A method, medium and system for constructing a digital twin system for nearshore overtopping disasters

Abstract

The present invention provides a method, medium and system for constructing a digital twin system for nearshore overtopping disasters, belonging to the field of image data processing or generation technology, including: first, using a drone to obtain multi-angle photos and elevation data of a target area, and building a three-dimensional real-scene model based on this. Then, using a real-time three-dimensional creation tool to simulate waves and hydrodynamics, while conducting real-time observations in the target area and inputting the data into the system. Next, a wave model and a storm surge model are constructed based on a high-resolution numerical model or an artificial intelligence model, and these models are used to forecast waves and storm surges for no less than 72 hours, while simulating historical and different recurrence period disaster scenarios and different disaster reduction measures scenarios. Finally, a disaster prevention and reduction scenario library is constructed based on the simulation results, which solves the problem that the existing technology cannot fully reflect the three-dimensional dynamic characteristics of the sea area and the effects of various disaster reduction measures.

Description

Technical Field

[0001] The present invention belongs to the technical field of image data processing or generation, and in particular, relates to a method, medium and system for constructing a digital twin system for nearshore overtopping disasters. Background Art

[0002] With the rapid development of social economy, the population density and infrastructure construction in coastal areas are increasing, and the need to combat natural disasters is becoming increasingly urgent. Marine disasters such as storm surges and coastal waves occur from time to time, causing serious casualties and property losses in coastal areas. Such disasters are not only highly destructive, but also have the characteristics of short occurrence time and difficulty in early warning, which brings huge challenges to disaster prevention and mitigation work.

[0003] At present, in the field of nearshore marine disaster prevention and control, traditional two-dimensional GIS data is used as the basis for modeling, which cannot fully reflect the three-dimensional dynamic characteristics of the sea area and the effects of various disaster reduction measures, making it difficult to better respond to and predict the situation of nearshore marine disasters. Summary of the invention

[0004] In view of this, the present invention provides a method, medium and system for constructing a digital twin system for nearshore overtopping disasters, which can solve the technical problems that the existing technology often uses traditional two-dimensional GIS data as the basis for modeling, cannot fully reflect the three-dimensional dynamic characteristics of the sea area and the effects of various disaster reduction measures, and is difficult to better respond to and predict the situation of nearshore marine disasters.

[0005] The present invention is achieved in that:

[0006] A first aspect of the present invention provides a method for constructing a nearshore overbank disaster digital twin system, comprising the following steps:

[0007] S10. Obtain multi-angle photos and elevation data of the target area, specifically, use a drone with oblique photography and elevation measurement functions to conduct aerial survey to obtain multi-angle photos and elevation data of the target area; the implementation can be as follows: select a drone platform with oblique photography and elevation measurement functions, such as DJI Phantom 4RTK or Wingtra One, and install a high-precision camera sensor and RTK-GNSS positioning system; determine the appropriate route and altitude according to the actual situation of the target area, generally select a flight altitude of 50-100 meters to obtain clear oblique photography images and high-precision terrain data; accurately position and calibrate the drone to ensure the measurement accuracy of its position and attitude data; carry out autonomous or remote-controlled aerial survey tasks according to the established route, and record the images and terrain data of each voyage; perform quality inspection and preprocessing on the obtained raw data to eliminate outliers and noise.

[0008] S20. Constructing a three-dimensional real scene model based on the multi-angle photos and elevation data, including image preprocessing, feature extraction and matching, aerial triangulation, dense matching and point cloud generation, three-dimensional reconstruction, texture mapping and model optimization; the implementation can be as follows: performing preprocessing such as dedistortion and color correction on the original photos to improve image quality; using algorithms such as scale-invariant feature transform (SIFT) and speeded up robust features (SURF) to extract and match key feature points from adjacent photos; based on the feature point matching results, using a bundle adjustment algorithm to estimate the pose parameters of each photo to obtain the camera's exterior orientation elements; using semi-global matching (SGM), deep learning and other methods to generate dense three-dimensional point cloud data; using point cloud data to construct a triangular mesh model, and performing geometric refinement and topological optimization; attaching the original photos to the surface of the three-dimensional model to obtain a real texture effect; applying filtering, simplification and other technologies to improve the geometric accuracy and rendering effect of the three-dimensional model.

