanism; the first dimension is schedule deviation. The second dimension is quality deviation. ;like If it is a Boolean value or a normalized pass rate, then ;like Based on the comprehensive calculation of multiple indicators, then ,in For the quantity of quality indicators.

[0040] The engine maintains a process dependency matrix internally. The process-dependent influence matrix During the project initialization phase, based on the component topology relationships in the BIM model, the process logic determined by the construction organization design, and historical engineering data, data is pre-generated through expert scoring or data fitting and stored in the central control server; when a certain process... Appear or When (i.e., quality is not up to standard), the system calculates its impact on all subsequent processes. Overall impact index:

[0041] ;

[0042] in: That is, the time-dependent component directly adopts the element values ​​of the process-dependent influence matrix;

[0043] The quality affects the quantity. As a quality influencing factor, based on the preceding process. The determination of the type and severity of quality defects is based on the following table:

[0044] Table 1:

[0045]

[0046] Defect severity assessment criteria:

[0047] Slight: The measured value is between 90% and 95% of the standard threshold;

[0048] Moderate: Measured values ​​are between 80% and 90% of the standard threshold;

[0049] Serious: The measured value is 80% lower than the standard threshold;

[0050] This is the quality-schedule coupling coefficient (default value is 2.0, reflecting the amplification effect of quality defects on subsequent processes); if (For example (This threshold is set based on historical engineering data and expert experience), then the process is marked. This is a high-risk process, and an early warning event is generated.

[0051] Furthermore, the model correction engine responds to the output of the deviation analysis engine and dynamically updates the BIM model. The update operation is divided into two categories: geometric correction and attribute correction. Geometric correction is applicable to situations where the geometric shape of the entity changes due to construction deviations, such as over-excavation or under-filling of roadbed slopes. In this case, the system calls the latest point cloud data obtained by the mobile laser scanner, aligns it with the original BIM model through the ICP (IterativeClosestPoint) algorithm, generates a deviation chromatogram, and automatically reconstructs areas exceeding the tolerance zone (e.g., ±5cm) into new IfcEarthworksFill or IfcEarthworksCut entities, replacing the corresponding parts of the original model. Attribute correction is for changes in quality status, such as updating the QualityStatus attribute in the IfcPropertySet of a certain road structure layer from Pending to Failed, and attaching a failure reason code (such as CompactionInsufficient) and a reference link to the measured data.

[0052] All model correction operations are recorded through transaction logs and version snapshots are generated to ensure that the model evolution process is auditable; the corrected BIM model is synchronized in real time to the 3D visualization engine of the digital twin platform for managers to view.

[0053] As the core closed-loop component of this invention, the plan rescheduling engine initiates an adaptive rescheduling process upon receiving a model correction event; the adaptive rescheduling process identifies the set of affected processes. Based on the revised resource availability (including manpower, equipment, and material inventory) and environmental constraints (such as the probability of rainfall in the 72-hour weather forecast), the following mixed-integer linear programming (MILP) problem is resolved:

[0054] ;

[0055] The constraints include:

[0056] Process logic constraints: ;

[0057] Resource capacity constraints: ;

[0058] Quality window constraint: If the process Quality depends on process If the condition is qualified, then ;

[0059] This is a waiting period for quality to stabilize (for example, after the base layer is compacted, you need to wait 24 hours before paving the next layer).

[0060] Key milestone conservation: The end of the total project period shall not be later than the contract completion date;

[0061] in: The process weight is 1.5 for the critical path and 1.0 for non-critical paths. For the first Excessive use of such resources The resource penalty coefficient is used; the solver is the commercial CPLEX22.1 solver, and the time limit is set to 300 seconds; the rearranged results are output as an updated XER file and sent back to the progress-driven engine to complete the closed loop.

[0062] This invention constructs and maintains a historical similar working condition database to improve the intelligence of decision-making. When generating high-risk warnings, the deviation analysis engine can simultaneously retrieve historical similar cases and their handling effects from the database for management reference. When solving optimization problems, the scheduling engine can also use successful scheduling schemes under similar historical working conditions as initial solutions or constraint references to accelerate the solution process. The method for constructing the historical similar working condition database is as follows:

[0063] Feature vector definition: The feature vector for each historical operating condition is... There are a total of 7 dimensions;

[0064] Feature standardization: Z-score standardization is performed on numerical features (temperature, humidity). ;

[0065] Similarity metric: Weighted cosine similarity is used.

