An intelligent monitoring system and method for bridge steel cofferdam construction based on digital twin

Abstract

The present invention relates to the field of digital twin technology, and in particular to a method and system for intelligent monitoring of bridge steel cofferdam construction based on digital twin. The method comprises the following steps: obtaining bridge steel cofferdam construction data; creating a full-element digital twin model according to the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model; obtaining real-time bridge steel cofferdam data, and performing fusion mapping according to the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model to obtain a bridge steel cofferdam dynamic model; performing real-time monitoring according to the bridge steel cofferdam dynamic model to obtain bridge steel cofferdam construction monitoring data, and performing dynamic risk assessment according to the bridge steel cofferdam construction monitoring data to obtain bridge steel cofferdam construction risk data, so as to perform intelligent monitoring operations for bridge steel cofferdam construction. The present invention continuously updates the digital twin scene through massive scenes and monitoring data, and realizes visualization, intelligent monitoring and early warning of the entire process of bridge steel cofferdam construction.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and in particular to a method and system for intelligent monitoring of bridge steel cofferdam construction based on digital twins. Background Art

[0002] The traditional bridge steel cofferdam construction monitoring and early warning mainly rely on total stations, levels, theodolites and other equipment to monitor its plane position, elevation, inclination, settlement and other data, and monitor according to the pre-set monitoring cycle and construction process. The structural safety data (stress, deformation, large volume concrete temperature, deflection, etc.) are monitored using strain gauges, displacement gauges, etc., and paper monitoring reports are compiled and reported to the construction, supervision and construction units. The traditional monitoring and early warning methods are difficult to collect data, and the data cannot be associated with the structural monitoring position in real time, which not only wastes a lot of manpower and material resources, increases the safety risks of monitoring personnel, but also cannot meet the accuracy and timeliness of steel cofferdam monitoring. It is impossible to quickly and real-timely grasp the stress and stability of the steel cofferdam structure, and it is impossible to achieve the purpose of construction safety risk early warning, which is contrary to the concept of pre-warning of construction safety risk and rapid disposal of emergency plans. It greatly increases the construction safety risk and the lag of safety early warning. It causes construction safety troubles for construction units and construction units. Summary of the invention

[0003] In order to solve the above-mentioned technical problems, the present invention proposes an intelligent monitoring method and system for bridge steel cofferdam construction based on digital twin to solve at least one of the above-mentioned technical problems.

[0004] The present application provides a method for intelligent monitoring of bridge steel cofferdam construction based on digital twins, the method comprising:

[0005] S1. Obtain the construction data of bridge steel cofferdam;

[0006] S2. Create a full-factor digital twin model based on the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model;

[0007] S3, obtaining real-time bridge steel cofferdam data, and performing fusion mapping based on the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model to obtain a bridge steel cofferdam dynamic model;

[0008] S4. Perform real-time monitoring based on the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data, and perform dynamic risk assessment based on the bridge steel cofferdam construction monitoring data to obtain the bridge steel cofferdam construction risk data to carry out intelligent monitoring operations for the bridge steel cofferdam construction.

[0009] In the present invention, the data of various sensors at the construction site are integrated to provide comprehensive and accurate construction status information. Using digital twin technology, a detailed virtual model is created to enhance the simulation and analysis capabilities of the construction process. Through real-time data mapping, the synchronization of the physical world and the virtual world is achieved, and dynamic and real-time construction status information is provided. Real-time monitoring data is used for intelligent risk assessment to timely discover and deal with potential risks in construction. Taking into full consideration the timeliness of bridge steel cofferdam construction monitoring, dynamic safety risk intelligent assessment, intelligent safety warning, and rapid disposal of emergency plans, through real-time acquisition, data cleaning, analysis and processing, and mapping of bridge steel cofferdam construction data based on digital twin + IOT equipment, stress, strain, displacement, settlement, and temperature data are linked to the structural position in real time, and compared and analyzed in real time according to monitoring standards, safety warnings are given to steel cofferdam structures throughout the construction process, and continuous learning, optimization and upgrading of IoT monitoring equipment are used to intelligently judge the safety status of the structure, and the system and algorithm are continuously upgraded and iterated, ultimately solving the problems of untimely and unintuitive monitoring of bridge steel cofferdam construction, high monitoring personnel costs, and high safety risks. By integrating emergency plan measures with intelligent monitoring, the system automatically starts the emergency plan when it determines that the emergency plan activation standards have been met, pushes emergency measures to relevant responsible personnel as soon as possible, and automatically starts corresponding disposal equipment to complete emergency disposal. Real-time visualization, intelligent, and convenient monitoring, early warning, and emergency disposal of steel cofferdams are achieved, reducing construction safety risks and improving construction efficiency.

[0010] Optionally, the present application further provides a bridge steel cofferdam construction intelligent monitoring system based on digital twin, which is used to execute the bridge steel cofferdam construction intelligent monitoring method based on digital twin as described above, and the bridge steel cofferdam construction intelligent monitoring system based on digital twin includes:

[0011] Bridge steel cofferdam construction data acquisition module, used to obtain bridge steel cofferdam construction data;

[0012] A full-factor digital twin model creation module is used to create a full-factor digital twin model based on the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model;

[0013] A real-time bridge steel cofferdam data fusion mapping model is used to obtain real-time bridge steel cofferdam data, and to perform fusion mapping based on the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model to obtain a bridge steel cofferdam dynamic model;

[0014] The intelligent monitoring module for bridge steel cofferdam construction is used to perform real-time monitoring based on the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data, and to perform dynamic risk assessment based on the bridge steel cofferdam construction monitoring data to obtain the bridge steel cofferdam construction risk data for intelligent monitoring of the bridge steel cofferdam construction.

[0015] The purpose of this invention is to develop an intelligent, visual, and digital monitoring system for steel cofferdams using technologies such as BIM, the Internet of Things, and digital twins. Through system development and application, dynamic intelligent assessment of safety risks throughout the construction process of bridge steel cofferdams, intelligent safety warnings, rapid disposal of emergency plans, real-time monitoring and analysis of stress conditions, multi-party collaborative progress control, and visual control of engineering projects can be achieved. It will effectively improve the efficiency and accuracy of steel cofferdam monitoring, reduce safety risks, and enhance the overall digital, intelligent, and scientific management and control of the project. Compared with existing monitoring methods and technologies, it has the following obvious advantages:

[0016] (1) “Visualization”: By mapping the real-time data collected by construction monitoring with the digital twin scene model, the steel cofferdam construction monitoring situation is displayed in real time in the visualization scene, creating a visual intelligent perception system for steel cofferdam construction, which greatly improves the efficiency of steel cofferdam monitoring and the visual intelligent perception capability.

[0017] (2) “Measurable”: By collecting the status data of the steel cofferdam in real time through Beidou, strain gauges, thermometers and other equipment, and conducting multiple comparative analyses with built-in standards, the steel cofferdam structure and position status can be monitored in real time, which greatly improves the accuracy and timeliness of monitoring and effectively reduces the safety risks of structural construction.

[0018] (3) “Controllable”: By integrating IoT data with digital twin scenarios, the steel cofferdam construction equipment can be automatically controlled based on the safety and position status data of the steel cofferdam monitoring structure, providing precise control for the steel cofferdam monitoring and construction, improving the accuracy and timeliness of the steel cofferdam monitoring and construction, and ensuring construction safety and quality.

[0019] (4) “Intelligent Early Warning” integrates digital twin scenarios and IoT data, has built-in security assessment and early warning algorithms, conducts real-time analysis and comparison of monitoring data, conducts real-time intelligent dynamic assessment of safety status, makes multiple comparisons between monitoring values ​​and early warning values, automatically activates multiple early warnings, and automatically initiates corresponding emergency plans, thereby improving safety management efficiency and reducing construction safety risks.

