Bridge construction monitoring method and system based on BIM

Abstract

The invention discloses a BIM-based bridge construction monitoring method and system, and belongs to the technical field of bridge construction. Construction time deviation characteristics and structure response deviation characteristics are extracted, a drift index is calculated to evaluate the binding error degree between data and components, and a differential processing strategy is implemented based on the drift level; and for high-risk data, a similarity scoring mechanism is introduced, an intelligent matching algorithm is adopted to identify an optimal component and complete rebinding, and finally accurate storage of the data and dynamic updating of a BIM model state are realized, so that the binding accuracy of the construction monitoring data, the risk identification capability and the model linkage efficiency are effectively improved, and the construction quality is improved. And the intelligent management and decision support level of bridge construction is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge construction, and in particular to a bridge construction monitoring method and system based on BIM. Background Art

[0002] BIM-based bridge construction monitoring utilizes Building Information Modeling (BIM) technology to provide digital, visual, and intelligent real-time monitoring and management of various construction activities during the bridge construction process. By integrating design drawings, construction progress, equipment status, and quality inspection data into BIM models, dynamic tracking, problem warnings, and collaborative decision-making throughout the construction process are achieved, thereby improving construction efficiency and ensuring project quality and safety.

[0003] The existing technology has the following shortcomings:

[0004] In BIM-based bridge construction monitoring, when real-time data is bound to model components using metadata such as geolocation, timestamps, and component IDs, data drift can lead to component binding errors. Because on-site GPS signals are affected by occlusion and reflection, sensor location information can shift, causing data to be mistakenly bound to adjacent or incorrect bridge components. This can lead to serious consequences such as misjudgment of risk warnings, overlooking real risks, and distorting historical data, compromising project safety and decision-making accuracy. Summary of the Invention

[0005] The purpose of the present invention is to provide a bridge construction monitoring method and system based on BIM to address the shortcomings of the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a BIM-based bridge construction monitoring method, comprising:

[0007] Acquire real-time sensor data from the bridge construction site, including timestamps, component IDs, geographic location coordinates, and physical response data;

[0008] Extract data drift related features from sensor data, including construction time deviation features and structural response deviation features;

[0009] Calculating a drift index based on the construction time deviation characteristics and the structural response deviation characteristics to evaluate the degree of binding error between sensor data and target components;

[0010] Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk. Differentiated processing is performed on data of different drift levels:

[0011] If the drift level is high risk, a similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component;

[0012] The sensor data after binding confirmation is written into the database, and the status information of the corresponding components in the BIM model is updated.

[0013] Preferably, obtaining real-time sensor data from the bridge construction site includes: deploying multiple types of sensors at the components of the bridge structure, and uniquely numbering each sensor or establishing an initial binding relationship with the component ID; the sensors include stress gauges, strain gauges, displacement gauges, accelerometers, temperature sensors, humidity sensors and GPS positioning modules.

[0014] Preferably, the construction time deviation feature is used to determine the degree of time deviation between the timestamp and the construction time plan of the corresponding component in the BIM model, and the extraction method is:

[0015] Construct the construction process dependency graph G(V,E), where node V represents a bridge component or construction process; directed edge E represents each edge (v i →v j ) represents the component v i It is v j Pre-construction conditions; for the target component C bound to the sensor data t , extract its: all direct predecessor nodes P(C t ): Must be in C t Completed before construction; all direct successor nodes S(C t ): In C t Construction can only be carried out after that; the sensor records the timestamp T data ; Obtain from BIM construction plan or historical records: Predecessor node completion time set: Successor node start time set: Define the construction time deviation value ΔT dep , which indicates the degree of position deviation of the target component in the dependency path, and is expressed as: Where, is the actual completion time of the i-th predecessor component, The planned or actual start time of the jth successor component.

[0016] Preferably, the structural response deviation feature is used to determine the degree of deviation between the physical response data of the sensor and the component simulation value or the historical data of the adjacent component, and the extraction method is:

[0017] Assume that the sensor is in a certain time window t∈[t1,t n] The response data collected in is: X={x1,x2,...,x n}; where x n represents the stress or strain value, n is the total number of response data; at the same time, the simulation data distribution of the target component is obtained as follows: Y = {y1, y2, ..., y m}; m is the total number of simulation data; sort X and Y respectively and normalize them to form the empirical distribution function F X 、F Y , calculate its structural response deviation value W(X,Y), the expression is: z is an integration variable representing any point on the response range.