[0009] S30. Use real-time 3D creation tools to simulate waves and hydrodynamics, where JONSWAP spectrum is used for wave simulation and simplified continuity equation and momentum equation are used for hydrodynamics simulation. This can be implemented as follows: import the constructed 3D real-scene model into real-time 3D creation tools such as Unity or UnrealEngine, set the basic properties of the scene such as terrain, material, etc.; select the wave simulation algorithm based on JONSWAP spectrum, and set the effective wave height of the wave according to the measured data or forecast results. , spectral peak period and peak elevation factor etc. parameters;

[0010] Using simplified forms of the continuity and momentum equations and , , simulate the hydrodynamic process; through real-time rendering technology, visualize the dynamic changes of waves and hydrodynamics.

[0011] S40. Carry out real-time observation in the digital twin area or nearby areas, and use buoys or bottom-mounted ADCP observation instruments to obtain real-time observation data, including wave height and water level. This can be implemented as follows: deploy buoys and bottom-mounted ADCP and other observation instruments in the target sea area to ensure their stability and data transmission reliability; use high-frequency (1-2Hz) sampling to record time series data such as water level, flow velocity, and flow direction; perform quality control on the original observation data, eliminate outliers and noise, and make necessary corrections and interpolations; transmit the processed measured data to the digital twin system in real time, and associate it with the simulation scene in real time.

[0012] S50. The real-time observation data is transmitted to the digital twin system so that the twin scene and the observation data are linked in real time. The implementation can be as follows: establish a data communication interface to realize the real-time transmission and interaction of the observation equipment data and the digital twin system; design a data assimilation algorithm, such as Kalman filtering or variational method, to assimilate the measured data into the simulation scene and correct the model prediction results; develop a visual dashboard to intuitively display the comparison between the digital twin scene and the measured data to provide support for management decisions; establish an automatic calibration mechanism to dynamically adjust the parameters and boundary conditions of the digital twin model according to the feedback of the measured data to improve the prediction accuracy.

[0013] S60. Use high-resolution numerical models or artificial intelligence models to construct the wave model and storm surge model of the digital twin area. This can be implemented as follows: For the use of high-resolution numerical models, determine the scope of the computational domain and the unstructured triangular grid, and encrypt the grid distribution in the nearshore area; use the global tidal model to provide astronomical tide levels, use the global wave model to provide wave spectrum parameters, and use a high-resolution meteorological model to provide wind field and air pressure data; select the fourth-order Runge-Kutta time format, MUSCL space format, and TVD format for the convection term discretization scheme; perform sensitivity analysis and automatic calibration of model parameters to minimize the deviation between simulated values ​​and observed values; couple the wave model and storm surge model to consider the impact of waves on storm surges, such as radiation stress and wave-induced mean flow. For the method of using artificial intelligence models, collect long-term series of wave and tide observation data, as well as corresponding meteorological and marine environmental data; perform data preprocessing, including outlier removal, standardization and feature engineering; select algorithms such as support vector regression (SVR), gradient boosting decision tree (GBDT), long short-term memory network (LSTM) or temporal convolutional network (TCN), train and optimize the model; evaluate the model performance on the test set, calculate indicators such as RMSE, MAE and correlation coefficient, and consider ensemble learning to improve prediction stability.

[0014] S70. Use the ocean wave model and storm surge model to forecast ocean waves and storm surges, with a forecast validity period of no less than 72 hours and a forecast result time interval of no more than 3 hours. The implementation may be as follows: obtain real-time meteorological and boundary condition data, such as wind field, air pressure field and astronomical tide level; input these data into the ocean wave model and storm surge model, and run numerical simulations. The ocean wave model includes an energy balance equation, a source term equation, a radiation stress equation and a bottom friction equation, and the storm surge model includes a continuity equation, a momentum equation, a pressure equation and a Coriolis force equation; output simulation results, such as wave height, wave period, tide level, flow velocity and other parameters; display the forecast results through a visual interface, and update them at intervals of 3 hours to provide decision support for management departments.

