Water conservancy project construction safety anomaly monitoring method and system based on digital twinning

Abstract

This invention discloses a method and system for monitoring construction safety anomalies in water conservancy projects based on digital twins, relating to the field of water conservancy engineering technology. The method includes: simulating geographic engineering data, environmental equipment data, and personnel data at the construction site; constructing and optimizing a digital twin model using a twin difference network to achieve dynamic tracking and simulation model display; using the digital twin model to perform data optimization perception and processing on the geographic engineering data, environmental equipment data, and personnel data respectively, obtaining first, second, and third feature data; optimizing and fusing the first, second, and third feature data through the digital twin model to obtain a comprehensive feature value; calculating optimization coefficients; performing dynamic simulation analysis using a multi-time-series method based on the construction stage and adjusting the time window; determining the level using the optimization coefficients; automatically adjusting the construction project maintenance progress according to the level and performing real-time prediction; and sending the results to a visual interface for construction personnel.

Description

Technical Field

[0001] This invention relates to the field of water conservancy engineering technology, specifically to a method and system for monitoring anomalies in water conservancy engineering construction safety based on digital twins. Background Technology

[0002] In the field of water conservancy engineering, with the continuous advancement of technology, monitoring models and methods are constantly being optimized and verified, but some problems still exist.

[0003] Current construction site monitoring primarily relies on building digital twin models of the construction site to simulate facilities and the environment, dynamically track and display data in three dimensions, and optimize resource management through simulation. However, using a single type of digital twin model to monitor data often fails to yield diverse data analysis results. Combining multi-time-series analysis methods with digital twin models for data verification allows for accurate and efficient comprehensive analysis and prediction. The predicted data then provides timely solutions, enabling proactive and effective preventative measures.

[0004] To address this, a method and system for monitoring anomalies in water conservancy engineering construction safety based on digital twins are proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for monitoring safety anomalies in water conservancy engineering construction based on digital twins, used for monitoring and handling safety anomalies at water conservancy engineering construction sites. The specific implementation steps include: simulating geographic engineering data, environmental equipment data, and personnel data at the construction site; constructing and optimizing a digital twin model using a twin difference network to achieve dynamic tracking and simulation model display; secondly, using the digital twin model to optimize and process the geographic engineering data, environmental equipment data, and personnel data respectively, obtaining first, second, and third feature data; optimizing and fusing the first, second, and third feature data through the digital twin model to obtain a comprehensive feature value, and using a multi-time series method for analysis and optimization to obtain optimization coefficients; determining the level using the optimization coefficients; automatically adjusting the maintenance progress of the construction project according to the level, making real-time predictions, and sending the results to a visual interface for construction personnel. This invention can accurately collect data from water conservancy engineering construction sites, achieve real-time monitoring, and promptly propose response methods, thereby improving the safety and monitoring efficiency of water conservancy engineering construction.

[0006] A digital twin-based method for monitoring anomalies in water conservancy engineering construction safety includes: simulating geographic engineering data, environmental equipment data, and personnel data at the construction site; constructing and optimizing a digital twin model using a twin difference network to achieve dynamic tracking and simulation model display; using the digital twin model to optimize and process the geographic engineering data, environmental equipment data, and personnel data respectively to obtain first, second, and third feature data; optimizing and fusing the first, second, and third feature data through the digital twin model to obtain a comprehensive feature value, and using a multi-time series method for analysis and optimization to obtain optimization coefficients; using the optimization coefficients to determine the level; automatically adjusting the maintenance schedule of the construction project according to the level, making real-time predictions, and sending the results to the construction personnel's visualization interface.

[0007] Preferably, the collection of the first geographic engineering data, the second environmental equipment data, and the third personnel data includes:

[0008] Data imported using a geographic information system and engineering design data are set as geographic engineering data; sensors installed at the construction site to monitor the environment and operating equipment parameters are set as environmental equipment data; smart monitoring wristbands worn by construction workers monitor their location information, movement status, heart rate, and blood pressure values ​​as personnel data; all data are recorded in the historical database of the digital twin model.