[0018] Preferably, a drift index is calculated based on the construction time deviation characteristics and the structural response deviation characteristics to evaluate the degree of binding error between the sensor data and the target component, specifically including:

[0019] The construction time deviation value and the structural response deviation value are normalized so that they are both between [0, 1]. The drift index is calculated based on the normalized construction time deviation value and structural response deviation value.

[0020] Preferably, based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk, and differentiated processing is performed on data of different drift levels, specifically including:

[0021] Comparing the obtained drift index with a gradient standard threshold, where the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and comparing the drift index with the first standard threshold and the second standard threshold respectively;

[0022] If the drift index is greater than a second standard threshold, the sensor data is determined to be at a high-risk drift level, triggering a binding correction process;

[0023] If the drift index is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the sensor data is determined to be at a medium-risk drift level, and the data is marked as requiring confirmation, and is manually reviewed or semi-automatically verified in combination with auxiliary information;

[0024] If the drift index is less than the first standard threshold, the sensor data is judged to be at a low-risk drift level, and its binding relationship is considered credible. The binding is automatically completed and included in the BIM model update process.

[0025] Preferably, if the drift level is high risk, a similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component, specifically including:

[0026] For sensor data with a high risk of drift, construct a data feature vector that fuses multiple modalities For all candidate components C in the BIM model i , respectively extract the existing construction time deviation characteristics and structural response deviation characteristics, and map them into component vectors Forming a standard candidate library d is the total number of components;

[0027] Use the multimodal embedding model S to map the input vector to a unified embedding space: Where, is the multimodal embedding vector of the current sensor data, Candidate component C i The standard multimodal embedding vector of ;

[0028] For each candidate component C i , calculate its vector similarity score S with the current sensor data i : Where, is a vector The Euclidean norm of , is a vector The Euclidean norm of the bound component is obtained; the score set of all components is obtained, and the component with the highest score is selected. If the score is greater than the confidence threshold, the original bound component is replaced to complete the correction; otherwise, the sensor data is marked as low-confidence binding and enters the manual review process.

[0029] The present invention also provides a BIM-based bridge construction monitoring system, comprising a data acquisition module, a drift feature extraction module, an evaluation module, a hierarchical processing module, a component rebinding module, and a BIM model updating module;

[0030] Data acquisition module: acquires real-time sensor data from the bridge construction site, including timestamps, component IDs, geographic coordinates, and physical response data;

[0031] Drift feature extraction module: extracts data drift related features from sensor data, including construction time deviation features and structural response deviation features;

[0032] Evaluation module: Calculates a drift index based on the construction time deviation characteristics and the structural response deviation characteristics, and is used to evaluate the degree of binding error between the sensor data and the target component;

[0033] Grading processing module: Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk, and data of different drift levels are processed differently:

[0034] Component rebinding module: If the drift level is high risk, the similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component;

[0035] BIM model update module: writes the sensor data after binding confirmation into the database and updates the status information of the corresponding components in the BIM model.

[0036] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0037] 1. This invention incorporates construction time deviation characteristics and structural response deviation characteristics to establish a drift index evaluation mechanism, enabling quantitative assessment of sensor data and BIM component binding errors. This overcomes the limitations of existing binding methods that rely on component IDs and geographic locations, effectively addressing technical challenges such as component binding errors, misjudgment of risk warnings, and distorted monitoring data caused by GPS drift. By categorizing and differentially processing sensor data by risk level, this approach not only improves data binding accuracy but also enhances intelligent identification and dynamic management capabilities during bridge construction.

[0038] 2. This invention introduces a multimodal embedding matching algorithm to perform component re-identification on high-risk data. By integrating structural response, temporal, and spatial information, it establishes a unified embedding space, achieving high-precision component similarity scoring and binding correction, ensuring the consistency and reliability of monitoring data and BIM models. Combining database storage with a synchronous BIM model status update mechanism, a data-driven closed-loop monitoring system is constructed, significantly enhancing the intelligence, visualization, and decision-making support capabilities of bridge construction monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0040] Figure 1 This is a mind map of the method of the present invention.

[0041] Figure 2 This is a mind map of the system modules of the present invention. DETAILED DESCRIPTION

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0043] Example 1, please refer to Figure 1 As shown, the BIM-based bridge construction monitoring method described in this embodiment includes:

[0044] Acquire real-time sensor data from the bridge construction site, including timestamps, component IDs, geographic location coordinates, and physical response data;

[0045] Extract data drift related features from sensor data, including construction time deviation features and structural response deviation features;

[0046] Calculating a drift index based on the construction time deviation characteristics and the structural response deviation characteristics to evaluate the degree of binding error between sensor data and target components;

[0047] Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk. Differentiated processing is performed on data of different drift levels:

[0048] If the drift level is high risk, a similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component;

[0049] The sensor data after binding confirmation is written into the database, and the status information of the corresponding components in the BIM model is updated.