[0015] S80. Use the wave model and storm surge model to simulate historical hourly wave and storm surge scenarios for no less than 2 years, historical typical wave and storm surge disaster scenarios, wave and storm surge disaster scenarios with different return periods, and wave and storm surge scenarios under different disaster reduction measures. The implementation can be as follows: collect typical wave and storm surge events that have occurred in history and obtain corresponding meteorological and boundary condition data; input these data into the wave model and storm surge model to reproduce historical ocean dynamic processes; set reasonable boundary conditions for extreme events that may occur in the future and simulate the wave and storm surge disasters caused by them; evaluate different disaster reduction measures, such as seawall reinforcement, sandbag stacking, etc., and simulate their impact on disaster scenarios; organize the simulation results into a scenario library, including various scenarios under historical events, extreme situations and disaster reduction measures.

[0016] S90. Construct a disaster prevention and mitigation scenario library based on the simulation results, and use marine environment simulation technology to display different disaster prevention and mitigation scenarios. The implementation can be as follows: organize the various wave and storm surge disaster scenarios obtained by the simulation in step S80, and classify and sort them according to the probability of occurrence, degree of impact, etc.; use 3D modeling and rendering technologies such as Unity or UnrealEngine to make these scenarios into interactive simulation scenes; add elements such as beaches, buildings, and transportation facilities to the scenes, and give them real physical properties; simulate disaster processes such as wave impact, tidal flooding, and storm damage, and show their impact on the surrounding environment; design corresponding disaster prevention and mitigation measures for different scenarios, such as seawall reinforcement, sandbag stacking, seagrass beds, etc., and perform simulation to verify their effects; integrate the completed scenario library into the visual interface of the digital twin system to provide managers and the public with immersive disaster risk auxiliary decision support.

[0017] Among them, the wave simulation adopts JONSWAP spectrum and is specifically expressed as follows:

[0018] ;

[0019] ;

[0020] In the formula, is the spectral density, in units of ; is the frequency, in units of ; is the effective wave height, in units of ; is the peak period, in units of ; is the peak enhancement factor, dimensionless, ranging from 1 to 7, with an average value of 3.3; is the spectrum width parameter, when hour, ,when hour, ,in is the spectrum peak frequency, in units of .

[0021] Parameter acquisition method:

[0022] and It is obtained through wave observation equipment (such as wave buoys). The specific steps include: 1) deploying wave buoys; 2) recording the time series of water surface height; 3) calculating the effective wave height and spectral peak period through statistical analysis.

[0023] The specific steps include: 1) obtaining the measured spectrum; 2) fitting the JONSWAP spectrum model using the nonlinear least squares method; 3) extracting the best fit value.

[0024] The hydrodynamic simulation uses simplified continuity equation and momentum equation as follows:

[0025] Continuity equation:

[0026] ;

[0027] Momentum equation:

[0028] ;

[0029] ;

[0030] In the formula, is the water depth in ; is the time, the unit is ; and They are and The velocity component of the direction, in units of ; is the acceleration due to gravity, take 9.8 ; is the water level in ; and They are and The external force term in the direction of .

[0031] Parameter acquisition method:

[0032] It is obtained through seabed topography measurement. The specific steps include: 1) using a multi-beam echo sounder to measure the seabed topography; 2) data processing and interpolation to obtain gridded water depth data.

[0033] and It is obtained by measuring with a flow meter (such as ADCP). The specific steps include: 1) deploying ADCP; 2) recording flow velocity data of different water layers; 3) processing the data to obtain the horizontal flow velocity component.

[0034] It is obtained through measurement at a tide station or water level gauge. The specific steps include: 1) installing tide observation equipment; 2) continuously recording water level data; 3) data quality control and processing.