[0009] Preferably, the process of obtaining the first feature in the step includes:

[0010] The geographic engineering data is compared and verified with standard values ​​using a digital twin model. Normal ranges and thresholds are set, and the actual deviations are compared to determine whether the deviations exceed the thresholds. If the actual values ​​exceed the set thresholds, they are marked as abnormal. The geographic engineering data is analyzed through the model, and volume, size, and location are recorded as feature types. The value of each feature type is a feature value. The deviations of each feature value are weighted and summed according to certain weights and set as the first feature.

[0011] Preferably, the second feature implementation process in step two includes:

[0012] The environmental equipment data is compared and verified with standard values ​​using a digital twin model. Normal ranges and thresholds are set, and the actual values ​​are compared to determine whether they exceed the thresholds. If they exceed the set thresholds, they are marked as abnormal. The environmental equipment data is analyzed through the model, and the flow rate and equipment pressure values ​​are recorded as feature types, and the value of the feature type is the feature value. The feature values ​​and deviations are weighted and summed according to a certain weight and set as the second feature.

[0013] Preferably, the process of obtaining the third feature in the step includes:

[0014] The personnel data is compared and verified with the standard values ​​using a digital twin model to determine whether the values ​​exceed the threshold. If the actual value exceeds the set threshold, it is marked as abnormal. The personnel data is analyzed by the model, and the location information, status, heart rate, and blood pressure of the construction personnel are recorded as feature types, and the value of the feature type is the feature value. The deviation of each feature value is weighted and summed according to a certain weight and set as the third feature.

[0015] Preferably, the process of obtaining the comprehensive feature value includes:

[0016] Based on the first feature, the second feature, and the third feature, a comprehensive feature value is obtained by multiplying the feature value by the feature value variation according to the weights, and then input into the digital twin model.

[0017] Preferably, the process of obtaining the optimization coefficients includes:

[0018] Based on the comprehensive feature value, analyze its value in different time periods, calculate the average value of the comprehensive feature value in several time periods, record it as the optimization coefficient, and input it into the digital twin model.

[0019] Preferably, the visualization interface includes:

[0020] Based on the optimization coefficients obtained from the analysis, the digital twin model sends the analysis and optimization prediction results to the smart bracelet. The bracelet then initiates vibrations of varying degrees, and its small LCD screen brightens to display the specific risk content, risk type, and recommended measures.

[0021] Preferably, the water conservancy project construction safety anomaly monitoring system based on digital twin includes:

[0022] Data extraction module: used to collect geographic engineering data, environmental equipment data, and personnel data at the construction site;

[0023] Digital Twin Model Module: This module integrates and optimizes data to build a digital twin model; it updates the digital twin model in real time, extracts data for multi-dimensional analysis, identifies any abnormal changes, and predicts future development trends in real time.

[0024] Feature fusion module: The first feature, second feature and third feature are fused using a weighted summation method to form a comprehensive feature value and obtain the optimization coefficient;

[0025] Automatic optimization decision support module: Optimizes construction plans and resource scheduling based on optimization coefficients, and provides real-time notifications;

[0026] Visual interface module: Sends the obtained results to the construction workers' wristbands.

[0027] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0028] The comprehensive feature value proposed in this invention, through a weighted summation method that integrates multiple data features, not only covers traditional geographical and engineering elements but also incorporates dynamic environmental factors, equipment operating status, and real-time personnel conditions, providing a richer and more comprehensive data foundation for construction safety monitoring. Compared with traditional monitoring methods based on a single data source, its innovation lies in the deep integration of multi-source heterogeneous data, breaking down data silos and thus more accurately reflecting the complex situation at the construction site.

[0029] The optimization coefficient proposed in this invention achieves a quantitative assessment of construction safety risks by analyzing the comprehensive characteristic values ​​at different time periods. Unlike previous qualitative or simple quantitative assessment methods, this method can dynamically adjust the risk level based on real-time data, providing construction safety management with an intuitive, accurate, and dynamic risk quantification tool. This reduces safety risks and resource waste caused by unreasonable decisions. This innovation shifts construction safety management from experience-based judgment to data-driven scientific decision-making, improving the timeliness and accuracy of risk assessment.