[0066] ;

[0067] Where the weight vector ;

[0068] Matching threshold: when When the condition is similar, it is determined to be a similar working condition.

[0069] Furthermore, the device described in this invention includes a central control server, edge computing nodes, a field sensor network, a supervisory mobile terminal, and a digital twin visualization terminal. The central control server is equipped with an Intel Xeon Gold 6348 processor (28 cores and 56 threads), 256GB DDR4 ECC memory, and 2TB NVMe SSD storage. It runs an Ubuntu 22.04 LTS operating system and deploys a Docker containerized microservice architecture. Each engine runs as an independent container and communicates via the gRPC protocol. The edge computing nodes use NVIDIA Jetson AGX Orin modules, are equipped with 32GB LPDDR5 memory, run a custom Linux kernel, and support a lightweight quality assessment model accelerated by TensorRT. The field sensor network interfaces with the edge computing nodes via the OPCUA protocol to ensure data semantic interoperability. The supervisory mobile terminal is a ruggedized Android tablet with a dedicated APP installed. It supports NFC tag scanning to confirm the completion of the process and uses a camera to capture quality evidence images. The images are hash-encrypted and uploaded to a blockchain evidence storage node based on the Hyperledger Fabric 2.5 architecture to ensure that the acceptance data is tamper-proof.

[0070] As another embodiment of the present invention, the digital twin visualization terminal integrates the WebGL rendering engine and the CesiumJS geospatial framework to achieve seamless fusion of the BIM model and real geographic coordinates; users can click on any component in the 3D scene to view the quadruple state vector, historical deviation curve, quality inspection report and rearrangement plan Gantt chart in real time; the system also provides an API interface to support data exchange with the owner's ERP system, the supervisor's quality management system and the government supervision platform, and the exchange format follows the IFC4.3 standard and JSON-LD semantic description specification.

[0071] Furthermore, this invention adopts a zero-trust architecture for data security; all device access requires two-way TLS authentication, data transmission is encrypted with AES-256 throughout, and sensitive operations (such as model correction and plan rescheduling) require multi-factor approval (at least two authorized personnel complete the electronic signature based on the SM2 elliptic curve public key cryptography algorithm approved by the State Cryptography Administration through the supervisor's mobile terminal); system logs are synchronized to the SIEM (Security Information and Event Management) platform in real time. The SIEM platform is built on ElasticStack and loads the MITREATT&CK framework rule base for detecting suspicious behavior such as abnormal logins and batch data export.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] At the system modeling level, continuous state representation of discrete construction data is realized. Traditional management methods treat progress and quality as independent time series and discrete detection points. This invention maps each BIM component to a state point with clear physical meaning through a quadruplet state vector. It transforms the engineering management problem into a computable and observable dynamic system, enabling a complete coupled state description to be obtained at any moment in the construction process, breaking through the limitations of data fragmentation and state ambiguity in traditional methods.

[0074] At the control logic level, a bidirectional coupled nonlinear feedback mechanism is constructed; the critical chain dynamic scheduling algorithm and process dependency influence matrix designed in this invention, for the first time, incorporate schedule deviations. With quality deviation Through nonlinear functions Coupled quantification was performed; the chain effect of schedule delays compressing process time during construction, leading to quality risks and further delays due to quality rework was simulated, enabling the system to make predictive interventions in the early stages of deviation propagation, rather than responding passively afterward.

[0075] At the decision-making level, global optimal solutions under constraints are achieved; by formalizing the scheduling reordering problem into a mixed-integer linear programming model, the multi-constraint combination optimization problem of construction scheduling is transformed into a standard problem in which a mathematically optimal solution can be obtained in finite time; the quality window constraint in the model... For the first time, process waiting time has been incorporated into the optimization framework, including: The waiting period for quality stabilization is determined based on material type, environmental conditions, and construction specifications. For example, after the base layer is compacted, it is necessary to wait 24 hours before the lower layer can be laid. This ensures that the newly generated plan not only meets time and resource constraints but also guarantees the conditions for quality formation in each process.