[0020] (5) "Intelligent emergency response" is achieved through real-time linkage between digital twin model scenarios and intelligent assessment, safety warning, and monitoring data. Through the digital twin model, the digital twin model can intelligently learn safety emergency response measures (personnel evacuation, reduction of pumping rate, backflow in the cofferdam). When the safety emergency start-up standard is triggered, the emergency plan is automatically started and the emergency measures are automatically pushed to the relevant responsible persons to carry out emergency measures such as organizing personnel evacuation, automatic broadcasting, and alarm bells. After the evacuation of personnel, the equipment is automatically controlled to carry out emergency measures such as water pumping backflow, realizing intelligent handling of emergency plans, improving the efficiency of emergency time processing, and reducing the possibility of safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings:

[0022] Figure 1 A flowchart of a method for intelligent monitoring of bridge steel cofferdam construction based on digital twins according to an embodiment is shown;

[0023] Figure 2 A flow chart showing the steps of a method for collecting data for bridge steel cofferdam construction according to an embodiment is shown;

[0024] Figure 3 A flowchart of the steps of a method for creating a full-factor digital twin model according to an embodiment is shown;

[0025] Figure 4 A flowchart showing a method for fusion mapping of a dynamic model of a bridge steel cofferdam according to an embodiment is shown;

[0026] Figure 5 A flowchart showing a method for dynamic risk assessment of bridge steel cofferdam construction according to an embodiment is shown;

[0027] Figure 6 A diagram showing the effect of an intelligent monitoring method for bridge steel cofferdam construction according to an embodiment is shown;

[0028] Figure 7 A flow chart of intelligent monitoring operation of bridge steel cofferdam construction according to one embodiment is shown;

[0029] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0030] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0031] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0032] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0033] With the continuous development of new technologies such as 5G, Internet of Things, cloud computing, big data, artificial intelligence, GIS, BIM, etc., technical guarantees are provided for the intelligent, rapid and scientific monitoring of super-large bridge and super-large deep-water foundation steel cofferdam. It makes it possible to realize intelligent safety early warning, establishment of safety emergency system, real-time monitoring and analysis of stress conditions, multi-party collaborative progress control, and visual control of engineering projects throughout the construction process of bridge steel cofferdam. It realizes intelligent and scientific monitoring of safety, progress, quality and other factors of deep-water foundation steel cofferdam and foundation construction in special environments. Reducing safety risks and changing passive control to active early warning will make a qualitative leap in safety monitoring, risk assessment, progress control, and emergency disposal management of deep-water foundation steel cofferdam construction.

[0034] Implementation path:

[0035] (1) Using the open source digital twin model scenario engine, build a digital twin scenario of all elements of the bridge steel cofferdam (structure, monitoring equipment, etc.) to form a digital twin model scenario of the bridge steel cofferdam;

[0036] (2) During the construction of the bridge steel cofferdam, install IoT devices such as Beidou, water level gauges, strain gauges, displacement gauges, and thermometers to collect real-time monitoring data of the entire process of steel cofferdam construction and clean the data (discrete data, mutation data, etc.);

[0037] (3) Through the development of data fusion sharing interface protocols and equipment coding, the steel cofferdam monitoring data is mapped to the corresponding position of the steel cofferdam digital twin scene model, so as to realize the visual monitoring of the steel cofferdam construction and accurately display the status of the steel cofferdam in real time;

[0038] (4) By establishing an intelligent assessment algorithm for the dynamic safety risk of steel cofferdams, the current safety risk level can be assessed by automatically analyzing and comparing the real-time dynamic monitoring data of steel cofferdams;

[0039] (5) Automatically start safety warnings based on the comparison between monitoring data and early warning system data, and initiate corresponding emergency plans based on the expected system. At the same time, the corresponding equipment can be automatically controlled in the digital twin scene to handle the early warning situation (automatically open / close cooling water pipes, broadcast announcements, water pump backflow, alarms, and pumping rate control based on thermometer data);

[0040] (6) Through massive scenes and monitoring data, the digital twin scene algorithm is continuously updated and iterated, ultimately achieving visualization, intelligent, real-time, accurate monitoring and early warning of the entire bridge steel cofferdam construction process.

[0041] 1. Real-time dynamic massive data collection and processing

[0042] Using IoT monitoring equipment such as Beidou, strain gauges, displacement gauges, thermometers, and water level gauges, real-time data on force, deformation, displacement, concrete temperature, and water level are collected during the entire process of steel cofferdam construction. The collection frequency is 10 minutes per time. All data from the entire process of steel cofferdam construction are collected into a unified database, and the data of each system is processed during the collection process to ensure the accuracy and integrity of the data. This avoids deviations in data analysis and processing during the subsequent data fusion process.

[0043] For the structured data collected into the database, the data is cleaned and organized according to the application requirements and standards of steel cofferdam monitoring. The main process is to deduplicate, verify and repair all data through data algorithms. For data obtained from unstructured data sources, it is necessary to convert them into a structured format to ensure the accuracy and completeness of the data, so as to serve the subsequent steel cofferdam monitoring and analysis processing.

[0044] 2. Creation of a full-factor digital twin model

[0045] Using digital twin scene construction technology, the steel cofferdam structure model of the bridge is truly restored according to the design drawing of the bridge steel cofferdam, including cofferdam ribs, compartment plates, inner and outer wall plates, supports and other components, 1:1 three-dimensional true restoration, and the digital twin infrastructure scene is constructed. The monitoring equipment is selected according to the monitoring plan, and the monitoring equipment model is constructed according to the equipment type, structure, location and other information. According to the environment where the bridge is located, models such as water flow environment, GIS information, and main structure of the bridge are constructed, and the actual location information is matched with the steel cofferdam structure model to construct a full-factor digital twin model of the steel cofferdam.

[0046] 3. Integration of massive real-time dynamic monitoring data and digital twin models

[0047] A coding system for monitoring equipment and steel cofferdam structure models is established based on the monitoring data interface standards and information such as the name, type, and spatial location attributes of the monitoring equipment. Through the data interface standards and coding system, a unified data transmission interface and general protocol are established to integrate massive IoT data with the digital twin model, and map the physical space data in the digital twin scene to achieve real-time linkage between the physical world and the virtual world of steel cofferdam construction.

[0048] 4. Full-process monitoring of steel cofferdam based on digital twin

[0049] The monitoring data is stored in the digital twin scene model in real time, and the monitoring equipment model can be selected at any time to view the real-time monitoring data. Set the warning threshold to ensure the normal operation of the steel cofferdam. Once the threshold is exceeded, an automatic alarm is pushed to the relevant person in charge. The person in charge makes proper handling based on the feedback data from the platform to reduce safety hazards. Through data fusion, directly select the model component in the digital twin platform to obtain on-site construction data, measurement point data records and change trends, comprehensively judge the construction safety status, simplify the information acquisition process, facilitate information exchange among users, and improve communication efficiency. During the lifting and lowering process of the steel cofferdam and in daily conditions, the real-time data of the steel cofferdam construction is collected and analyzed, problems that arise during the construction process are prevented, and abnormal parts of the monitoring data are promptly fed back to the relevant person in charge for processing. The problem loop is realized in the platform, and dynamic reminders are given for overdue problems. Realize the full process, all elements, and real-time dynamic monitoring of the steel cofferdam.

[0050] 5. Intelligent dynamic assessment of security risks

[0051] ① Establish a dynamic safety risk assessment index system based on the steel cofferdam monitoring standards and construction site monitoring measurements;

[0052] ②Use the analytic hierarchy process to determine the weight coefficient of each dynamic monitoring indicator in the safety risk assessment;

[0053] ③ A safety risk assessment algorithm model is built into the digital twin monitoring system. The digital twin monitoring system automatically compares and calculates the unacceptable value and satisfactory value based on the real-time monitoring data, and automatically calculates the safety risk level of the current steel cofferdam;

[0054] ④ Take risk control measures during the construction process according to the safety risk level, including strengthening monitoring and measurement, temporary reinforcement measures, etc.;

[0055] ⑤ After taking risk control measures during construction, re-evaluate the safety risks. When high risks continue to occur, the system will push measures such as changes in construction methods, optimization of construction processes, and strengthening measures to ensure structural safety;

[0056] ⑥When high risk turns into low risk, it will automatically enter the next stage of risk intelligent assessment.

[0057] The intelligent assessment process of safety risks during the construction process is displayed in real time in the digital twin scene.

[0058] The specific algorithm of the intelligent evaluation algorithm model is:

[0059] 1) Calculate the safety risk assessment scores of each indicator:

[0060] Where: X iis the actual value of the i-th (i=1, 2, 3…m) evaluation index; X hi is the satisfaction value of the i-th indicator; X si is the unallowed value of the ith indicator; d i is the safety risk assessment score of the ith evaluation indicator.

[0061] 2) According to the weight of each evaluation index, calculate the total score of the safety risk assessment of the steel cofferdam:

[0062] Where: D is the total score of the safety risk assessment of the steel cofferdam; ω i is the normalized weight coefficient of the i-th evaluation index.

[0063] Table 1 Safety risk intelligent assessment construction suggestion table

[0064]

[0065] 6. Intelligent early warning of security risks

[0066] A dual warning mechanism and a three-level warning system are established within the system based on the digital twin algorithm model and intelligent safety risk assessment. Intelligent analysis and warning are carried out based on real-time automatic comparison and analysis of monitoring data to improve the reliability and timeliness of structural warnings.