[0030] 3. The visualization interface module proposed in this invention displays the assessment report through a visual interface. Construction personnel can determine different risk levels by observing the vibration frequency of the wristband, achieving efficient transmission and communication of risk information. Its innovation lies in presenting complex assessment results to construction personnel in an easy-to-understand way, ensuring effective information delivery. This visualization interface also supports interactive functions, allowing users to easily understand the risk situation of specific areas or equipment, providing strong support for on-site decision-making. Attached Figure Description

[0031] Figure 1 This is a flowchart of the water conservancy project construction safety anomaly monitoring method based on digital twin proposed in an embodiment of this invention application;

[0032] Figure 2 This is a flowchart illustrating the process of obtaining comprehensive feature values ​​and optimization coefficients as proposed in an embodiment of this invention.

[0033] Figure 3 This is a schematic diagram of the structure of a water conservancy engineering construction safety anomaly monitoring system based on digital twins, as proposed in an embodiment of this invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] During the construction of water conservancy projects, the construction site environment is complex and there are various safety risks. If there are problems such as incomplete monitoring data, poor real-time performance, and untimely early warning, it will be difficult to meet the safety requirements of modern water conservancy construction. Therefore, developing a method and system that can monitor water conservancy construction safety anomalies in real time, comprehensively and intelligently is of great practical significance.

[0036] This invention relates to the field of water conservancy engineering technology, specifically to a method and system for monitoring anomalies in water conservancy engineering construction safety based on digital twins. It is applicable to water conservancy engineering construction scenarios such as dam construction and river regulation, improving the safety and monitoring efficiency of water conservancy engineering construction. To illustrate the effectiveness of the method and system of this invention, detailed descriptions will be provided in conjunction with the accompanying drawings and the following embodiments. Example 1

[0037] This application discloses a method for monitoring construction safety anomalies in water conservancy projects based on digital twins. The method detects construction safety anomalies during the construction of a large-scale dam water conservancy project (see reference). Figure 1 The flowchart of the method for monitoring anomalies in water conservancy engineering construction safety based on digital twins is as follows: It simulates geographic engineering data, environmental equipment data, and personnel data at the construction site; a digital twin model is constructed and optimized using a twin difference network to achieve dynamic tracking and simulation model display; the digital twin model is used to optimize and process the geographic engineering data, environmental equipment data, and personnel data respectively to obtain first, second, and third feature data; the first, second, and third feature data are optimized and fused through the digital twin model to obtain a comprehensive feature value, and an optimization coefficient is obtained by multi-time series analysis; the optimization coefficient is used to determine the level; the maintenance schedule of the construction project is automatically adjusted according to the level, and real-time prediction is performed, with the results sent to the construction personnel's visualization interface.

[0038] Furthermore, geographic engineering data, environmental equipment data, and personnel data from the water conservancy project construction site are collected and input into the digital twin model. Corresponding to the above steps, the specific implementation includes:

[0039] Specifically, in the initial stage of construction, data is imported into the Geographic Information System (GIS) and combined with satellite remote sensing technology to obtain large-scale and complex geographic information, acquiring high-resolution on-site images. After processing, topographic data is obtained and recorded as geoengineering data. Topographic maps and engineering design data of the construction site are also acquired. Sensors for monitoring environmental parameters are installed at the construction site, and sensors for monitoring the operation of construction equipment are installed on the equipment. Data is collected in real time through a sensor network and recorded as environmental equipment data. After processing the collected geoengineering and environmental equipment data, digital twin models are used to achieve dynamic tracking and simulation models to display complex terrain. Augmented reality (AR) technology is used to construct multi-scene sandboxes to assist construction personnel in positioning, intuitively display construction points that are difficult to visually detect, simulate operation procedures and safety regulations under different scenarios, and demonstrate the effectiveness of different emergency plans. The location information, movement status, heart rate, and blood pressure values ​​displayed on the wristbands worn by each construction worker are recorded as personnel data. A twin differential network algorithm spatial positioning and data fusion module is used to perform coordinate transformation, data format unification, missing value imputation, and digitization processing on the collected geoengineering data, converting engineering design drawings and documents into a computable format. Time series data of different devices and monitoring parameters of environmental equipment are aligned, comparing data differences and trends at adjacent time points, and automatically adjusting the starting point and sampling frequency parameters of each time series to keep environmental equipment data synchronized in the time dimension, filling in missing data and recording the normal operation and fault status of equipment. Personnel data is located and recorded to form a complete dataset of personnel activity trajectories and statuses.

