Search

CN-121998407-A - Digital twinning-based real-time early warning system for safety risk of building construction site

CN121998407ACN 121998407 ACN121998407 ACN 121998407ACN-121998407-A

Abstract

The application discloses a digital twin-based real-time early warning system for safety risks of a building construction site, which comprises a data acquisition layer, a data fusion processing layer, a digital twin dynamic mapping layer, a safety risk identification and early warning layer and an early warning output and response layer, wherein a technical path for extracting risk characteristics from a digital mirror image by the digital twin dynamic mapping layer is defined, an improved D-S evidence theory and long-short-term memory network model combined data fusion model is established, and the deep interaction between a twin model and a risk assessment model is realized. The structure belongs to the technical field of building construction, and can clearly define a risk feature extraction mechanism of a digital twin dynamic mapping layer, so that real-time identification and efficient treatment of safety risks are realized.

Inventors

  • Dai Ayong

Assignees

  • 杭州拓佳科技有限公司

Dates

Publication Date
20260508
Application Date
20251219

Claims (10)

  1. 1. The digital twinning-based real-time early warning system for the safety risk of the construction site is characterized by comprising a data acquisition layer, a data fusion processing layer, a digital twinning dynamic mapping layer, a safety risk identification and early warning layer and an early warning output and response layer; The data acquisition layer is used as a data input end of the system, acquires multi-source heterogeneous data of a construction site, wherein the multi-source heterogeneous data comprises physical entity data, environment data, behavior data and business data, and outputs the acquired data to the data fusion processing layer in real time; The data fusion processing layer carries out cleaning, standardization and fusion processing on the multi-source heterogeneous data, adopts an improved D-S evidence theory to combine with a long-period memory network model to eliminate data uncertainty and realize space-time alignment, and outputs processed unified characteristic data to the digital twin dynamic mapping layer; The digital twin dynamic mapping layer builds a digital twin model based on a building information model and a Geographic Information System (GIS), wherein the digital twin model comprises three dimensions of geometry, physics and behavior, updates the model state in real time based on input characteristic data, forms a digital mirror image of a construction site and outputs risk characteristic data to the safety risk recognition and early warning layer; The safety risk recognition and early warning layer recognizes entity risk through an improved YOLOv model, evaluates risk level by combining with a fuzzy level comprehensive evaluation model, generates a grading early warning signal according to an evaluation result and outputs the grading early warning signal to the early warning output and response layer; And the early warning output and response layer outputs early warning information through a plurality of terminals and generates a treatment scheme, and meanwhile, a risk treatment result is collected and fed back to the digital twin dynamic mapping layer to finish closed loop updating.
  2. 2. The digital twinning-based real-time early warning system for safety risk of building construction sites according to claim 1, wherein the data acquisition layer acquires multi-source heterogeneous data of the construction sites, and the system comprises the following modules: The system comprises an Internet of things sensor module, a data acquisition module and a data processing module, wherein the Internet of things sensor module acquires physical entity data and environment data, the physical entity data and the environment data comprise a force sensor and a stress sensor, the force sensor acquires equipment operation parameters, and a temperature, humidity, dust and wind speed sensor acquires environment parameters and outputs numerical data according to preset frequency; The video acquisition module comprises camera equipment for acquiring a construction site image and outputting image data for identifying personnel behaviors and equipment states; the positioning module is used for acquiring the positions of personnel and mobile equipment through the ultra-wideband positioning tag and outputting position data; and the business interface module is used for interfacing the construction management system through a standardized interface and outputting corresponding business data.
  3. 3. The digital twinning-based real-time early warning system for safety risk of construction sites of buildings according to claim 1, wherein the standardized output of the data fusion layer to the data comprises the following contents: Receiving multi-source heterogeneous data of a data acquisition layer, carrying out standardization processing on the digital data, carrying out coding processing on text data, carrying out frame extraction on image data, removing abnormal data, complementing missing data, and outputting standardized data; Mapping the standardized data to a unified space-time coordinate system, processing time sequence features through a long-short-term memory network model, and outputting a time-aligned multidimensional feature sequence; And introducing an evidence weight coefficient to improve a D-S evidence theory, solving the problem of data conflict, and outputting fused data with credibility.