[0076] This invention achieves a fundamental breakthrough in three dimensions: system modeling, control logic, and optimization decision-making. It upgrades the traditional experience-driven, post-event correction management model into a model-driven, predictive intervention intelligent control system, thereby improving the inherent controllability and overall robustness of the highway construction process. Attached Figure Description

[0077] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0078] Figure 1This invention provides a method and process for integrated progress and quality control of highway construction using BIM and digital twin technologies.

[0079] Figure 2 This is a data interaction flowchart of the schedule-quality joint state space model and the five-element collaborative control engine in this invention.

[0080] Figure 3 This is a schematic diagram of the hardware deployment of the device of the present invention. Detailed Implementation

[0081] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0082] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0083] Example 1: See Figures 1-3 This embodiment discloses a method and device for integrated progress and quality control of highway construction using BIM-digital twin. By constructing a joint progress-quality state space model with BIM components as basic units and deploying a five-element collaborative control architecture consisting of a progress-driven engine, a quality verification engine, a deviation analysis engine, a model correction engine, and a schedule rescheduling engine, dynamic, closed-loop, and two-way linkage control of the entire highway construction process can be achieved.

[0084] In a typical implementation scenario of this invention, the Kth section of a new highway construction project, with a total length of 12.6 kilometers, includes specialized engineering works such as roadbed, pavement, bridges and culverts, drainage, and traffic safety facilities. BIM modeling software (such as Autodesk Civil 3D or Bentley OpenRoads) is used to construct a three-dimensional model for all disciplines. All components are assigned a unique IfcGloballyUniqueId (GUID) according to the ISO12006-2 standard and exported as an IFC4.3 format file, serving as the initial geometric and semantic foundation for the digital twin platform. Simultaneously, the master schedule is prepared using Primavera P6, containing a total of 1,842 processes, covering logical relationships, resource allocation, schedule constraints, and critical path identification. Finally, it is exported in XER file format for use by the schedule-driven engine.

[0085] The central control server is deployed in the project's data center, with hardware including an Intel Xeon Gold 6348 processor (28 cores, 56 threads), 256GB DDR4 ECC memory, 2TB NVMe SSD storage, and running Ubuntu 22.04 LTS operating system. The system adopts a Docker containerized microservice architecture, encapsulating the progress-driven engine, deviation analysis engine, model correction engine, and schedule rescheduling engine into independent containers. These containers communicate efficiently via the gRPC protocol and share a unified data bus. The unified data bus is built on Apache Kafka, supporting a high-throughput, low-latency message queue mechanism to ensure the real-time performance and consistency of status updates and command issuance. The BIM model database uses a MongoDB sharded cluster to store the geometric topology, attribute sets, and version history of all components. The distributed time-series database uses InfluxDB 2.7 to persistently store the four-tuple state vector and its timestamp sequence for each component.

[0086] The schedule-driven engine parses the XER file, extracting the process network diagram, predecessor-successor relationships, resource requirements, and original planned time. The engine's internal critical chain scheduler dynamically calculates each incomplete process using the following formula. Theoretical completion time :

[0087] ;

[0088] in: Obtained by forward traversal of the network graph. For process intervals (such as the concrete needing to be cured for 72 hours before the formwork can be removed). The floating time allocated based on the buffer management strategy is dynamically adjusted according to the buffer consumption rate of the path in which the process is located; the specific allocation rules are as follows:

[0089] Processes located on the critical chain, The initial value is allocated from the project buffer proportionally to the duration of the work process, calculated using the following formula: ;in For process The planned duration, This is the sum of the durations of all processes on the critical chain. This serves as the overall buffer zone for the project.

[0090] Non-critical chain processes, Allocate 50% of the free float, with the maximum value not exceeding the time difference with the critical chain;

[0091] The scheduler automatically triggers a global recalculation every 24 hours, or immediately upon receiving an external event (such as a red rainstorm warning issued by the meteorological bureau). The recalculated plan generates a process execution instruction flow at the daily granularity, pushes it to the progress visualization module of the digital twin platform, and synchronizes it to the task to-do list on the supervisor's mobile terminal.