[0040] This application's embodiments acquire data through various means, including high-precision measuring equipment, sensors, and monitoring wristbands. These devices can acquire real-time data on geospatial engineering, environmental equipment parameters, and personnel status. This real-time data acquisition capability enables the system to promptly reflect the latest situation at the construction site, providing dynamic and real-time data support for construction safety monitoring. It avoids errors and omissions that may arise from a single data source, providing a rich and comprehensive information foundation for subsequent integrated analysis and risk assessment, ensuring the accuracy and reliability of the data.

[0041] Furthermore, the geographic engineering data, environmental equipment data, and personnel data are analyzed separately to obtain the first feature, the second feature, and the third feature. The specific implementation of these steps includes:

[0042] Based on the geoengineering data, it is compared with standard engineering design data. According to engineering design specifications and construction requirements, normal ranges and thresholds are set. The deviation between the actual building dimensions and the design dimensions is compared to determine if the deviation exceeds the threshold. If the actual value exceeds the set threshold, it is marked as abnormal; if the actual value does not exceed the set threshold, it is marked as normal. Through analysis of the geoengineering data, the building's volume, size, and location are recorded as feature types; feature values ​​are the values ​​of the feature types. The volume weight is set to 0.3, the size weight to 0.3, and the location weight to 0.4. The weighted sum of all feature values ​​is set as the first feature 'a'.

[0043] This application's embodiments, by comparing geoengineering data with standard engineering design data, can accurately extract key volume, dimension, and location parameters during construction. This precise feature extraction method helps to promptly identify potential structural deviations during construction, and by setting normal ranges and thresholds, and comparing deviations in real time, the system can immediately mark an anomaly once the actual value exceeds the set threshold. This achieves real-time monitoring and timely early warning of the construction process. Furthermore, by weighted summing of the deviations of each feature value to obtain the first feature, multi-dimensional feature analysis is achieved. This quantitative evaluation method makes the measurement of anomaly severity more scientific and accurate, providing a more detailed decision-making basis for construction safety management.

[0044] Based on the environmental equipment data, it is compared with standard values. According to engineering design specifications and construction requirements, normal ranges and thresholds are set. The actual water level is compared to determine if it exceeds the threshold, or if the flow rate exceeds the threshold, or if the pressure exceeds the threshold. If the actual value exceeds the set threshold, it is marked as abnormal. Because the environmental equipment data changes rapidly, when the actual value exceeds the set threshold, an adaptive data acquisition strategy is adopted to automatically initiate emergency drone inspections, increasing the frequency of water level and flow rate monitoring. Key monitoring data of the environmental equipment is stored as evidence. Environmental data includes timestamps, location, and sensor IDs; equipment data includes timestamps, location, equipment ID, and operating status. This adaptive risk handling method facilitates more timely marking of anomalies and timely acquisition of information on sudden anomalies. By analyzing environmental equipment data, the current water level, flow velocity, and pressure value are recorded as feature types, and the feature value is the value of the feature type. Through analysis, the weight of the current water level is set to 0.3, the weight of the flow velocity is set to 0.3, the weight of the pressure value is set to 0.2, the weight of the state value is set to 0.2, and the weighted sum of each feature value is set as the second feature b.