[0067] 1) Dual early warning mechanism

[0068] The warning value of each monitoring device is determined based on structural digital simulation analysis, material properties, monitoring specifications, construction conditions, etc. The monitoring data is collected in real time through the monitoring equipment, and the monitoring data is compared and analyzed with the warning value in real time. The data of adjacent monitoring devices is automatically obtained through the digital twin scene model for comparative analysis. Double verification is performed to intelligently judge the current warning situation. When it is judged that the warning situation is real, a warning is issued. If the warning situation is false, the cause of the data abnormality (equipment abnormality, etc.) is automatically analyzed and the warning situation is ignored.

[0069] 2) Three-level early warning system

[0070] A three-level early warning system is established according to different situations. The early warning monitoring data and safety risk level are compared and analyzed with the early warning value in real time. The early warning level is intelligently judged and pushed to the corresponding responsible personnel according to the early warning level. Level 1 early warning: When the monitoring value reaches 80% of the early warning value, it is pushed to the front-line technical personnel of the construction. Level 2 early warning: When the monitoring value reaches the early warning value, it is pushed to the leaders in charge of the construction unit and the on-site supervisor. Level 3 early warning: When the monitoring value reaches the bearing limit of the structural material, it is pushed to the project leaders of each participating unit and the corresponding emergency plan is activated.

[0071] 6. Intelligent handling of emergency plans

[0072] The emergency measures library and emergency plan triggering standards are built into the system, and the emergency plan measures are associated with the monitoring equipment. When the structural safety is monitored to meet the emergency plan and early warning triggering conditions, the digital twin monitoring system will automatically start the emergency plan, quickly push the corresponding measures of the emergency plan to the relevant responsible persons, and automatically start the relevant emergency disposal equipment to deal with the emergency situation. For example, when it is monitored that the head difference between the inside and outside of the steel cofferdam is large, the structural stress is large, it meets the emergency plan triggering conditions, and meets the personnel evacuation standards, it will automatically start broadcasting and alarms to guide the construction personnel to evacuate. After the personnel evacuate, the digital twin model will automatically control the water pump to start backflow, balance the head difference between the inside and outside, and ensure the safety of the structure. When it is monitored that the internal temperature of the concrete of the foundation in the steel cofferdam rises sharply, the digital twin model will automatically control the cooling water pipe switch to reduce the internal temperature of the concrete. Realize real-time linkage and intelligent control between the physical world and the virtual world, improve the efficiency of emergency disposal in the entire process of steel cofferdam construction, and avoid safety accidents.

[0073] 7. Algorithm model update iteration and intelligent control based on digital twins

[0074] 1) A two-way comparative analysis is conducted based on the monitoring data during the monitoring process of the steel cofferdam and the structural safety status of the steel cofferdam, and the monitoring data is transmitted to the algorithm model based on the digital twin in real time. The big data algorithm model uses the digital twin model to reversely and intelligently control the pumping rate of the steel cofferdam in the pumping condition (the pumping condition is the most dangerous) in real time to ensure the structural safety during the pumping condition, and save the data in real time to the big data model for subsequent application in related engineering cases.

[0075] 2) Based on the monitoring data during the steel cofferdam monitoring process, the current safety status of the steel cofferdam structure is intelligently judged. When the intelligent monitoring detects that the structure is severely deformed and reaches the safety warning standard and emergency response measure standard, the digital twin model will automatically remotely control the construction site broadcast to broadcast emergency measures, call on construction personnel to organize evacuation, and automatically remotely control water pumps and other equipment to pump water back into the steel cofferdam after the personnel evacuate, eliminating the head difference and ensuring the safety of the steel cofferdam structure. Realize intelligent, automated, and remote control of emergency response during steel cofferdam construction. All monitoring data and response measures in the process are stored in the digital twin model to provide reference for subsequent construction and similar engineering construction.

[0076] The present invention creatively constructs a digital twin scene for all-weather, all-element, and full-process safety monitoring of bridge steel cofferdams, realizes real-time interconnection between the physical world and virtual reality of bridge steel cofferdams, improves the level of visualization and intelligent monitoring of steel cofferdams, effectively reduces the safety risks and monitoring costs of steel cofferdam monitoring personnel, and greatly improves the timeliness and accuracy of steel cofferdam monitoring.

[0077] A fusion application of digital twin technology and intelligent algorithms was proposed to automatically monitor the safety status of the steel cofferdam structure in real time. Combined with the intelligent analysis model, the safety risk level of the steel cofferdam structure was automatically judged and automatic early warning was issued based on the results of real-time data monitoring and artificial intelligence analysis. This greatly improved the timeliness and effectiveness of safety early warnings, put safety early warnings in place, and effectively ensured the safety of steel cofferdam construction.

[0078] It has pioneered the intelligent and rapid handling of emergency plans for the entire process of bridge steel cofferdam construction based on digital twins. According to the system's intelligent safety risk assessment and intelligent safety warning, the digital twin system is used to intelligently and automatically control related equipment to quickly carry out emergency plan handling, improve the efficiency of safety emergency response, reduce the safety risks of construction and monitoring operations, and help the development of steel cofferdam construction monitoring towards digitalization, intelligence, convenience and automation.

[0079] See also Figures 1 to 7 The present application provides a method for intelligent monitoring of bridge steel cofferdam construction based on digital twins, the method comprising:

[0080] S1. Obtain the construction data of bridge steel cofferdam;

[0081] Specifically, IoT devices such as stress gauges, strain gauges, displacement gauges, thermometers, and water level gauges are installed at the construction site. Data is automatically collected every 10 minutes and uploaded to the central database. Python scripts are used to deduplicate, fill in gaps, verify, and correct the collected structured data. Natural language processing (NLP) and image recognition technology are used to convert on-site photos, videos, and text records into structured data. The cleaned structured data and the converted structured data are merged to form a complete construction data set.

[0082] S2. Create a full-factor digital twin model based on the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model;

[0083] Specifically, obtain the design drawings, structural parameters and material information of the bridge steel cofferdam from the design department. Use CAD or BIM software to build a basic three-dimensional model of the bridge steel cofferdam. On the basis of the basic model, add cofferdam ribs, bulkheads, inner and outer wall panels, supports and other components. Use the environmental parameters of the construction site (such as water level, flow rate, soil properties, meteorological conditions, etc.) to build an environmental model. Integrate the structural refinement model with the environmental model to form a complete digital twin model.

[0084] S3, obtaining real-time bridge steel cofferdam data, and performing fusion mapping based on the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model to obtain a bridge steel cofferdam dynamic model;

[0085] Specifically, the operation data of the bridge steel cofferdam is collected in real time through the Internet of Things devices. The real-time data is preprocessed by standardization, denoising, missing value filling and other operations. The preprocessed data is extracted using a deep learning algorithm. The mapping model is constructed and the mapping matrix is ​​generated by using historical data and multi-source data fusion technology. The real-time data is fused and mapped with the digital twin model, and the digital twin model is dynamically updated.

[0086] S4. Perform real-time monitoring based on the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data, and perform dynamic risk assessment based on the bridge steel cofferdam construction monitoring data to obtain the bridge steel cofferdam construction risk data to carry out intelligent monitoring operations for the bridge steel cofferdam construction.

[0087] Specifically, various sensor devices (such as stress gauges, strain gauges, displacement gauges, thermometers, water level gauges, etc.) are installed and configured for real-time data acquisition. Vishay strain gauges, Omega temperature sensors, and Trimble water level gauges are used. The data collected by the sensors are transmitted to the central database in real time through a wireless network. A visualization platform (such as Grafana) is used to display real-time data and monitor the construction status of the bridge steel cofferdam. Indicators and thresholds for safety risk assessment are set, including parameters such as stress, strain, displacement, temperature, and water level. Stress threshold: 200MPa, strain threshold: 0.005, displacement threshold: 5cm, temperature threshold: 30℃, water level threshold: 5 meters. According to different risk levels, corresponding risk control measures are formulated, such as adjusting the construction plan, strengthening monitoring, and strengthening measures. According to the formulated risk control measures, specific control operations are implemented, and the implementation process and results are recorded. Real-time monitoring of changes in risk levels during construction is carried out, and the change data is recorded. () Timestamp, risk level, control measures): (2024-05-21, 10:00, 3, enhanced monitoring), (2024-05-21 10:10, 2, regular monitoring). Establish an intelligent early warning system and set early warning thresholds and processing rules for different risk levels. Push early warning information to relevant persons in charge via SMS, email or mobile application to ensure timely handling of risks. Record all early warning processing processes and results, analyze the data, and optimize the early warning system and construction management.