[0092] The on-site sensor network covers the entire construction area, specifically including: three intelligent road rollers (model: Volvo SD110D) equipped with built-in Kistler 8704B50 accelerometers and Trimble R12 GNSS receivers, with a sampling frequency of 10Hz and a positioning accuracy of ±1.5cm; two sets of FLIRA 8580sc infrared thermal imaging arrays installed on both sides of the asphalt paver, with a spatial resolution of 640×480, a temperature measurement range of 0–300℃, and an accuracy of ±1.5℃; and sensors at key sections of bridge piers and box girders. HBMKYOWAKFG-5-120-C1-11L1M2R strain gauges and SensirionSHT45 temperature and humidity probes are pre-embedded, with a strain gauge range of ±5000με and a nonlinearity error of <0.1%FS; a FAROFocusS350 mobile 3D laser scanner scans the completed roadbed section daily, with a point cloud density of 1200 points / square meter and a distance accuracy of ±0.8mm; all devices are connected to the edge computing node via industrial Ethernet (backbone) or 5G private network (mobile devices).

[0093] The edge computing nodes utilize NVIDIA Jetson AGX Orin modules, equipped with 32GB of LPDDR5 memory, running a custom Linux kernel (kernel version 5.15.83-rt47), and integrating the TensorRT 8.6 inference engine. The quality verification engine is deployed on the edge computing nodes, connecting to the sensor network via the OPCUA protocol interface to achieve device discovery, data subscription, and semantic mapping. When the progress-driven engine issues a list of procedures to be verified, the quality verification engine listens for procedure confirmation events on the supervisor's mobile terminal. For example, when a supervisor uses a ruggedized Android tablet (model: Panasonic Toughpad FZ-M1) to scan the NFC tag of a roadbed filling section (GUID: 2x$Vz9qYH0e8n$aJv$pjQaP) and clicks the confirmation button, the system records... And immediately initiate the quality verification process.

[0094] For this roadbed section (320m long, 26m wide), the quality verification engine extracted... Compaction trajectory data within the time window, including: For the process duration of this step, in this example, the roadbed filling... The roller's travel path was reconstructed based on GNSS trajectory, and the road section was divided into 1m×1m grid cells, totaling 8,320 cells; the number of roller passes was counted for each cell. Calculate its mean Standard deviation coefficient of variation The design specifications require a minimum number of compaction passes. Maximum permissible coefficient of variation The number of qualified units was determined to be 7,980, with a pass rate of [percentage missing]. It is higher than the threshold of 95%, therefore The value is assigned to 1; the result, along with metadata such as the original point cloud, temperature field image, and strain time series curve, is encrypted with AES-256 and written to the distributed time series database, triggering the deviation analysis engine.

[0095] The deviation analysis engine monitors state vector changes in real time; the aforementioned roadbed section, due to and (less than the threshold) No warning was triggered; however, in another embodiment, a certain asphalt lower layer paving section (GUID:3y$Wk0rZI1f9o#bKw$qkRbQbQ) was found to have substandard base layer compaction. This led to rework. , The engine then invokes the process dependency influence matrix. This matrix was pre-calculated and generated based on the BIM model topology and construction organization design during the project initialization phase; the subsequent upper layer paving process... Physical length Total length of the entire line And it is located on the critical path, here Take 0.4, Therefore, we take 0.6.

[0096] ;

[0097] According to the table of quality defect types, insufficient compaction of the base layer is classified as insufficient compaction defect. The measured compaction degree of 89% is lower than the standard threshold of 90%, which is considered a medium defect. The quality impact factor is obtained from the table. ;Calculation yields:

[0098] Time affects the components ;

[0099] Mass influence component ;

[0100] Coupling coefficient ;

[0101] The overall impact index is as follows:

[0102] ;

[0103] Calculated Much greater than the threshold The system immediately processed the process. Marking an event as high-risk generates an early warning event, which is then pushed to the alarm panel of the digital twin visualization terminal.

[0104] The model correction engine responds to early warning events and initiates dynamic updates to the BIM model. For geometric deviations, the system calls the latest laser scan point cloud and aligns it with the original BIM model using the ICP algorithm. For example, in a cut section, the measured slope is 0.8m more excavated than designed, exceeding the tolerance zone by ±5cm. The engine automatically identifies this area, generates a new IfcEarthworksCut entity, replaces the corresponding part of the original model, and updates the geometric boundaries. For attribute correction, such as the asphalt sub-layer mentioned above, the engine updates the QualityStatus attribute in IfcPropertySet from Pending to Failed and adds the IfcPropertySingleValue attribute.