[0045] This application embodiment, through in-depth analysis of environmental equipment data, accurately extracts key environmental parameters such as water level, flow rate, and pressure, enabling timely detection of minor anomalies in environmental changes. By comparing the monitored environmental equipment data with standard values ​​in real time, it can immediately determine whether parameters such as actual water level exceed set thresholds. Once the threshold is exceeded, the system will mark the anomaly and issue an early warning in real time. Furthermore, the deviations of each feature value are weighted and summed to obtain a second feature, achieving multi-dimensional analysis. This allows the system to more comprehensively identify anomalies in environmental equipment data, avoiding misjudgments or omissions caused by focusing on only a single feature, and enhancing the ability to identify construction safety risks.

[0046] Based on the personnel data, existing parameters are compared with normal vital signs of construction workers. Normal ranges and thresholds for each body posture value are set according to individual physical conditions. The actual value is compared to the individual's actual situation to determine if it exceeds the threshold. If the actual value exceeds the set threshold, it is marked as abnormal. When analyzing personnel data, UWB precise positioning and posture recognition technology are introduced. An LSTM network is used to analyze abnormal patterns in personnel trajectories to determine if there is prolonged lingering in high-risk areas. A dynamic safety density threshold is defined. Based on personnel location clustering results and equipment movement trajectories, collision probability is calculated in real time, triggering spatial partitioning warnings. This provides a more accurate display of the construction workers' status and effectively monitors the occurrence of safety accidents. By analyzing third-party personnel data, the location information, movement status, heart rate, and blood pressure values ​​of construction workers are compared and recorded as feature types. The feature value is the feature type. The weight of location information is set to 0.2, the weight of movement status is set to 0.2, the weight of heart rate is set to 0.3, and the weight of blood pressure value is set to 0.3. The weighted sum of all feature values ​​is set as the third feature c.

[0047] This application embodiment comprehensively understands the real-time status of construction workers by monitoring their location, movement status, heart rate, blood pressure, and other multi-dimensional data. Personalized risk assessments are performed based on the specific data of each worker, making the risk assessments more accurate and targeted. Furthermore, the third feature enables the system to comprehensively monitor the situation of construction workers, ensuring their safety.

[0048] Furthermore, the first, second, and third features are fused to obtain a comprehensive feature value, and the comprehensive feature value is analyzed to obtain a risk assessment coefficient. The risk level is determined using the risk assessment coefficient, and an assessment report is generated. The specific implementation of the above steps includes:

[0049] Based on the first, second, and third features, and with weight ratios of 0.3, 0.4, and 0.3 respectively, the sum of the products of the eigenvalue variation and the weights is calculated to obtain the comprehensive eigenvalue. The calculation formula is expressed as:

[0050]

[0051] in, , and Let ω1, ω2, and ω3 be the first feature, the second feature, and the third feature, respectively, and let ω1, ω2, and ω3 be the weights of the first feature, the second feature, and the third feature, respectively. .

[0052] In this embodiment, the amplitudes of the first, second, and third features are multiplied and summed according to their weights to form a comprehensive feature value. This provides a unified quantitative standard to comprehensively reflect the overall risk status of the construction site. This makes subsequent optimization coefficient calculations, risk level determinations, and construction plan adjustments more consistent and comparable, facilitating analysis and decision-making by management personnel. In different water conservancy construction sites, the weights of each feature can be flexibly adjusted according to the specific characteristics and needs of the project.

[0053] Furthermore, based on the comprehensive feature value Analyze its values ​​over different time periods, and take the average of the comprehensive characteristic values ​​over several time periods, which is denoted as the optimization coefficient. For the acquisition process, please refer to [link / reference]. Figure 2 The flowchart for obtaining the comprehensive eigenvalues ​​and optimization coefficients is expressed by the following formula:

[0054] ;

[0055] in, To optimize the coefficients, The comprehensive feature value is represented by n, which is the number of data points within the time window. Based on the standardized optimization, the risk level classification criteria are set as follows:

[0056] Low risk: 0≤ <0.3;

[0057] Medium risk: 0.3≤ <0.7;

[0058] High risk: 0.7≤ ≤1;

[0059] The calculated standardized optimization coefficients are compared with the set thresholds to determine the current risk level. An assessment report is then generated, which includes the risk level, main risk factors and their sources, the location of the risk, and optimization coefficient information.