[0088] The intelligent monitoring operation of bridge steel cofferdam construction is, specifically, to carry out intelligent early warning and emergency response of bridge steel cofferdam. It is further explained that the early warning is carried out through the intelligent early warning of safety risks based on the digital twin model and the emergency plan is handled through the intelligent disposal of the emergency plan based on the digital twin model. Intelligent early warning in the intelligent early warning of safety risks based on the digital twin model: double early warning mechanism->determination of early warning value->real-time comparison and analysis of monitoring data and early warning value->intelligent mutual verification of adjacent equipment monitoring data->true and false judgment of early warning situation. If it is judged to be false, the early warning is ignored; if it is judged to be true, it enters the three-level early warning system. The three-level early warning system includes a first-level early warning and two second-level early warnings. A first-level early warning is: if the early warning value reaches 80%, it will be sent to the front-line technical personnel of the construction via SMS / app; two second-level early warnings are: 1. If the early warning value is reached, it will be sent to the construction unit’s responsible leader and on-site supervisor via SMS / app; 2. If the limit value of the structural material is reached, it will be sent to the project leader of each participating unit via SMS / app. Intelligent handling of emergency plans in the intelligent handling of emergency plans based on digital twin models: emergency plan measures library, emergency plan trigger conditions -> emergency plan trigger -> automatic push of emergency plan measures -> intelligent control of equipment based on digital twin models -> automatic broadcasting, alarm bells, automatic backflow of water pumps, and intelligent control of pumping rate.

[0089] In the present invention, the data of various sensors at the construction site are integrated to provide comprehensive and accurate construction status information. Using digital twin technology, a detailed virtual model is created to enhance the simulation and analysis capabilities of the construction process. Through real-time data mapping, the synchronization of the physical world and the virtual world is achieved, and dynamic and real-time construction status information is provided. Real-time monitoring data is used for intelligent risk assessment to timely discover and deal with potential risks in construction. Taking into full consideration the timeliness of bridge steel cofferdam construction monitoring, dynamic safety risk intelligent assessment, intelligent safety warning, and rapid disposal of emergency plans, through real-time acquisition, data cleaning, analysis and processing, and mapping of bridge steel cofferdam construction data based on digital twin + IOT equipment, stress, strain, displacement, settlement, and temperature data are linked to the structural position in real time, and compared and analyzed in real time according to monitoring standards, safety warnings are given to steel cofferdam structures throughout the construction process, and continuous learning, optimization and upgrading of IoT monitoring equipment are used to intelligently judge the safety status of the structure, and the system and algorithm are continuously upgraded and iterated, ultimately solving the problems of untimely and unintuitive monitoring of bridge steel cofferdam construction, high monitoring personnel costs, and high safety risks. By integrating emergency plan measures with intelligent monitoring, the system automatically starts the emergency plan when it determines that the emergency plan activation standards have been met, pushes emergency measures to relevant responsible personnel as soon as possible, and automatically starts corresponding disposal equipment to complete emergency disposal. Real-time visualization, intelligent, and convenient monitoring, early warning, and emergency disposal of steel cofferdams are achieved, reducing construction safety risks and improving construction efficiency.

[0090] Optionally, S1 is specifically:

[0091] S11. Use the Internet of Things monitoring equipment to collect the entire process of steel cofferdam construction in real time to obtain data on the entire process of steel cofferdam construction;

[0092] Specifically, various sensors such as stress gauges, strain gauges, displacement gauges, thermometers and water level gauges are installed at the construction site. Stress gauges are installed at key structural nodes, displacement gauges are installed at multiple key locations of the bridge, and thermometers and water level gauges are installed at appropriate locations. The collection frequency of the IoT devices is configured to collect data every 10 minutes. The collected data is uploaded to the central database in real time through the wireless network for storage and processing.

[0093] S12, performing data cleaning on the steel cofferdam construction structured data of the entire process data of the steel cofferdam construction to obtain the steel cofferdam construction structured cleaned data;

[0094] Specifically, write Python scripts to remove duplicates, fill in missing values, and perform data verification and correction on the collected structured data. Perform quality checks on the cleaned data to ensure data accuracy and consistency. Use data analysis tools (such as Pandas) to perform data statistics and visualization, and check data distribution and outliers.

[0095] S13, performing structured conversion on the unstructured data of the steel cofferdam construction of the whole process data of the steel cofferdam construction to obtain structured conversion data of the steel cofferdam construction;

[0096] Specifically, NLP technology is used to convert text records into structured data. Image recognition technology is used to convert photos and videos into structured data. Text data processed by NLP and data processed by image recognition are integrated to form structured converted data. Data from different sources are merged to generate a unified structured data set.

[0097] S14, integrating the steel cofferdam construction structured cleaning data and the steel cofferdam construction structured conversion data to obtain the bridge steel cofferdam construction data.

[0098] Specifically, the cleaned structured data is merged with the structured transformed data to form a complete construction data set. The merged data is checked for inconsistencies or conflicts and any necessary corrections are made. Data analysis tools are used to check the integrity and consistency of the data to ensure that the timestamps and other key fields of all data points are consistent.

[0099] The present invention not only ensures the comprehensiveness and real-time nature of data collection, but also improves the quality and availability of data through data cleaning and conversion. Data integration enables all data to be managed and used in a unified manner, providing a solid foundation for the subsequent creation of digital twin models and real-time monitoring. These steps combined provide reliable data support and guarantee for the intelligent monitoring of bridge steel cofferdam construction, improving the overall safety, efficiency and management level of construction.

[0100] Optionally, S2 is specifically:

[0101] S21. Obtaining bridge steel cofferdam design data;

[0102] Specifically, the design drawings, structural parameters and material information of the bridge steel cofferdam are obtained from the design department. Design parameters: the bridge length is 200 meters, the width is 30 meters, the main material is reinforced concrete, and the design strength is 500MPa. Design drawings: including bridge plan, section, and structural detail drawings. Use an SQL database (such as MySQL) to store design parameters and drawing paths.

[0103] S22, constructing a bridge steel cofferdam basic model according to the bridge steel cofferdam design data to obtain a bridge steel cofferdam basic model;

[0104] Specifically, use CAD or BIM software to build a basic 3D model of the bridge steel cofferdam according to the design data. Import the design drawings in Revit. Create a 3D model of the bridge according to the drawings and design parameters, including the foundation structure, main bridge body, supporting structure, etc.

[0105] S23, structurally refining the bridge steel cofferdam foundation model to obtain a bridge steel cofferdam structural refinement model;

[0106] Specifically, on the basis of the basic model, add detailed components such as cofferdam ribs, bulkheads, inner and outer wall panels, and supports. Continue to use Autodesk Revit for detailed modeling. Load the basic model in Revit. According to the design drawings and construction specifications, add various detailed components such as ribs, bulkheads, etc. Ensure that the position, size, and material of all components meet the design requirements.

[0107] S24, constructing a bridge steel cofferdam environmental model according to the environmental parameter data in the bridge steel cofferdam construction data to obtain a bridge steel cofferdam environmental model;

[0108] Specifically, collect environmental parameter data of the construction site, including water level, flow rate, soil properties, meteorological conditions, etc. Water level: 3.2 meters, flow rate: 0.5 meters per second, soil type: clay, meteorological conditions: sunny, temperature 25°C. Use environmental modeling software to build an environmental model based on the collected environmental parameter data. Use ArcGIS or InfraWorks for environmental modeling. Import environmental parameter data into ArcGIS. Generate a hydrological model based on water level, flow rate and other data. Generate geological and meteorological models based on soil properties and meteorological conditions. Integrate different environmental sub-models into a complete environmental model. Import each sub-model into InfraWorks for integration and correction. Ensure the coherence and consistency of the model.

[0109] S25. The detailed model of the bridge steel cofferdam structure and the environmental model of the bridge steel cofferdam are integrated to obtain a digital twin model of the bridge steel cofferdam.

[0110] Specifically, the refined structural model and environmental model are imported into a unified digital twin platform for fusion. Use Autodesk Revit or Bentley iTwin for model fusion. Import the refined structural model in Revit. Import the environmental model and align it with the structural model. Ensure that environmental parameters and structural details are accurately displayed on the same platform.

[0111] The present invention not only ensures the comprehensiveness and accuracy of model construction, but also makes the model closer to the actual situation by comprehensively considering environmental factors. Model fusion enables all details and environmental factors to be managed and used in a unified manner, providing a solid foundation for subsequent real-time monitoring, simulation analysis and construction optimization. These steps combined provide reliable technical support for the intelligent monitoring of bridge steel cofferdam construction, and improve the safety, efficiency and management level of the overall construction.

[0112] Optionally, S3 is specifically:

[0113] S31, obtaining real-time bridge steel cofferdam data;

[0114] Specifically, various sensors (such as stress gauges, strain gauges, displacement gauges, thermometers and water level gauges) are used to collect the operation data of the bridge steel cofferdam in real time. Vishay's strain gauges, Omega's temperature sensors, and Trimble's water level gauges are used. The MQTT protocol is used for data transmission.