  4. 4. The digital twinning-based real-time early warning system for the safety risk of the construction site of the building, according to claim 1, is characterized in that the digital twinning dynamic mapping layer output risk characteristic data specifically comprises the following contents: building a building element model through a building information model, building a terrain and safety area model through a geographic information system, fusing the terrain and safety area model into a unified basic model through coordinate conversion, and outputting a twin substrate containing full element information; The output unified characteristic data of the data fusion processing layer is used for triggering the position and the parameter through an event, synchronously updating the geometric position, the physical attribute and the behavior state of the model by matching with the timing inspection, and outputting a real-time twin mirror image; The static characteristics comprise dangerous area range and equipment rated parameters, the dynamic characteristics comprise personnel positions and equipment operation parameters, and standardized risk characteristic vectors are output.
  5. 5. The digital twin-based real-time early warning system for safety risk of building construction site according to claim 4, wherein the digital twin dynamic mapping layer further comprises a light-weight rendering module for receiving the twin model updated in real time and performing light-weight processing, and outputting the interactive rendering model through a three-dimensional engine.
  6. 6. The digital twinning-based real-time early warning system for the safety risk of the construction site of the building according to claim 1, wherein the safety risk recognition and early warning layer evaluates the risk level and generates a hierarchical early warning signal according to the evaluation result, and the system comprises the following contents: Receiving risk feature vectors output by the digital twin dynamic mapping layer, optimizing YOLOv models of anchor frames through K-means clustering, identifying risk targets such as personnel violations, equipment anomalies and the like by combining an attention mechanism, and outputting entity risk results; determining index weight and fuzzy number to determine membership degree through an entity risk result by using an analytic hierarchy process algorithm, and calculating and outputting quantized risk level by using a fuzzy hierarchy comprehensive evaluation model; And receiving the quantized risk level, triggering early warning according to a preset rule, and synchronously outputting risk association information, wherein the risk association information comprises a risk position and a risk type.
  7. 7. The digital twinning-based real-time early warning system for safety risk of construction sites of building according to claim 1, wherein the early warning output and response layer generating a treatment scheme and completing closed loop updating comprises the following contents: receiving the early warning signal and synchronously outputting the early warning signal through an output device; receiving risk type and grade information, calling a preset treatment scheme library through a rule engine, and outputting a differential treatment strategy; And outputting feedback data to the digital twin dynamic mapping layer through the treatment result manually recorded or automatically monitored, updating the model state and storing the treatment record.
  8. 8. The real-time early warning system for the safety risk of the building construction site based on the digital twin system according to claim 7 is characterized in that a preset scheme library arranged in a rule engine in the early warning output and response layer comprises the corresponding relation between risk types and disposal measures, the scheme library is updated according to construction scene changes, the scheme library is transmitted to the digital twin model before being output according to the disposal strategy, feasibility is verified through simulation, if the feasibility is achieved, the scheme library is output, and if the feasibility is not achieved, the scheme library is terminated.
  9. 9. The digital twinning-based real-time early warning system for the safety risk of the construction site of the building according to claim 1, further comprising an edge and cloud cooperation module, wherein the edge and cloud cooperation module is used as a supporting layer for data transmission and storage, and layered processing is realized: The system comprises an edge gateway, a data acquisition layer, a cloud end and a cloud end, wherein the edge gateway is deployed on a construction site, receives multi-source heterogeneous data output by the data acquisition layer, performs local simplifying processing, outputs simplified data to the cloud end, and reduces transmission pressure; the cloud node adopts an elastic server to receive the simplified data and the running logs of each layer output by the edge gateway, stores the data such as a twin model, a treatment record and the like, and outputs storage service to support interaction of each layer.
  10. 10. A computer storage medium, characterized in that it has stored thereon a program which, when executed by a processor, implements a digital twinning based real-time early warning system for security risk of construction sites according to claims 1-9.