[0060] This application's embodiments calculate the average of comprehensive characteristic values ​​across multiple time periods, reducing the impact of data anomalies at a single time point on the evaluation results. This comprehensive analysis method using multi-time period data improves the reliability of the optimization coefficients, avoids misjudgments caused by short-term data fluctuations, and enhances the foresight of construction safety management.

[0061] The time window for adjusting the monitoring optimization coefficient is adjusted according to the risk level. If the optimization coefficient is less than the low-risk threshold, the time window is adjusted to 1.5-2 times the default length. If the optimization coefficient is between the low-risk and high-risk thresholds, the time window length remains unchanged. If the optimization coefficient is greater than the high-risk threshold, the time window is adjusted to 0.5-0.7 times the default length. Each time the time window length is adjusted, it does not jump directly to the target length, but gradually approaches the target length in a certain step size to reduce resource consumption, optimize monitoring efficiency, and avoid excessive process occupation.

[0062] This application's embodiments calculate optimization coefficients and dynamically adjust the time window length according to the risk level of the construction stage. The time window is shortened during high-risk periods and extended during low-risk periods. The trend of window data is predicted through a model. The weight of the comprehensive feature value is dynamically adjusted according to changes in weather and terrain at the construction site, thus upgrading the multi-time series analysis method from a general algorithm to a technical solution with professional barriers in water conservancy engineering.

[0063] Furthermore, the construction and maintenance schedule is dynamically adjusted based on the risk level and sent to a visual interface. The specific implementation of the above steps includes:

[0064] Construction workers are equipped with smart wristbands that allow them to view the latest safety status and risk information related to their work area at any time. Received information is highlighted and can directly display text, numbers, and icons. Information is transmitted via vibration at different levels based on risk level: long vibrations indicate high risk, continuous short vibrations indicate medium risk, and a single short vibration indicates low risk. Based on the risk level results, the construction sequence is adjusted, prioritizing high-risk areas. Using a twin differential network algorithm, the difference between current equipment operating data and normal operating data is extracted based on the risk assessment coefficient, and the data is processed in real time to modify construction methods. Workers are assigned to high-risk areas according to the construction schedule, and safety inspections of the construction site are strengthened. Equipment allocation and material supply plans are optimized, and the supply of waterproofing materials is increased, with emergency supplies stockpiled in advance.

[0065] This application embodiment uses a construction worker wristband to promptly issue warnings to construction workers through vibration and high-brightness display, ensuring that construction workers can immediately notice abnormal situations and take corresponding countermeasures. This enables them to take correct actions quickly in emergency situations, reducing confusion and blind responses, and improving the efficiency of emergency response.

[0066] This application embodiment utilizes a digital twin-based method for monitoring construction safety anomalies in water conservancy projects, achieving the monitoring of construction safety anomalies in digital twin water conservancy projects. The specific process mainly includes the following steps: First, simulating the geographical engineering data, environmental equipment data, and personnel data of the construction site, constructing and optimizing a digital twin model using a twin difference network to achieve dynamic tracking and simulation model display; then, using the digital twin model to perform data optimization perception and processing on the geographical engineering data, environmental equipment data, and personnel data respectively, obtaining first, second, and third feature data; finally, optimizing and fusing the first, second, and third feature data through the digital twin model to obtain a comprehensive feature value, and using a multi-time series method for analysis and optimization to obtain optimization coefficients; finally, using the optimization coefficients to determine the level; and then automatically adjusting the maintenance progress of the construction project according to the level, performing real-time prediction, and sending the results to the construction personnel's visual interface. This invention proposes a comprehensive feature value, a risk assessment coefficient, and a visualization interface module for the above process; the comprehensive feature value provides a more comprehensive data foundation for construction safety monitoring; the risk assessment coefficient provides an intuitive and dynamic risk quantification tool, improving the timeliness and accuracy of risk assessment; the visualization interface module presents the assessment results to construction personnel in an easy-to-understand way, ensuring the effective communication of information.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents. Example 2