[0035] On the basis of the above technical solution, the method for constructing a nearshore overbank disaster digital twin system of the present invention can also be improved as follows:

[0036] The ocean wave model includes an energy balance equation, a source term equation, a radiation stress equation and a bottom friction equation.

[0037] Furthermore, the energy balance equation is used to describe the propagation and change of wave energy, the input variables include the wave energy spectrum and the wave propagation speed, and the output variable is the spatiotemporal change of the wave energy spectrum;

[0038] The source term equation is used to describe the growth and dissipation of wave energy, the input variables include wind speed, water depth and wave energy spectrum, and the output variable is the growth and dissipation rate of wave energy;

[0039] The radiation stress equation is used to calculate the horizontal force caused by waves, the input variables include wave energy spectrum and water depth, and the output variable is radiation stress;

[0040] The bottom friction equation is used to calculate the energy dissipation of waves on the seabed, the input variables include wave parameters and bottom roughness, and the output variable is the bottom friction coefficient.

[0041] Specifically, the ocean wave model includes the following equations:

[0042] 1. Energy balance equation:

[0043] ;

[0044] In the formula, is the wavenumber action density spectrum, in units of ; is the time, the unit is ; is the group velocity vector, in units of ; is the relative angular frequency, in units of ; , and are wind input term, nonlinear interaction term and dissipation term, respectively, and the unit is .

[0045] 2. Source term equation:

[0046] ;

[0047] ;

[0048] ;

[0049] In the formula, is the directional spectrum, in units of ; is the frequency, in units of ; is the wave direction, in units of ; and is the wind input parameter, which is calculated from wind speed and wave steepness; is the nonlinear interaction operator; is the dissipation coefficient; is the wave number, in units of .

[0050] 3. Radiation stress equation:

[0051] ;

[0052] ;

[0053] ;

[0054] In the formula, , , and is the radiation stress tensor component in ; is the water density, take 1025 ; is the ratio of group velocity to phase velocity, .

[0055] 4. Bottom friction equation:

[0056] ;

[0057] In the formula, is the bottom friction stress, in units of ; is the wave friction factor, which is calculated by the bottom roughness and the thickness of the wave boundary layer; is the bottom track speed in , calculated using linear wave theory.

[0058] Furthermore, the storm surge model includes a continuity equation, a momentum equation, a pressure equation, a Coriolis force equation and a boundary condition equation.

[0059] Furthermore, the continuity equation is used to describe the conservation of water mass, the input variables include flow velocity and water depth, and the output variable is the rate of change of water level;

[0060] The momentum equation is used to describe the movement of water bodies, the input variables include flow velocity, water depth and external force, and the output variable is the rate of change of flow velocity;

[0061] The pressure equation is used to calculate the pressure distribution inside the water body, the input variables include water depth and density, and the output variable is pressure;

[0062] The Coriolis force equation is used to describe the effect of the earth's rotation on the movement of water bodies. The input variables include flow velocity and latitude, and the output variable is the Coriolis force.

[0063] The boundary condition equation is used to describe the hydrodynamic characteristics of the model boundary. The input variables include the boundary type and external forcing, and the output variable is the water level or flow velocity on the boundary.

[0064] Specifically, the storm surge model includes the following equations:

[0065] 1. Continuity equation:

[0066] ;

[0067] 2. Momentum equation:

[0068] ;

[0069] ;

[0070] In the formula, is the Coriolis parameter, , is the Earth's rotation angular velocity, is latitude; and is the wind stress component, in units of ; and is the bottom friction stress component, in units of ; is the horizontal eddy viscosity coefficient, in units of .

[0071] 3. Pressure equation:

[0072] ;

[0073] In the formula, is the pressure, in units of ; is the vertical coordinate, positive upward, unit is .

[0074] 4. Coriolis force equation:

[0075] ;

[0076] In the formula, is the Coriolis force, with units of ; is the vertical unit vector; is the horizontal velocity vector.

[0077] 5. Boundary condition equation:

[0078] Open Boundary: ;

[0079] Land borders: ;

[0080] In the formula, is the astronomical tide level, obtained by harmonic analysis method; The water increase for storm surge is calculated by inverse pressure effect and wind stress; is the normal velocity component perpendicular to the boundary.