Description

Digital twinning-based real-time early warning system for safety risk of building construction site Technical Field The invention relates to the technical field of building construction, in particular to a digital twinning-based real-time early warning system for safety risk of a building construction site. Background The construction site has complex environment, frequent personnel flow and various equipment types, the safety risk presents burstiness, coupling and dynamic characteristics, the traditional safety management relies on manual inspection and fixed threshold monitoring, and the problems of risk identification lag, high false alarm rate, data fragmentation and the like exist. The existing digital twin-based security management system realizes digital mapping of a construction site, but has key technical defects: 1. The multi-source heterogeneous data has weak fusion capability, can not effectively integrate data such as sensors, videos, positioning and the like, and has insufficient data value mining; 2. the digital twin model is updated and the risk characteristics are extracted and disjointed, the model is only used as a visual carrier, and a technical path for acquiring the risk characteristics from the digital mirror image is not clear; 3. the risk assessment model and the twin model are not interacted enough, and a linkage mechanism of data, the model and assessment is not formed; 4. The early warning response lacks dynamic adaptation to the twinning model, and the treatment scheme feasibility cannot be verified by the model. Therefore, development of a digital twin-based real-time early warning system for safety risk of a building construction site is needed to solve the problems in the prior art. Disclosure of Invention The invention aims to provide a digital twinning-based real-time early warning system for the safety risk of a building construction site, which can clearly determine a risk feature extraction mechanism of a digital twinning dynamic mapping layer, realize real-time identification and efficient disposal of the safety risk, has a simple structure and is convenient to use, so as to solve the problems in the background technology. In order to achieve the above purpose, the present invention provides the following technical solutions: the digital twinning-based real-time early warning system for the safety risk of the construction site comprises a data acquisition layer, a data fusion processing layer, a digital twinning dynamic mapping layer, a safety risk identification and early warning layer and an early warning output and response layer; The data acquisition layer is used as a data input end of the system, acquires multi-source heterogeneous data of a construction site, wherein the multi-source heterogeneous data comprises physical entity data, environment data, behavior data and business data, and outputs the acquired data to the data fusion processing layer in real time; The data fusion processing layer carries out cleaning, standardization and fusion processing on the multi-source heterogeneous data, adopts an improved D-S evidence theory to combine with a long-period memory network model to eliminate data uncertainty and realize space-time alignment, and outputs processed unified characteristic data to the digital twin dynamic mapping layer; The digital twin dynamic mapping layer builds a digital twin model based on a building information model and a Geographic Information System (GIS), wherein the digital twin model comprises three dimensions of geometry, physics and behavior, updates the model state in real time based on input characteristic data, forms a digital mirror image of a construction site and outputs risk characteristic data to the safety risk recognition and early warning layer; the safety risk identification and early warning layer identifies entity risk through an improved YOLOv model, evaluates risk level by combining a analytic hierarchy process-fuzzy comprehensive evaluation model, generates a grading early warning signal according to an evaluation result and outputs the grading early warning signal to the early warning output and response layer; And the early warning output and response layer outputs early warning information through a plurality of terminals and generates a treatment scheme, and meanwhile, a risk treatment result is collected and fed back to the digital twin dynamic mapping layer to finish closed loop updating. Preferably, the data acquisition layer acquires multi-source heterogeneous data of a construction site, wherein the multi-source heterogeneous data comprises the following modules: The system comprises an Internet of things sensor module, a data acquisition module and a data processing module, wherein the Internet of things sensor module acquires physical entity data and environment data, the physical entity data and the environment data comprise a force sensor and a stress sensor, the force sensor acquires equipment operation