[0115] S32, performing planning processing according to the real-time bridge steel cofferdam data to obtain the planned data of the bridge steel cofferdam;

[0116] Specifically, perform preprocessing operations such as standardization, denoising, and missing value filling on real-time data. Convert the preprocessed data into a format suitable for subsequent processing and analysis. Use Pandas to convert data formats.

[0117] S33, performing deep multi-dimensional feature extraction on the planned data of the bridge steel cofferdam to obtain characteristic data of the bridge steel cofferdam;

[0118] Specifically, use TensorFlow or PyTorch for feature extraction. Save the extracted feature data as a new dataset for subsequent use. import tensorflow as tf; #define model model = tf.keras.Sequential([tf.keras.layers.Dense(128, activation = 'relu', input_shape = (real_time_data.shape[1],)), tf.keras.layers.Dense(64, activation = 'relu'), tf.keras.layers.Dense(32, activation = 'relu')]) #compile model model.compile(optimizer = 'adam', loss = 'mean_squared_error') #train model (using part of historical data) model.fit(historical_data, historical_labels, epochs = 10, batch_size = 32) #feature extraction features = model.predict(real_time_data).

[0119] S34, generating mapping data through a preset physical monitoring digital twin mapping model according to the characteristic data of the bridge steel cofferdam to obtain mapping matrix data;

[0120] Specifically, historical data and multi-source data fusion technology are used to build and train the mapping model. According to the trained mapping model, the extracted feature data is mapped to the corresponding part of the digital twin model. The mapping data is generated using linear regression coefficients.

[0121] S35. According to the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model, a fusion mapping is performed through mapping matrix data to obtain a dynamic model of the bridge steel cofferdam.

[0122] Specifically, real-time data is fused and mapped with the digital twin model, and the digital twin model is dynamically updated. Use Python scripts to implement data fusion and model update. def update_digital_twin(real_time_data,digital_twin_model,mapping_matrix):#Get the current real-time data features features = model.predict(real_time_data)#Generate mapping data mapped_data = np.dot(features,mapping_matrix.T)#Update the digital twin model digital_twin_model.update(mapped_data); return digital_twin_model; #Real-time update digital_twin_model = update_digital_twin(real_time_data,digital_twin_model,mapping_matrix).

[0123] In the present invention, multi-dimensional data is collected in real time through a variety of sensors to ensure the comprehensiveness and real-time nature of the data, providing a reliable basis for subsequent processing and analysis. Data quality is improved through data preprocessing to ensure the consistency and reliability of input data, providing high-quality data input for deep feature extraction. Multi-dimensional feature extraction is performed using deep learning algorithms to capture complex patterns and relationships in the data, improving the understanding and prediction capabilities of the operating status of bridge steel cofferdams. Through a preset mapping model, physical data is accurately mapped to the corresponding parts of the digital twin model to achieve accurate association and real-time synchronization between the physical and virtual models. Through the fusion mapping of real-time data and the digital twin model, the model is dynamically updated to achieve synchronization between the physical world and the virtual world, providing accurate real-time monitoring and analysis.

[0124] Optionally, S4 is specifically:

[0125] S41, performing real-time monitoring according to the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data;

[0126] Specifically, the construction process of the bridge steel cofferdam is monitored in real time through IoT devices and digital twin models, and various construction data are collected, including stress, strain, displacement, temperature, and water level. The monitoring data is stored in a central database in real time, and the real-time status of the construction site is displayed through a visualization platform.

[0127] S42, calculating the safety risk level according to the bridge steel cofferdam construction monitoring data using the preset safety risk assessment index data to obtain safety risk level data;

[0128] Specifically, according to industry standards and construction specifications, safety risk assessment indicators are set, including thresholds for stress, strain, displacement, temperature and water level. Stress threshold: 200MPa, strain threshold: 0.005, displacement threshold: 5cm, temperature threshold: 30℃, water level threshold: 5m.

[0129] S43, performing risk control operations according to the safety risk level data to obtain safety risk level change data;

[0130] Specifically, formulate corresponding risk control measures according to the risk level, such as adjusting the construction plan, strengthening monitoring, and strengthening measures. Low risk (1-2): Routine monitoring, no special measures are required. Medium risk (3-4): Strengthen monitoring and increase the frequency of inspections. High risk (5): Immediately suspend construction for detailed inspection and reinforcement. Implement specific control operations according to the formulated risk control measures, and record the implementation process and results. Install additional sensors to strengthen monitoring. Adjust the construction plan to reduce the intensity of operations in high-risk areas. Monitor the changes in risk levels during the construction process in real time and record the changes.

[0131] S44. Perform intelligent early warning processing of safety risks based on the safety risk level change data to obtain the risk data of bridge steel cofferdam construction so as to carry out intelligent monitoring of bridge steel cofferdam construction.

[0132] Specifically, an intelligent early warning system is established, and early warning thresholds and processing rules are set for different risk levels. Risk level change data is analyzed in real time, and once the risk level exceeds the preset threshold, an early warning is automatically triggered. Early warning information is pushed to relevant persons in charge via SMS, email or mobile applications to ensure timely handling of risks. All early warning processing processes and results are recorded, and the data is analyzed to optimize the early warning system and construction management.

[0133] The present invention uses a digital twin model for real-time monitoring to ensure the comprehensiveness and real-time nature of construction data, and to provide reliable data support for subsequent risk assessment and control. The real-time monitoring data is analyzed and calculated using preset safety risk assessment indicators to accurately assess the safety risk level of the construction status, and to promptly identify and reflect construction risks. Through the intelligent early warning system, early warning processing is performed based on risk level change data, and real-time early warning information is sent to relevant personnel to prompt potential risks and response measures, thereby improving the timeliness and accuracy of early warnings. Based on the real-time risk assessment results, corresponding risk control measures are formulated and implemented to ensure the safety and stability of the construction process, and to improve the level of construction management and emergency response capabilities.

[0134] Optionally, the steps for constructing the physical monitoring digital twin mapping model preset in S34 are specifically as follows:

[0135] Obtain historical bridge steel cofferdam data;

[0136] Specifically, historical data accumulated during the construction and operation of the bridge steel cofferdam are collected, including stress, strain, displacement, temperature, water level, etc.

[0137] Multi-source data fusion is performed based on historical bridge steel cofferdam data to obtain multi-source fusion data of bridge steel cofferdam data;

[0138] Specifically, the data from different sensors and monitoring devices are fused to eliminate redundant information and unify the data format. The fused data is cleaned to remove noise and outliers to ensure data accuracy.

[0139] Perform deep feature extraction on multi-source fusion data of bridge steel cofferdam data to obtain deep feature data of bridge steel cofferdam;

[0140] Specifically, deep feature extraction is performed through a deep learning model, and an autoencoder is selected as a deep learning model because it can effectively extract data features in unsupervised learning. The autoencoder consists of an encoder and a decoder, and learns the feature representation of data through the process of compressing and restoring data.

[0141] A bridge steel cofferdam monitoring dataset, containing the following features: Feature 1: Timestamp, Feature 2: Temperature, Feature 3: Water level, Feature 4: Environmental stress, Feature 5: Environmental displacement. After preprocessing and standardization, the dataset is stored in the merged_data.csv file. Timestamp, temperature, water level, environmental stress, environmental displacement, 1, 25.0, 3.2, 150, 2.5, 2, 24.8, 3.1, 148, 2.4, ... Standardized data: scaled_data[:5], output: array([[-0.5, 0.1, 0.2, 1.1, -0.3], [-0.3, 0.0, 0.1, 0.9, -0.4], .... Autoencoder encoding output (low-dimensional feature representation): encoded_data[:5]. Output: tensor([[0.3, -0.2, 0.1, ..., -0.1, 0.2, 0.3], [0.2, -0.1, 0.0, ..., -0.2, 0.1, 0.4], .... Output data, deep feature data: feature 1, feature 2, ..., feature 32, 0.3, -0.2, ..., 0.3, 0.2, -0.1, ..., 0.4, ....

[0142] Conduct time series analysis on the depth characteristic data of bridge steel cofferdam to obtain the characteristic time series analysis data of bridge steel cofferdam;

[0143] Specifically, time series analysis methods (such as LSTM and GRU) are used to perform time series analysis on deep feature data to capture the time dependency and dynamic change patterns of the data.

[0144] An adaptive mapping model is constructed based on the characteristic time series analysis data of the bridge steel cofferdam, and an adaptive mapping model of the bridge steel cofferdam is obtained;

[0145] Specifically, an adaptive neural network model is used to build a mapping model that can adapt to different construction environments and conditions based on time series analysis data. Deep learning technology is used to train the adaptive mapping model. The trained adaptive mapping model is verified to ensure the accuracy and generalization ability of the model. The model performance is evaluated through cross-validation.