[0068] In Example 1, the method of the present invention was used to monitor construction safety anomalies in a large-scale dam water conservancy project. In this embodiment, the digital twin-based water conservancy project construction safety system proposed in this invention is applied to the monitoring of construction safety anomalies in a Type B dam water conservancy project; see reference... Figure 3 A schematic diagram of the structure of a water conservancy project construction safety anomaly monitoring system based on digital twins. The water conservancy project construction safety system based on digital twins includes: a data extraction module, a digital twin model module, a feature fusion module, an automatic optimization decision support module, and a visualization interface module.

[0069] Furthermore, the data acquisition module employs lightweight methods, combining small-scale UAV mapping and digitization of design drawings to acquire primary geographic engineering data; it utilizes low-cost but high-precision sensors to monitor water levels, flow rates, and the status of small equipment to acquire environmental equipment data. Construction workers wear wristbands, providing basic personal information and simple activity data, thereby acquiring personnel data.

[0070] Furthermore, the digital twin model module integrates the collected data to construct a digital twin model; it updates the digital twin model in real time, extracts data for analysis, and determines whether there are any abnormal changes.

[0071] Furthermore, the feature fusion module fuses the first feature a, the second feature b, and the third feature c using a weighted summation method to form a comprehensive feature. Analyze and synthesize characteristics The optimization coefficients are obtained, and the specific implementation process includes:

[0072] Based on the obtained first feature a, second feature b, and third feature c, and with weight ratios of 0.4, 0.4, and 0.2 respectively, the sum of the product of the eigenvalue variation and the weights is calculated to obtain the comprehensive feature. Its calculation formula is expressed as:

[0073]

[0074] in, For comprehensive eigenvalues, , and Let ω1, ω2, and ω3 be the first feature, the second feature, and the third feature, respectively, and let ω1, ω2, and ω3 be the weights of the first feature, the second feature, and the third feature, respectively. .

[0075] Furthermore, based on the comprehensive feature value, its values ​​at different time periods are analyzed, and the average value of the comprehensive feature value over one minute is taken as the optimization coefficient. For a flowchart of how the optimization coefficient is obtained, please refer to [link to flowchart]. Figure 2 Its calculation formula is expressed as:

[0076] ;

[0077] in, To optimize the coefficients, The comprehensive feature value is n, which is the number of data points within the time window.

[0078] Furthermore, based on the calculated optimization coefficients, risk level classification criteria are set to determine the risk level:

[0079] Low risk: 0≤ <0.4;

[0080] Medium risk: 0.4≤ <0.8;

[0081] High risk: 0.8≤ ≤1;

[0082] Furthermore, based on the risk level results, the construction sequence is adjusted, prioritizing construction tasks in high-risk areas. The Siamese differential network algorithm in deep learning is used to adjust the time window for monitoring and optimizing the risk level. If the optimization coefficient is less than the low-risk threshold, the time window is adjusted to 1.5-2 times the default length; if the optimization coefficient is between the low-risk and high-risk thresholds, the time window length remains unchanged; if the optimization coefficient is greater than the high-risk threshold, the time window is adjusted to 0.5-0.7 times the default length. Each time the time window length is adjusted, it does not jump directly to the target length, but gradually approaches the target length in steps to reduce resource consumption, optimize monitoring efficiency, avoid excessive process occupancy, extract the differences between the current construction equipment operating data and the data during normal operation, process the data in real time, and modify the construction method accordingly. Based on the construction schedule requirements, corresponding workers are assigned to high-risk areas for construction, and safety inspections of the construction site are strengthened. Equipment allocation and material supply plans are optimized, and the supply of waterproofing materials is increased, with emergency supplies stockpiled in advance.