[0050] Access real-time sensor data from the bridge construction site, including:

[0051] Sensor layout and numbering: Various types of sensors are pre-installed at key components of the bridge structure (such as piers, main beams, suspension cable anchor sections, etc.), including but not limited to stress gauges, strain gauges, displacement gauges, accelerometers, temperature sensors, humidity sensors and GPS positioning modules. Each sensor is uniquely numbered or initially bound to the component ID.

[0052] Data field collection definition:

[0053] Each sensor collects data in real time and generates a data packet that includes the following:

[0054] Timestamp: indicates the acquisition time of the data record, accurate to seconds or milliseconds;

[0055] Component ID: indicates the component number that the sensor is initially bound or identified to (if pre-bound);

[0056] Geographic coordinates: spatial coordinates provided by GPS or high-precision RTK positioning modules, in the format of longitude, latitude and optional elevation data;

[0057] Physical response data: Depending on the sensor type, it may include indicators such as structural stress value, strain value, acceleration value, displacement, temperature, humidity, and voltage change;

[0058] Data uplink and identification encapsulation: The data is uploaded to the data receiving server or BIM monitoring platform via wired or wireless communication (such as LoRa, NB-IoT, 4G / 5G network); each data packet is accompanied by a unique data identifier (dataUID) for subsequent tracking and error correction.

[0059] Data preprocessing and caching mechanism: Before entering the main processing flow, the uploaded data is first placed in the cache processing area for preprocessing operations such as format verification, missing value filling, and outlier removal to ensure that the data quality meets the standards and meets the subsequent component binding calculation requirements.

[0060] Multi-source synchronization and time alignment mechanism: For data packets from different sensor sources, the system adopts a unified time synchronization mechanism (such as based on the NTP protocol or local master clock alignment) to ensure a unified time base when cross-analyzing multiple sensors.

[0061] The construction time deviation feature is used to determine the degree of time deviation between the timestamp and the construction time plan of the corresponding component in the BIM model. The extraction method is:

[0062] The bridge construction process is abstracted into a directed acyclic graph (DAG), where each node represents a construction component or process, and each edge represents the dependency relationship between previous and subsequent construction steps. By analyzing the timestamps of sensor data and the completion times of other components in the dependency path, the degree of time deviation within the overall construction logic is assessed, thereby determining whether the data binding is appropriate.

[0063] Construct a construction process dependency graph G(V,E), where each node V represents a bridge component or construction process, such as a pier, beam segment, or cable tower.

[0064] Directed edge E represents each edge (v i →v j ) represents the component v i It is v j Pre-construction conditions;

[0065] Target component C for sensor data binding t , extract it:

[0066] All direct predecessor nodes P(C t ): Must be in C t Completed before construction;

[0067] All direct successor nodes S(C t ): In C t Construction can only begin after that.

[0068] Sensor recording timestamp T data ;

[0069] Obtained from BIM construction plan or history: Predecessor node completion time set: Successor node start time set:

[0070] Define the construction time deviation value ΔT dep , which indicates the degree of position deviation of the target component in the dependency path, and is expressed as: Where, is the actual completion time of the i-th predecessor component, The planned or actual start time of the jth successor component.

[0071] A larger construction time deviation indicates that the sensor data timestamp deviates further from the target component's planned construction time window in the BIM model. This generally indicates a higher degree of misbinding error between the data and the target component. In other words, the sensor data likely originates from another component or construction phase, significantly increasing the risk of misbinding.

[0072] Conversely, if the construction time deviation is close to zero, it indicates that the timestamp of the sensor data is highly consistent with the planned construction time of the target component. In this case, the binding sequence is more reasonable, the probability of misbinding is lower, and the credibility is higher. This feature can be used as one of the key criteria for data drift detection.

[0073] The structural response deviation feature is used to determine the degree of deviation between the sensor's physical response data and the component simulation value or the historical data of adjacent components. The extraction method is:

[0074] Assume that the sensor is in a certain time window t∈[t1,t n ] The response data collected in is: X={x1,x2,...,x n}; where x n represents the stress or strain value, and n is the total number of response data;

[0075] At the same time, the simulation data distribution of the target component is obtained as follows: Y = {y1, y2, ..., y m}; m is the total number of simulation data;

[0076] Sort X and Y separately and normalize them to form the empirical distribution function F X 、F Y , calculate its structural response deviation value W(X,Y), the expression is: z is an integration variable representing any point on the response range.