[0105] FailureReasonCode=CompactionInsufficient;

[0106] MeasurementDataRef=https: / / tsdb.example.com / measure / 3y$Wk0rZI1f9o#bKw$qkRbQbQ;

[0107] All corrections are logged to the transaction log and a version snapshot (version number: v20240615.1432) is generated to ensure auditability.

[0108] After receiving a model correction event, the scheduling engine identifies the set of affected processes. Regarding the above cases, The process includes 12 steps: top layer paving, road marking, and installation of traffic safety facilities. The engine uses the CPLEX 22.1 solver to solve the following mixed-integer linear programming problem within 300 seconds:

[0109] ;

[0110] Among them: critical path process weight Non-critical values ​​are set to 1.0; resource penalty coefficient. The constraints include: the top layer must be laid after the base layer repair is completed and 24 hours have passed (quality window constraint). Only after these conditions are met can the paving begin; no more than two pavers can be used per day; there will be no rainfall in the next 72 hours (the probability of rainfall returned by the weather API is <10%); the solution results will postpone the paving time of the upper layer by 2 days, and by adjusting the night shifts, the total construction period will be ensured to not exceed the contract completion date; the updated XER file will be sent back to the progress-driven engine to complete the closed loop.

[0111] The digital twin visualization terminal is built on WebGL and CesiumJS. Users can click on any component in the 3D scene to bring up an information panel that displays its current state vector and historical deviation curves (such as those from the past 7 days). and The system includes trend analysis, quality inspection reports (including heatmaps, point cloud comparison charts), and rearrangement plan Gantt charts. Simultaneously, the system exchanges cost data with the owner's ERP system via external system API interfaces and synchronizes acceptance records with the supervisor's quality management system. Data formats strictly adhere to IFC4.3 and JSON-LD specifications.

[0112] Regarding data security, all device access requires two-way TLS authentication, with certificates issued by an internal CA. Sensitive operations such as model modification and plan rescheduling require at least two authorized personnel (such as the project manager and the chief supervising engineer) to complete electronic signatures (based on the national cryptographic SM2 algorithm) on the supervisor's mobile terminal. All operation logs are synchronized to the SIEM security monitoring platform in real time. The SIEM security monitoring platform is built on ElasticStack and loads the MITREATT&CK framework rule base, which can detect abnormal logins and suspicious behaviors such as batch data export.

[0113] To verify the technical effect of the present invention, a 3-kilometer test section (Example 1) in the K section of this project was selected for comparison with an adjacent 3-kilometer control section (Comparative Example 1) that did not use the present system. The geological conditions, construction teams and material supply of the two sections were the same. The example deployed the complete system of the present invention, while the comparative example adopted the traditional BIM + manual progress tracking mode.

[0114] The key metrics are compared in the table below:

[0115] Table 2:

[0116]

[0117] Among them, deviation response time refers to the average time from the occurrence of a quality problem to the issuance of a new plan; the implementation example has the ability to automatically trigger, edge computing and optimize solutions, which significantly shortens the response cycle; although the frequency of plan rearrangement has increased, all of them are preventive fine-tuning, avoiding large-scale delays.

[0118] The core inventiveness and technological breakthrough of this invention do not stem from the simple integration of BIM and digital twin technologies, but rather from the construction of a progress-quality coupled dynamic system that can be formally calculated and has closed-loop feedback capabilities, thus achieving a paradigm shift.

[0119] Traditional methods treat the construction process as a static network composed of discrete tasks, with schedule and quality belonging to two parallel data streams, and their linkage relying on human experience and post-event review. This invention, however, starts from the technical contradictions, where schedule (time resource consumption) and quality (material state formation) are constrained by the same energy input process at the physical level. It abstracts each BIM component into a dynamic entity with a four-dimensional state vector (planned time, actual time, quality standard, and measured score), so that the originally separate schedule execution and quality verification data can be observed and calculated synchronously in a unified state space.

[0120] The collaborative control architecture of this invention is a dedicated state observer and feedback controller for this coupled dynamic system. The five-element engine is not a simple list of functional modules, but constitutes a complete control loop: progress drive acts as a feedforward command generator; quality verification acts as a state observer; deviation analysis realizes the propagation prediction and root cause location of state deviations through the quantified process dependency influence matrix and comprehensive influence index; model correction is the online calibration of the system model; and plan rescheduling is an optimized feedback control based on the updated model and multiple constraints.