[0083] Furthermore, the visualization interface displays information and notifies construction personnel through highlighting and different vibration frequencies, enabling them to obtain information in real time.

[0084] The construction workers' wristbands transmit information through vibrations of varying degrees based on the risk level. Continuous long vibrations indicate high risk, continuous short vibrations indicate medium risk, and a single short vibration indicates low risk, as shown in Table 1.

[0085] Table 1. Calculation of Risk Assessment Coefficients and Level Classification, and Visualization Module Display

[0086]

[0087] This application embodiment utilizes a digital twin-based water conservancy engineering construction safety anomaly monitoring system. This system integrates data acquisition, digital twin, feature fusion, automatic optimization decision support, and visualization interface modules for modular data management. It enables real-time model updates, conducts accurate risk assessments of construction projects, provides effective decision support, and offers intuitive visualization.

[0088] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring construction safety anomalies in water conservancy projects based on digital twins, characterized in that, Includes the following steps: The system simulates geographical engineering data, environmental equipment data, and personnel data from a construction site, and uses twin differential networks to build and optimize a digital twin model, enabling dynamic tracking and simulation model display. Digital twin models are used to optimize and process geographic engineering data, environmental equipment data, and personnel data to obtain first, second, and third feature data. The first feature is obtained by comparing and verifying the geographic engineering data with standard values ​​using digital twin models, and volume, size, and location are recorded as feature types. The value of the feature type is the feature value. The deviation of each feature value is weighted and summed to set as the first feature. The second feature uses a digital twin model to compare and verify the environmental equipment data with standard values. Flow rate and equipment pressure are recorded as feature types, and the value of the feature type is the feature value. The deviations of each feature value are weighted and summed to set as the second feature. The third feature uses a digital twin model to compare and verify the personnel data with standard values. The location information, movement status, heart rate, and blood pressure of the construction personnel are recorded as feature types, and the value of the feature type is the feature value. The deviation of each feature value is weighted and summed to set as the third feature. The data of the first, second, and third features are optimized and fused using a digital twin model. The feature values ​​of each feature are multiplied and summed according to their weights to obtain the comprehensive feature value. The specific formula is as follows: ; in, , and These are the first feature, the second feature, and the third feature, respectively. , and The weights of the first feature, the second feature, and the third feature are respectively. + + =1; Based on the comprehensive feature values, the average value of the comprehensive feature values ​​over several time periods is calculated to obtain the optimization coefficient. The specific formula is as follows: ; in, To optimize the coefficients, For the comprehensive feature value, n is the number of data points within the time window; based on the construction stage, a multi-time series method is used for dynamic scenario simulation analysis, and the time window length of the optimization coefficient is dynamically adjusted according to the environmental and equipment load parameters of the construction stage, shortening the time window during high-risk periods and extending the time window during low-risk periods; The optimization coefficients are used to determine the level; the maintenance schedule of the construction project is automatically adjusted according to the level, and real-time prediction is made, with the results sent to the construction personnel's visual interface.

2. The method for monitoring construction safety anomalies in water conservancy projects based on digital twins according to claim 1, characterized in that, The geographic engineering data, environmental equipment data, and personnel data include: Data imported using a geographic information system and engineering design data are set as geographic engineering data; sensors installed at the construction site to monitor the environment and operating equipment parameters are set as environmental equipment data; smart monitoring wristbands worn by construction workers monitor their location information, movement status, heart rate, and blood pressure values ​​as personnel data; all data are recorded in the historical database of the digital twin model.

3. The method for monitoring construction safety anomalies in water conservancy projects based on digital twins according to claim 1, characterized in that, The process of sending the results to the construction personnel's visual interface includes: Based on the optimization coefficients obtained from the analysis, the digital twin model sends the analysis and optimization prediction results to the smart bracelet. The bracelet then vibrates to varying degrees, and its small LCD screen brightens to display the specific simulation optimization results.

Images