[0077] The larger the structural response deviation value, the more significant the difference between the physical response data currently collected by the sensor and the simulation prediction value of the target component or the historical response value of the adjacent component, indicating that its structural behavior does not meet expectations. There may be problems such as sensor misbinding, data drift, or abnormality of the component itself, and the degree of binding error is high.

[0078] On the contrary, the smaller the structural response deviation value, the more consistent the sensor data is with the response characteristics of the target component, which conforms to the structural force logic and historical laws, indicating that the data is more likely to be correctly bound to the target component, with higher credibility and lower risk of misbinding.

[0079] Based on the construction time deviation characteristics and the structural response deviation characteristics, a drift index is calculated to evaluate the degree of binding error between the sensor data and the target component, specifically including:

[0080] The construction time deviation value and the structural response deviation value are normalized so that they are both between [0, 1]. The drift index is calculated based on the normalized construction time deviation value and structural response deviation value.

[0081] For example, the present invention can use the following formula to calculate the drift index, and the calculation expression is: Where F is the drift index, ΔT dep is the construction time deviation value, W(X,Y) is the structural response deviation value, a1 and a2 are the weight coefficients of the construction time deviation value and the structural response deviation value (which can be optimized based on experimental experience or machine learning), and a1 and a2 are both greater than 0.

[0082] Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk. Differentiated processing is performed on data of different drift levels, specifically including:

[0083] Comparing the obtained drift index with a gradient standard threshold, where the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and comparing the drift index with the first standard threshold and the second standard threshold respectively;

[0084] If the drift index is greater than the second standard threshold, the sensor data is judged to be at a high-risk drift level, triggering a binding correction process, which specifically includes recalculating component matching similarity, reference structure response trends, construction sequence and other characteristics, re-identifying the most likely binding components, and replacing the original binding relationship;

[0085] If the drift index is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the sensor data is determined to be at a medium-risk drift level, and the system marks the data as requiring confirmation. The accuracy of the bound components is confirmed through manual review or semi-automatic verification combined with auxiliary information (such as video records, personnel operation logs, and construction location information).

[0086] If the drift index is less than the first standard threshold, the sensor data is judged to be at a low-risk drift level, and its binding relationship is considered credible. The binding is automatically completed and incorporated into the BIM model update process for subsequent visual monitoring and risk assessment.

[0087] In the present invention, by grading the drift index, quantitative evaluation and differentiated response of data binding quality are achieved, which effectively improves the intelligence level and data credibility of the BIM model in bridge construction monitoring.

[0088] If the drift level is high risk, the similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component, specifically including:

[0089] For sensor data with a high risk of drift, construct a data feature vector that fuses multiple modalities Including but not limited to:

[0090] Mode 1: structural response mode;

[0091] Mode 2: Time mode;

[0092] Mode 3: spatial mode (sensor geographic location, original binding component coordinates);

[0093] Mode 4: Drift Index.

[0094] The above data are spliced into a unified high-dimensional feature vector, which represents the multimodal description of the current data to be bound.

[0095] For all candidate components C in the BIM model i , respectively extract the existing construction time deviation characteristics and structural response deviation characteristics, and map them into component vectors Forming a standard candidate library d is the total number of components;

[0096] Use a pre-trained or online-trained multimodal embedding model S, such as one based on a Transformer or MLP structure, to map the input vector to a unified embedding space: Where, is the multimodal embedding vector of the current sensor data, Candidate component C i The standard multimodal embedding vector of ;

[0097] For each candidate component C i , calculate its vector similarity score S with the current sensor data i (Cosine similarity can be used): Where, is a vector The Euclidean norm of , is a vector The Euclidean norm of the bound component is obtained; the score set of all components is obtained, and the component with the highest score is selected. If the score is greater than the confidence threshold, the original bound component is replaced to complete the correction; otherwise, the sensor data is marked as low-confidence binding and enters the manual review process.