[0121] This invention upgrades construction management from an open-loop model relying on static planning and discrete checks to a closed-loop adaptive control model that senses status in real time, predicts the impact of deviations, dynamically corrects the model, and optimizes future instructions. The system no longer simply answers whether there is a delay or whether it is up to standard, but proactively calculates and issues warnings: to what extent will the current schedule deviation trigger what kind of quality risk, what antecedent factors are responsible for the current quality defects, and how will they cascade into subsequent processes? Therefore, this invention is not an enhancement of existing technology, but rather, by constructing a new system model and control logic, it resolves the fundamental technical contradiction between schedule and quality—where data originates from the same source but decision-making logic is fragmented—providing a calculable and implementable solution for the real-time collaborative management of complex engineering systems.

[0122] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0123] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for integrated progress and quality control of highway construction using BIM and digital twin technologies, characterized in that... Includes the following steps: Step 1: Construct a schedule-quality joint state space model using BIM components as basic units. Each BIM component is uniquely identified by the IfcGloballyUniqueId defined in the ISO 12006-2 standard and associated with a four-tuple state vector. ; in: This is the theoretical completion timestamp for the component in the schedule. This is the actual completion timestamp. The set of quality indicator thresholds specified in the design specifications. The overall quality score is calculated based on on-site measured data; Step 2: Deploy a five-element collaborative control architecture consisting of a progress-driven engine, a quality verification engine, a deviation analysis engine, a model correction engine, and a schedule rescheduling engine. Each engine achieves status synchronization and command interaction through a unified data bus. Step 3: The progress-driven engine dynamically generates the process execution instruction flow based on the critical chain method, and updates the... Pushed to the digital twin platform; Step 4: After receiving the process completion confirmation signal, the quality verification engine performs a process verification based on multi-source sensor data collected by the on-site sensor network. Perform the calculations and write the results into the joint state space model; Step 5: The deviation analysis engine is based on schedule deviation. With quality deviation Perform calculations, where For quality deviation based on The calculation yields a comprehensive impact index for subsequent processes, which is then calculated based on a preset process dependency impact matrix. A high-risk warning is generated when the comprehensive impact index exceeds a threshold. Step Six: The model correction engine performs geometric or attribute corrections on the BIM model based on the deviation type and generates a version snapshot; Step 7: The schedule rescheduling engine solves the mixed-integer linear programming problem based on the modified model, resource availability, and environmental constraints to generate an optimized schedule, which is then fed back to the schedule-driven engine to form a closed-loop control process.

2. The method for integrated progress and quality control of highway construction using BIM-digital twin technology as described in claim 1, characterized in that, The progress-driven engine dynamically calculates The formula is: ； in: Number the current process. For the collection of preceding processes, This refers to the process interval time. The earliest possible start time without considering resource constraints. Floating time allocated based on a buffer management strategy; For process The planned completion timestamp; the progress drive engine recalculates all incomplete processes every 24 hours or upon receiving an external trigger signal. .

3. The method for integrated progress and quality control of highway construction using BIM-digital twin technology as described in claim 1, characterized in that: For the roadbed compaction process, the quality judgment algorithm executed by the quality verification engine includes: Reconstructing the roller pass distribution map based on GNSS trajectory; The road section to be verified was divided into 1m×1m grid cells, and the average number of compaction passes for each cell was calculated. ;like And coefficient of variation If so, the unit is deemed qualified; Calculate the pass rate of the entire road section. ; in: The average number of compaction passes per grid cell. To meet the minimum number of compaction passes required by the design specifications, for standard deviation for The mean, The coefficient of variation is 1. The maximum permissible coefficient of variation, For the road section qualification rate, The pass rate threshold For overall quality scoring; like ,but Assign a value of 1 if the value is 1, otherwise assign a value of 0.

4. The method for integrated progress and quality control of highway construction using BIM-digital twin technology as described in claim 1, characterized in that, The field sensor network includes: an intelligent road roller with a built-in accelerometer and GNSS positioning module, an infrared thermal imaging array, an embedded strain gauge, a temperature and humidity probe, and a mobile 3D laser scanner. All sensors are connected to the edge computing node via an industrial Ethernet or 5G private network and achieve semantic interoperability through the OPCUA protocol interface.