[0146] The multi-task joint learning is performed on the adaptive mapping model of bridge steel cofferdam to obtain the multi-task learning model of bridge steel cofferdam;

[0147] Specifically, we use multi-task learning technology to simultaneously learn and optimize multiple related tasks to improve the overall performance of the model. from tensorflow.keras.models import Model; fromtensorflow.keras.layers import Input,Dense; #define input layerinput_layer=Input(shape=(timeseries_features.sha pe[1],))#Define shared layer shared_layer=Dense(128,activation='relu')(input_layer)#Define task 1 output layer task1_output=Dense(1,name='task1_output')(shared_layer)#Define task 2 output layer task2_output=Dense(1,name='task2_output')(shared_layer)#Define multi-task model model=Model(inputs=input_layer,outputs=[task1_output,task2_output])#Compile model model.compile(optimizer='adam',loss='mean_squared_error')#Train model model.fit(timeseries_features,[task1_labels,task2_labels],epochs=10,batch_size=32)#Generate multi-task learning model multi_task_model=model.

[0148] Specifically, prepare datasets for multiple tasks, standardize and split them to ensure data consistency. Build a multi-task learning model to learn the objective functions of multiple tasks simultaneously. Train the model using the training data and optimize the objective functions of multiple tasks simultaneously. Evaluate the model performance using the test data and calculate the prediction error for each task. Divide the dataset into training and test sets and standardize them. Build a multi-task learning model with shared layers and task-specific layers. Train the model using the training data and optimize the objective functions of the stress prediction task and the displacement prediction task simultaneously. Evaluate the model performance using the test data and calculate the prediction error for each task.

[0149] Specifically, different task layers are added to the output layer of the adaptive mapping model of the bridge steel cofferdam to obtain different task learning models, which are trained until the results converge to obtain a multi-task learning model of the bridge steel cofferdam.

[0150] Conduct uncertainty assessment on the multi-task learning model of bridge steel cofferdams and obtain uncertainty assessment report data;

[0151] Specifically, uncertainty quantification methods are used to evaluate the uncertainty of the prediction results of the multi-task learning model, identify and quantify the uncertainty in the prediction. Bayesian deep learning methods are used for uncertainty quantification.

[0152] Specifically, Bayesian neural networks or Monte Carlo Dropout methods are used to perform uncertainty assessment on multi-task learning models to quantify the uncertainty in predictions.

[0153] The multi-task learning model of bridge steel cofferdam is integrated according to the uncertainty assessment report data to obtain the integrated model of bridge steel cofferdam;

[0154] Specifically, based on the uncertainty assessment results, multiple models are used for weighted averaging or stacking to improve the robustness and generalization ability of the model.

[0155] According to the integrated model of bridge steel cofferdam, a dynamic mapping matrix output layer is added to obtain the physical monitoring digital twin mapping model.

[0156] Specifically, the output layer of the integrated model is dynamically mapped with the physical monitoring system to form a digital twin system, which updates and feeds back the physical monitoring data in real time. The output layer of the integrated model is dynamically mapped with the physical monitoring system to form a digital twin system, which updates and feeds back the physical monitoring data in real time.

[0157] class PhysicalMonitoringDigitalTwinModel(nn.Module):def__init__(self,base_model,output_dim):super(PhysicalMonitoringDigitalTwinModel,self).__init__();self.base_model=base_model;self.output_layer=nn.Linear(base_model.hidden_dim,output_dim);def forward(self,x):base_output=self.base_model.shared_layer(x);final_output=self.output_layer(base_output);returnfinal_output#Create digital twin modeloutput_dim=2#Assume that two monitoring parameters are output: stress and displacementdigital_twin_model=PhysicalMonitoringDigitalTwinModel(model,output_dim)#Mapping matrix generationdef generate_mapping_matrix(digital_twin_model,x):digital_twin_model.eval()with torch.no_grad():mapping_matrix = digital_twin_model(x).numpy(); return mapping_matrix #Dynamic mapping matrix output mapping_matrix = generate_mapping_matrix(digital_twin_model, X_test_tensor) #Save mapping matrix np.savetxt('mapping_matrix.csv', mapping_matrix, delimiter = ',').

[0158] In the present invention, multi-source data is integrated to form a comprehensive and consistent data set, which improves the comprehensiveness and quality of the data and provides a solid foundation for subsequent processing. Deep learning and time series analysis methods are used to capture complex patterns and dynamic changes in the data, and improve the understanding and prediction capabilities of the model. Through adaptive mapping models and multi-task learning, the adaptability, generalization ability and comprehensive performance of the model are improved to ensure its reliability in different environments. Uncertainty assessment methods are used to identify and quantify uncertainties in predictions, and model integration is used to improve the accuracy and stability of predictions. On the basis of the integrated model, a dynamic mapping matrix output layer is added to realize the dynamic mapping of physical monitoring data and digital twin models, providing accurate real-time monitoring and analysis capabilities.

[0159] Optionally, the adaptive mapping model is constructed as follows:

[0160] Data enhancement is performed based on the characteristic time series analysis data of the bridge steel cofferdam to obtain characteristic time series enhanced data;

[0161] Specifically, data enhancement is performed on the original time series data using methods including random noise addition, data smoothing, time axis expansion, etc.

[0162] Perform graph construction on the feature time series enhanced data to obtain feature time series graph data;

[0163] Specifically, the enhanced time series data is constructed as a graph data structure, where nodes represent the features of time points and edges represent time series relationships.

[0164] Performing self-attention graph weighted calculation on the feature time series graph data to obtain feature time series graph weighted data;

[0165] Specifically, a self-attention mechanism is introduced into the graph structure data, the self-attention weight of each node is calculated, and different weight values ​​are assigned to different nodes.

[0166] Perform graph convolution calculation on the weighted data of the feature time series graph to obtain the convolution layer data of the feature time series graph;

[0167] Specifically, a graph convolutional network (GCN) is applied to perform convolution operations on the weighted graph data to extract the spatial relationship features between nodes.

[0168] Performing graph pooling layer processing on the feature time series graph convolution layer data to obtain feature time series graph pooling layer data;

[0169] Specifically, pooling operations (maximum pooling layer, minimum pooling layer, and average pooling layer) are performed on the convolutional graph data to reduce the feature dimension and retain important features.

[0170] Perform full connection calculation on the feature time series pooling layer data to obtain the feature time series fully connected layer data;

[0171] Specifically, the pooled feature data is input into the fully connected layer for feature combination and mapping to generate the final feature representation.

[0172] The feature time series fully connected layer data is iteratively trained to obtain an adaptive mapping model.

[0173] Specifically,

[0174] In the present invention, data enhancement technology is used to increase the diversity and richness of the data set, improve the robustness of the model, and prevent overfitting. Through the application of graph construction and graph convolutional networks, the complex spatial and temporal relationships in time series data can be captured, and the data representation ability can be improved. The self-attention mechanism is introduced to enhance the focus on important features through weighted calculations, thereby improving the model's ability to capture key information and accuracy. Through graph convolution calculations and graph pooling operations, local and global features in graph structured data can be efficiently extracted, data dimensions can be reduced, and computing efficiency can be improved. Feature data is combined and mapped through a fully connected layer to enhance the model's expressiveness and predictive performance. The model is adaptively optimized through iterative training, and model parameters are gradually adjusted to improve the model's predictive accuracy and generalization ability.

[0175] Optionally, the self-attention graph weighted calculation is specifically:

[0176] Performing self-attention hierarchy graph weighted calculation on the feature time series graph data to obtain first feature time series graph weighted data;

[0177] Specifically, a hierarchical self-attention mechanism is applied in the graph structure to calculate the importance weight of each node at different levels.

[0178] Perform self-attention cross-graph weighted calculation on the feature time series graph data to obtain second feature time series graph weighted data;

[0179] Specifically, a self-attention mechanism is introduced between different graph structure data to calculate the association weights between different graphs and generate cross-graph feature weighted data.

[0180] The first characteristic timing diagram weighted data and the second characteristic timing diagram weighted data are concatenated to obtain characteristic timing diagram weighted data;

[0181] Specifically, the first characteristic timing diagram weighted data and the second characteristic timing diagram weighted data are concatenated to form final characteristic timing diagram weighted data.

[0182] The weighted calculation of the self-attention hierarchy graph is as follows:

[0183] Perform self-attention global graph weighted calculation on the feature time series graph data to obtain global weighted data of the feature time series graph;

[0184] Specifically, the self-attention mechanism is applied in the graph structure to calculate the importance weight of each node in the global scope.

[0185] Extracting time series variation features according to the characteristic time series graph data to obtain characteristic time series graph variation feature data;

[0186] Specifically, time series variation features are extracted from the characteristic time series graph data to capture the characteristics of data changing over time.

[0187] Extract the change point according to the change feature data of the characteristic time series diagram to obtain the change point data;

[0188] Specifically, the time series change feature data is analyzed, the change points in the data are identified, and the nodes where the data has changed significantly are marked.