[0098] Once the sensor data is confirmed to be bound to the component through steps such as drift assessment and similarity matching, the system performs data storage and synchronizes the BIM model status update operations, including:

[0099] The sensor data after binding confirmation is encapsulated into a structured data record, including but not limited to the following fields:

[0100] Sensor unique number (SensorID);

[0101] Data collection timestamp (Timestamp);

[0102] Confirm the bound component ID (ComponentID);

[0103] Measurement values (such as stress, strain, displacement, etc.);

[0104] Drift index and binding similarity score (for later tracking);

[0105] Binding confirmation status identification (such as: automatic binding / correction binding / manual review).

[0106] This data structure serves as the standard unit of persistent storage.

[0107] The system writes structured data records into the central construction monitoring database;

[0108] Data is archived in time series format, supporting indexing and querying by component, sensor, or time range;

[0109] At the same time, the logging and binding history tracking mechanisms are triggered for later data auditing and traceability analysis.

[0110] Based on the confirmed bound component ID, the system locates the corresponding component entity in the BIM model and updates the following information:

[0111] Component status property update:

[0112] Update the latest sensor response value;

[0113] Mark the health status and data update time of the current component;

[0114] If sensor data triggers a warning threshold, the status can also be marked as "warning" or "abnormal".

[0115] In the 3D BIM visualization interface, the component color is dynamically adjusted according to the status of the bound data, such as green for normal, yellow for attention, and red for abnormal; the real-time monitoring interface is automatically refreshed.

[0116] If the bound data meets the warning conditions, it will be automatically transferred to the risk assessment module; if the data is continuously abnormal, it can trigger component structure performance degradation analysis or operation and maintenance intervention suggestions; all data storage actions record a unique data ID to facilitate subsequent batch processing, model training or fault tracking.

[0117] This embodiment provides a bridge construction monitoring method based on BIM. By collecting sensor data from the construction site, construction time deviation characteristics and structural response deviation characteristics are extracted, and a drift index is calculated to evaluate the degree of binding error between the data and the target component. Based on the drift index, the data is divided into three risk levels: high, medium, and low. Intelligent algorithms such as multimodal embedding matching are used to re-identify and correct the binding of high-risk data. Finally, the confirmed bound data is written into the database, and the status information of the corresponding components in the BIM model is updated in real time, thereby realizing the precise binding of construction monitoring data and dynamic linkage of the model, and improving the level of intelligent perception and safety management and control during the bridge construction process.

[0118] Example 2, please refer to Figure 2 As shown, the BIM-based bridge construction monitoring system described in this embodiment includes a data acquisition module, a drift feature extraction module, an evaluation module, a hierarchical processing module, a component rebinding module, and a BIM model update module;

[0119] Data acquisition module: acquires real-time sensor data from the bridge construction site, including timestamps, component IDs, geographic coordinates, and physical response data;

[0120] Drift feature extraction module: extracts data drift related features from sensor data, including construction time deviation features and structural response deviation features;

[0121] Evaluation module: Calculates a drift index based on the construction time deviation characteristics and the structural response deviation characteristics, and is used to evaluate the degree of binding error between the sensor data and the target component;

[0122] Grading processing module: Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk, and data of different drift levels are processed differently:

[0123] Component rebinding module: If the drift level is high risk, the similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component;

[0124] BIM model update module: writes the sensor data after binding confirmation into the database and updates the status information of the corresponding components in the BIM model.

[0125] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0126] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0127] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0128] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A BIM-based bridge construction monitoring method, characterized by: include: Acquire real-time sensor data from the bridge construction site, including timestamps, component IDs, geographic location coordinates, and physical response data; Extract data drift related features from sensor data, including construction time deviation features and structural response deviation features; Calculating a drift index based on the construction time deviation characteristics and the structural response deviation characteristics to evaluate the degree of binding error between sensor data and target components; Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk. Differentiated processing is performed on data of different drift levels: If the drift level is high risk, a similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component; The sensor data after binding confirmation is written into the database, and the status information of the corresponding components in the BIM model is updated.

2. The BIM-based bridge construction monitoring method according to claim 1, characterized in that: Acquiring real-time sensor data from the bridge construction site includes: deploying multiple types of sensors at the components of the bridge structure and uniquely numbering each sensor or establishing an initial binding relationship with the component ID; the sensors include stress gauges, strain gauges, displacement gauges, accelerometers, temperature sensors, humidity sensors and GPS positioning modules.