5. The method for integrated progress and quality control of highway construction using BIM-digital twin technology as described in claim 1, characterized in that, The process dependency influence matrix Middle elements The calculation formula is: ； in: and For the weighting coefficients, satisfying The default values ​​are 0.4 and 0.6 respectively; For process The physical length; This refers to the total length of the project. As a key indicator of the process, if the process If it is located on the critical path, the value is 1; otherwise, it is 0. For process The set of all subsequent processes; For process For the process The influence coefficient; The overall impact index: ;in: ; This is a quality impact factor used to quantify the degree of influence of quality defects in preceding process i on subsequent process j. It is determined by referring to a preset rule table (as shown in Table 1) based on the type and severity of quality defects in preceding process i. This is the quality-schedule coupling coefficient, used to quantify the amplification effect of quality deviations on schedule. The default empirical value is 2.0, which can be adjusted within the range of 1.5 to 3.0 depending on the project type. For subsequent processes The comprehensive impact index; For schedule deviation, For quality deviation, The amount of time affects the weight. The quality affects the component.

6. The method for integrated progress and quality control of highway construction using BIM-digital twin technology as described in claim 1, characterized in that, The geometric correction performed by the model correction engine is as follows: it calls the point cloud data acquired by the mobile 3D laser scanner, aligns it with the original BIM model through the ICP algorithm, and automatically generates new IfcEarthworksFill or IfcEarthworksCut entities to replace the corresponding parts of the original model for areas exceeding the ±5cm tolerance zone; the attribute correction refers to: updating the QualityStatus attribute in the IfcPropertySet of the BIM component to Failed, and attaching the failure reason code and the reference link of the measured data.

7. The method for integrated progress and quality control of highway construction using BIM-digital twin technology as described in claim 1, characterized in that, The objective function of the mixed-integer linear programming problem solved by the reordering engine is: ； in: For the set of affected processes, For process The weights are 1.5 on critical paths and 1.0 on non-critical paths. For process Rearranged plan completion timestamps For process Originally planned to complete the timestamp, This is the resource penalty coefficient. For the number of resource types, For the first Excessive use of such resources This is a waiting period for quality to stabilize; The constraints include process logic constraints, resource capacity constraints, quality window constraints, and critical node conservation constraints. The quality window constraint requires that the start time of a subsequent process must not be earlier than the actual completion time of the preceding process plus the quality stabilization waiting period. , It is determined based on the type of material, environmental conditions, and construction specifications.

8. A highway construction BIM-digital twin integrated progress and quality linkage control device for implementing the method as described in any one of claims 1 to 7, characterized in that, include: The central control server is used to host the schedule-driven engine, deviation analysis engine, model correction engine, and schedule rescheduling engine. Edge computing nodes, using NVIDIA Jetson AGX Orin modules, are used to deploy the quality verification engine; The on-site sensor network includes a smart road roller, an infrared thermal imaging array, embedded strain gauges, temperature and humidity probes, and a mobile 3D laser scanner; the supervisor's mobile terminal is a ruggedized Android tablet that supports NFC tag scanning to confirm process completion and captures quality verification images to be uploaded to a blockchain evidence node based on the Hyperledger Fabric 2.5 architecture; the digital twin visualization terminal integrates a WebGL rendering engine and a CesiumJS geospatial framework for 3D scene interaction and status information display.

9. The highway construction BIM-digital twin integrated progress and quality linkage control device according to claim 8, characterized in that, The central control server connects the BIM model database and the distributed time series database through a unified data bus. The BIM model database uses MongoDB sharded clusters to store component geometry and attribute information, while the distributed time series database uses InfluxDB 2.7 to store quadruple state vectors and their time series.

10. The highway construction BIM-digital twin integrated progress and quality linkage control device according to claim 8, characterized in that, When the supervisory mobile terminal confirms the completion of the process, it needs to call the camera to capture on-site images. The images are then hash-encrypted and uploaded to a blockchain evidence storage node based on the Hyperledger Fabric 2.5 architecture to ensure that the acceptance data cannot be tampered with.

Images