[0189] Constructing a correlation graph according to the change point data to obtain the change point correlation graph data;

[0190] Specifically, based on the identified change point data, a correlation diagram between the change points is constructed to analyze the relationship between the change points.

[0191] Generate local boundary parameters according to the change point association graph data to obtain local boundary parameter data;

[0192] Specifically, local boundary parameters are extracted from the change point association graph data to describe the boundaries and relationships between the change points.

[0193] Perform self-attention local graph weighted calculation according to local boundary parameter data and feature time series graph data to obtain local weighted data of the feature time series graph;

[0194] Specifically, the self-attention mechanism is applied in the local scope to calculate the importance weight of the node in the local scope.

[0195] The global weighted data of the characteristic timing diagram and the local weighted data of the characteristic timing diagram are weightedly calculated to obtain the first characteristic timing diagram weighted data.

[0196] Specifically, the global weighted data and the local weighted data are weighted and calculated, and the global and local features are comprehensively considered.

[0197] In the present invention, different self-attention graph weighted calculations are combined to obtain deep and accurate feature time series graph weighted data, wherein the self-attention hierarchical graph weighted calculations are performed through global and local self-attention weighted calculations to comprehensively capture the global and local features in the graph data and provide a more comprehensive feature representation. Through time series change feature extraction and change point extraction, the dynamic changes of data over time are identified, and the model's ability to understand time-dependent characteristics is improved. Through cross-graph weighted calculations, the correlation information between different graphs is captured, and the model's comprehensive analysis ability for multi-graph data is enhanced. The self-attention mechanism is introduced, and the focus on important features is enhanced through weighted calculations, thereby improving the accuracy and depth of feature extraction. By splicing global and local weighted data, the richness and diversity of feature representation are enhanced, and the comprehensive performance of the model is improved. Through graph convolution and graph pooling operations, the data structure is simplified, the computational complexity is reduced, and the expression ability and computational efficiency of the model are improved. The self-attention cross-graph weighted calculation takes into account the inherent connection between different graphs, so that the self-attention weighted calculation has a more extensive and in-depth characteristic.

[0198] Optionally, the self-attention cross-graph weighted calculation is specifically:

[0199] Linearly mapping the characteristic timing diagram data to obtain characteristic timing diagram mapping data;

[0200] Specifically, the feature time series graph data is mapped to a new feature space through linear transformation to facilitate subsequent similarity calculation and attention weight calculation.

[0201] Perform node similarity calculation on the feature time series graph mapping data to obtain node similarity data;

[0202] Specifically, the similarities between nodes in the mapping feature matrix are calculated to form a node similarity matrix.

[0203] The self-attention weight is calculated based on the node similarity data and the feature time series graph mapping data to obtain the cross-graph attention weight data;

[0204] Specifically, based on the node similarity matrix, the Softmax function is used to calculate the attention weight matrix, and different weights are assigned to different nodes.

[0205] A cross-graph feature weighted calculation is performed on the feature time series graph data according to the cross-graph attention weight data to obtain second feature time series graph weighted data.

[0206] Specifically, the attention weight matrix is ​​applied to the node feature matrix to weight the node features and generate a weighted feature matrix.

[0207] In the present invention, the original feature data is converted into a unified feature representation through linear mapping, which is convenient for subsequent similarity calculation and attention weight calculation, and improves the processability of the data. Through node similarity calculation, the similarity relationship between nodes is captured, and basic data is provided for attention weight calculation, which enhances the model's understanding of node relationships. The self-attention mechanism is introduced to enhance the attention to important nodes and features by calculating the attention weights, thereby improving the accuracy and depth of feature extraction. Through weighted calculation, the node features are redistributed according to the weights to generate weighted feature data, thereby enhancing the diversity and accuracy of feature representation. Through cross-graph weighted calculation, the correlation information between different graphs is captured, and the model's comprehensive analysis capability for multi-graph data is improved.

[0208] Optionally, the present application further provides a bridge steel cofferdam construction intelligent monitoring system based on digital twin, which is used to execute the bridge steel cofferdam construction intelligent monitoring method based on digital twin as described above, and the bridge steel cofferdam construction intelligent monitoring system based on digital twin includes:

[0209] Bridge steel cofferdam construction data acquisition module, used to obtain bridge steel cofferdam construction data;

[0210] A full-factor digital twin model creation module is used to create a full-factor digital twin model based on the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model;

[0211] A real-time bridge steel cofferdam data fusion mapping model is used to obtain real-time bridge steel cofferdam data, and to perform fusion mapping based on the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model to obtain a bridge steel cofferdam dynamic model;

[0212] The intelligent monitoring module for bridge steel cofferdam construction is used to perform real-time monitoring based on the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data, and to perform dynamic risk assessment based on the bridge steel cofferdam construction monitoring data to obtain the bridge steel cofferdam construction risk data for intelligent monitoring of the bridge steel cofferdam construction.

[0213] The purpose of this invention is to develop an intelligent, visual, and digital monitoring system for steel cofferdams using technologies such as BIM, the Internet of Things, and digital twins. Through system development and application, dynamic intelligent assessment of safety risks throughout the construction process of bridge steel cofferdams, intelligent safety warnings, rapid disposal of emergency plans, real-time monitoring and analysis of stress conditions, multi-party collaborative progress control, and visual control of engineering projects can be achieved. It will effectively improve the efficiency and accuracy of steel cofferdam monitoring, reduce safety risks, and enhance the overall digital, intelligent, and scientific management and control of the project. Compared with existing monitoring methods and technologies, it has the following obvious advantages:

[0214] (1) “Visualization”: By mapping the real-time data collected by construction monitoring with the digital twin scene model, the steel cofferdam construction monitoring situation is displayed in real time in the visualization scene, creating a visual intelligent perception system for steel cofferdam construction, which greatly improves the efficiency of steel cofferdam monitoring and the visual intelligent perception capability.

[0215] (2) “Measurable”: By collecting the status data of the steel cofferdam in real time through Beidou, strain gauges, thermometers and other equipment, and conducting multiple comparative analyses with built-in standards, the steel cofferdam structure and position status can be monitored in real time, which greatly improves the accuracy and timeliness of monitoring and effectively reduces the safety risks of structural construction.

[0216] (3) “Controllable”: By integrating IoT data with digital twin scenarios, the steel cofferdam construction equipment can be automatically controlled based on the safety and position status data of the steel cofferdam monitoring structure, providing precise control for the steel cofferdam monitoring and construction, improving the accuracy and timeliness of the steel cofferdam monitoring and construction, and ensuring construction safety and quality.

[0217] (4) “Intelligent Early Warning” integrates digital twin scenarios and IoT data, has built-in security assessment and early warning algorithms, conducts real-time analysis and comparison of monitoring data, conducts real-time intelligent dynamic assessment of safety status, makes multiple comparisons between monitoring values ​​and early warning values, automatically activates multiple early warnings, and automatically initiates corresponding emergency plans, thereby improving safety management efficiency and reducing construction safety risks.

[0218] (5) "Intelligent emergency response" is achieved through real-time linkage between digital twin model scenarios and intelligent assessment, safety warning, and monitoring data. Through the digital twin model, the digital twin model can intelligently learn safety emergency response measures (personnel evacuation, reduction of pumping rate, backflow in the cofferdam). When the safety emergency start-up standard is triggered, the emergency plan is automatically started and the emergency measures are automatically pushed to the relevant responsible persons to carry out emergency measures such as organizing personnel evacuation, automatic broadcasting, and alarm bells. After the evacuation of personnel, the equipment is automatically controlled to carry out emergency measures such as water pumping backflow, realizing intelligent handling of emergency plans, improving the efficiency of emergency time processing, and reducing the possibility of safety accidents.