3. The BIM-based bridge construction monitoring method according to claim 1, characterized in that: The construction time deviation feature is used to determine the degree of time deviation between the timestamp and the construction time plan of the corresponding component in the BIM model. The extraction method is: Construct the construction process dependency graph G(V,E), where node V represents a bridge component or construction process; directed edge E represents each edge (v i →v j ) represents the component v i It is v j Pre-construction conditions; for the target component C bound to the sensor data t , extract its: all direct predecessor nodes P(C t ): Must be in C t Completed before construction; all direct successor nodes S(C t ): In C t Construction can only be carried out after that; the sensor records the timestamp T data ; Obtain from BIM construction plan or historical records: Predecessor node completion time set: Successor node start time set: Define the construction time deviation value ΔT dep , which indicates the degree of position deviation of the target component in the dependency path, and is expressed as: Where, is the actual completion time of the i-th predecessor component, The planned or actual start time of the jth successor component.

4. The BIM-based bridge construction monitoring method according to claim 3, characterized in that: The structural response deviation feature is used to determine the degree of deviation between the sensor's physical response data and the component simulation value or the historical data of adjacent components. The extraction method is: Assume that the sensor is in a certain time window t∈[t1,t n ] The response data collected in is: X={x1,x2,...,x n }; where x n represents the stress or strain value, n is the total number of response data; at the same time, the simulation data distribution of the target component is obtained as follows: Y = {y1, y2, ..., y m }; m is the total number of simulation data; sort X and Y respectively and normalize them to form the empirical distribution function F X 、F Y , calculate its structural response deviation value W(X,Y), the expression is: z is an integration variable representing any point on the response domain.

5. The BIM-based bridge construction monitoring method according to claim 4, characterized in that: Based on the construction time deviation characteristics and the structural response deviation characteristics, a drift index is calculated to evaluate the degree of binding error between the sensor data and the target component, specifically including: The construction time deviation value and the structural response deviation value are normalized so that they are both between [0, 1]. The drift index is calculated based on the normalized construction time deviation value and structural response deviation value.

6. The BIM-based bridge construction monitoring method according to claim 5, characterized in that: Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk. Differentiated processing is performed on data of different drift levels, specifically including: Comparing the obtained drift index with a gradient standard threshold, where the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and comparing the drift index with the first standard threshold and the second standard threshold respectively; If the drift index is greater than a second standard threshold, the sensor data is determined to be at a high-risk drift level, triggering a binding correction process; If the drift index is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the sensor data is determined to be at a medium-risk drift level, and the data is marked as requiring confirmation, and is manually reviewed or semi-automatically verified in combination with auxiliary information; If the drift index is less than the first standard threshold, the sensor data is judged to be at a low-risk drift level, and its binding relationship is considered credible. The binding is automatically completed and included in the BIM model update process.

7. The BIM-based bridge construction monitoring method according to claim 6, characterized in that: If the drift level is high risk, the similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component, specifically including: For sensor data with a high risk of drift, construct a data feature vector that fuses multiple modalities For all candidate components C in the BIM model i , respectively extract the existing construction time deviation characteristics and structural response deviation characteristics, and map them into component vectors Forming a standard candidate library d is the total number of components; Use the multimodal embedding model S to map the input vector to a unified embedding space: Where, is the multimodal embedding vector of the current sensor data, Candidate component C i The standard multimodal embedding vector of ; For each candidate component C i , calculate its vector similarity score S with the current sensor data i : Where, is a vector The Euclidean norm of , is a vector The Euclidean norm of the bound component is obtained; the score set of all components is obtained, and the component with the highest score is selected. If the score is greater than the confidence threshold, the original bound component is replaced to complete the correction; otherwise, the sensor data is marked as low-confidence binding and enters the manual review process.

8. A BIM-based bridge construction monitoring system, for implementing the BIM-based bridge construction monitoring method according to any one of claims 1 to 7, characterized in that: It includes data acquisition module, drift feature extraction module, evaluation module, hierarchical processing module, component rebinding module and BIM model update module; Data acquisition module: acquires real-time sensor data from the bridge construction site, including timestamps, component IDs, geographic coordinates, and physical response data; Drift feature extraction module: extracts data drift related features from sensor data, including construction time deviation features and structural response deviation features; Evaluation module: Calculates a drift index based on the construction time deviation characteristics and the structural response deviation characteristics, and is used to evaluate the degree of binding error between the sensor data and the target component; Grading processing module: Based on the drift index, the sensor data is divided into three drift levels: high risk, medium risk, and low risk, and data of different drift levels are processed differently: Component rebinding module: If the drift level is high risk, the similarity score of the target component is calculated based on the sensor data and the drift index, and the target component with the highest similarity score is selected to replace the original bound component; BIM model update module: writes the sensor data after binding confirmation into the database and updates the status information of the corresponding components in the BIM model.

Images