[0219] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0220] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent monitoring method for bridge steel cofferdam construction based on digital twin, characterized in that: The method comprises: S1. Obtain the construction data of bridge steel cofferdam; S2. Create a full-factor digital twin model based on the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model; S3, obtaining real-time bridge steel cofferdam data, and performing fusion mapping based on the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model to obtain a bridge steel cofferdam dynamic model; S4. Perform real-time monitoring based on the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data, and perform dynamic risk assessment based on the bridge steel cofferdam construction monitoring data to obtain the bridge steel cofferdam construction risk data to perform intelligent monitoring operations for the bridge steel cofferdam construction; Among them, S3 is specifically: S31, obtaining real-time bridge steel cofferdam data; S32, performing planning processing according to the real-time bridge steel cofferdam data to obtain the planned data of the bridge steel cofferdam; S33, performing deep multi-dimensional feature extraction on the planned data of the bridge steel cofferdam to obtain characteristic data of the bridge steel cofferdam; S34, generating mapping data according to the characteristic data of the bridge steel cofferdam through a preset physical monitoring digital twin mapping model to obtain mapping matrix data; S35, fusing and mapping the real-time bridge steel cofferdam data and the bridge steel cofferdam digital twin model through mapping matrix data to obtain a bridge steel cofferdam dynamic model; The specific steps for constructing the physical monitoring digital twin mapping model preset in S34 are as follows: Obtain historical bridge steel cofferdam data; Multi-source data fusion is performed based on historical bridge steel cofferdam data to obtain multi-source fusion data of bridge steel cofferdam data; Perform deep feature extraction on multi-source fusion data of bridge steel cofferdam data to obtain deep feature data of bridge steel cofferdam; Conduct time series analysis on the depth characteristic data of bridge steel cofferdam to obtain the characteristic time series analysis data of bridge steel cofferdam; An adaptive mapping model is constructed based on the characteristic time series analysis data of the bridge steel cofferdam, and an adaptive mapping model of the bridge steel cofferdam is obtained; The multi-task joint learning is performed on the adaptive mapping model of bridge steel cofferdam to obtain the multi-task learning model of bridge steel cofferdam; Conduct uncertainty assessment on the multi-task learning model of bridge steel cofferdams and obtain uncertainty assessment report data; The multi-task learning model of bridge steel cofferdam is integrated according to the uncertainty assessment report data to obtain the integrated model of bridge steel cofferdam; According to the integrated model of bridge steel cofferdam, a dynamic mapping matrix output layer is added to obtain the physical monitoring digital twin mapping model.

2. The method according to claim 1, characterized in that S1 is specifically: S11. Use the Internet of Things monitoring equipment to collect the entire process of steel cofferdam construction in real time to obtain data on the entire process of steel cofferdam construction; S12, performing data cleaning on the steel cofferdam construction structured data of the entire process data of the steel cofferdam construction to obtain the steel cofferdam construction structured cleaned data; S13, performing structured conversion on the unstructured data of the steel cofferdam construction of the whole process data of the steel cofferdam construction to obtain structured conversion data of the steel cofferdam construction; S14, integrating the steel cofferdam construction structured cleaning data and the steel cofferdam construction structured conversion data to obtain the bridge steel cofferdam construction data.

3. The method according to claim 1, characterized in that S2 is specifically: S21. Obtaining bridge steel cofferdam design data; S22, constructing a bridge steel cofferdam basic model according to the bridge steel cofferdam design data to obtain a bridge steel cofferdam basic model; S23, structurally refining the bridge steel cofferdam foundation model to obtain a bridge steel cofferdam structural refinement model; S24, constructing a bridge steel cofferdam environmental model according to the environmental parameter data in the bridge steel cofferdam construction data to obtain a bridge steel cofferdam environmental model; S25. The detailed model of the bridge steel cofferdam structure and the environmental model of the bridge steel cofferdam are integrated to obtain a digital twin model of the bridge steel cofferdam.

4. The method according to claim 1, characterized in that S4 is specifically: S41, performing real-time monitoring according to the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data; S42, calculating the safety risk level according to the bridge steel cofferdam construction monitoring data using the preset safety risk assessment index data to obtain safety risk level data; S43, performing risk control operations according to the safety risk level data to obtain safety risk level change data; S44. Perform intelligent early warning processing of safety risks based on the safety risk level change data to obtain the risk data of bridge steel cofferdam construction so as to carry out intelligent monitoring of bridge steel cofferdam construction.

5. The method according to claim 1, characterized in that The adaptive mapping model is constructed as follows: Data enhancement is performed based on the characteristic time series analysis data of the bridge steel cofferdam to obtain characteristic time series enhanced data; Perform graph construction on the feature time series enhanced data to obtain feature time series graph data; Performing self-attention graph weighted calculation on the feature time series graph data to obtain feature time series graph weighted data; Perform graph convolution calculation on the weighted data of the feature time series graph to obtain the convolution layer data of the feature time series graph; Performing graph pooling layer processing on the feature time series graph convolution layer data to obtain feature time series graph pooling layer data; Perform full connection calculation on the feature time series pooling layer data to obtain the feature time series fully connected layer data; The feature time series fully connected layer data is iteratively trained to obtain an adaptive mapping model.

6. The method according to claim 5, characterized in that The weighted calculation of the self-attention graph is specifically: Performing self-attention hierarchy graph weighted calculation on the feature time series graph data to obtain first feature time series graph weighted data; Perform self-attention cross-graph weighted calculation on the feature time series graph data to obtain second feature time series graph weighted data; The first characteristic timing diagram weighted data and the second characteristic timing diagram weighted data are concatenated to obtain characteristic timing diagram weighted data; The weighted calculation of the self-attention hierarchy graph is as follows: Perform self-attention global graph weighted calculation on the feature time series graph data to obtain global weighted data of the feature time series graph; Extracting time series variation features according to the characteristic time series graph data to obtain characteristic time series graph variation feature data; Extract the change point according to the change feature data of the characteristic time series diagram to obtain the change point data; Constructing a correlation graph according to the change point data to obtain the change point correlation graph data; Generate local boundary parameters according to the change point association graph data to obtain local boundary parameter data; Perform self-attention local graph weighted calculation according to local boundary parameter data and feature time series graph data to obtain local weighted data of the feature time series graph; The global weighted data of the characteristic timing diagram and the local weighted data of the characteristic timing diagram are weightedly calculated to obtain the first characteristic timing diagram weighted data.

7. The method according to claim 6, characterized in that The weighted calculation of self-attention across graphs is as follows: Linearly mapping the characteristic timing diagram data to obtain characteristic timing diagram mapping data; Perform node similarity calculation on the feature time series graph mapping data to obtain node similarity data; The self-attention weight is calculated based on the node similarity data and the feature time series graph mapping data to obtain the cross-graph attention weight data; A cross-graph feature weighted calculation is performed on the feature time series graph data according to the cross-graph attention weight data to obtain second feature time series graph weighted data.

8. An intelligent monitoring system for bridge steel cofferdam construction based on digital twin, characterized in that: Used to execute the intelligent monitoring method for bridge steel cofferdam construction based on digital twin as claimed in claim 1, the intelligent monitoring system for bridge steel cofferdam construction based on digital twin comprises: Bridge steel cofferdam construction data acquisition module, used to obtain bridge steel cofferdam construction data; A full-factor digital twin model creation module is used to create a full-factor digital twin model based on the bridge steel cofferdam construction data to obtain a bridge steel cofferdam digital twin model; A real-time bridge steel cofferdam data fusion mapping model is used to obtain real-time bridge steel cofferdam data; perform planning processing based on the real-time bridge steel cofferdam data to obtain planned data for the bridge steel cofferdam; perform deep multi-dimensional feature extraction on the planned data for the bridge steel cofferdam to obtain characteristic data for the bridge steel cofferdam; generate mapping data based on the characteristic data of the bridge steel cofferdam through a preset physical monitoring digital twin mapping model to obtain mapping matrix data; perform fusion mapping based on the real-time bridge steel cofferdam data and the digital twin model of the bridge steel cofferdam through mapping matrix data to obtain a dynamic model of the bridge steel cofferdam; the specific steps for constructing the preset physical monitoring digital twin mapping model are as follows: obtain historical bridge steel cofferdam data; perform multi-source data fusion based on the historical bridge steel cofferdam data to obtain multi-source fusion data of bridge steel cofferdam data; The multi-source fusion data of steel cofferdam data is used to extract deep features to obtain the deep feature data of bridge steel cofferdam; the time series analysis of the deep feature data of bridge steel cofferdam is performed to obtain the characteristic time series analysis data of bridge steel cofferdam; the adaptive mapping model is constructed based on the characteristic time series analysis data of bridge steel cofferdam to obtain the adaptive mapping model of bridge steel cofferdam; multi-task joint learning is performed on the adaptive mapping model of bridge steel cofferdam to obtain the multi-task learning model of bridge steel cofferdam; uncertainty assessment is performed on the multi-task learning model of bridge steel cofferdam to obtain uncertainty assessment report data; model integration is performed on the multi-task learning model of bridge steel cofferdam based on the uncertainty assessment report data to obtain the integrated model of bridge steel cofferdam; dynamic mapping matrix output layer is added based on the integrated model of bridge steel cofferdam to obtain the physical monitoring digital twin mapping model; The intelligent monitoring module for bridge steel cofferdam construction is used to perform real-time monitoring based on the dynamic model of the bridge steel cofferdam to obtain the bridge steel cofferdam construction monitoring data, and to perform dynamic risk assessment based on the bridge steel cofferdam construction monitoring data to obtain the bridge steel cofferdam construction risk data for intelligent monitoring of the bridge steel cofferdam construction.

Images