[0081] Parameter acquisition method:

[0082] Wind stress is calculated using meteorological observation data. The specific steps include: 1) obtaining wind speed and direction data; 2) using the wind stress formula Calculate, where is the air density, is the drag coefficient, is the wind speed vector at a height of 10 meters.

[0083] The bottom friction stress is calculated by the Manning formula. The specific steps include: 1) determining the bottom roughness; 2) using the formula Calculate, where is the Manning coefficient.

[0084] The horizontal eddy viscosity coefficient is calculated using the Smagorinsky formula:

[0085] ,in is the Smagorinsky constant, usually 0.1-0.2.

[0086] Furthermore, the high-resolution numerical model is specifically a data mapping relationship constructed based on a large amount of historical observation data and numerical simulation results, including the mapping relationship between wave characteristic parameters and environmental factors, and the mapping relationship between storm surge water levels and meteorological and ocean dynamic factors.

[0087] Furthermore, the specific steps of constructing the wave model and storm surge model of the digital twin area using the high-resolution numerical model include: determining the computational domain and grid division, setting initial conditions and boundary conditions, debugging and verifying model parameters, and coupling the wave model and storm surge model.

[0088] Furthermore, the specific steps of using artificial intelligence models to build the wave model and storm surge model of the digital twin area include: collecting historical observation data and numerical simulation results, data preprocessing and feature engineering, training models and optimizing hyperparameters, verifying model performance and performing integrated learning.

[0089] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are run in a computer, they are used to execute the above-mentioned method for constructing a digital twin system for nearshore overtopping disasters.

[0090] A third aspect of the present invention provides a system for constructing a digital twin system for nearshore overtopping disasters, which includes the above-mentioned computer-readable storage medium.

[0091] Compared with the prior art, the method, medium and system for constructing a nearshore overbank disaster digital twin system provided by the present invention have the following beneficial effects:

[0092] 1. Using the oblique photography and elevation measurement functions of drones, high-resolution three-dimensional real-life data of the target area is obtained, laying a solid foundation for the construction of the digital twin model. Compared with traditional two-dimensional GIS data, this method can more realistically reflect the geographical features and terrain changes of the sea area.

[0093] 2. Use advanced 3D reconstruction algorithms, such as feature extraction and matching, dense matching, and point cloud generation, to build a high-fidelity 3D real-scene model. This method can not only restore the geometric shape of the target area, but also attach real texture effects, greatly improving the realism of the digital twin model.

[0094] 3. For the two major types of nearshore disasters, sea waves and storm surges, high-precision numerical models based on physical mechanisms and data-driven are constructed. Among them, the sea wave model includes energy balance equations, source term equations, and radiation stress equations, which can accurately simulate the propagation and change process of waves; the storm surge model includes continuity equations, momentum equations, and pressure equations, which can reflect the dynamic process of surges and water level changes. These models can provide a reliable numerical basis for subsequent disaster scenario analysis.

[0095] 4. The dynamic integration of the digital twin system and real-time observation data is realized. By deploying buoys, ADCP and other observation equipment, measured data such as waves and water levels are obtained, and assimilated into the digital twin model in real time to correct the simulation results, greatly improving the accuracy and credibility of the forecast.

[0096] 5. Based on the constructed digital twin model, simulations of ocean dynamic processes at different time scales were carried out, including historical event reproduction, extreme situation prediction, and disaster reduction measures evaluation. These simulation results were compiled into a rich disaster prevention and reduction scenario library and displayed using three-dimensional visualization technology. Managers and the public can experience and understand the effects of disaster prevention measures under different disaster situations in person, providing valuable support for decision-making.

[0097] In general, the method for constructing a digital twin system for nearshore overtopping disasters proposed in the present invention makes full use of multi-source remote sensing data and advanced ocean dynamics simulation technology to construct a high-fidelity digital twin model, realizes real-time fusion with measured data, supports multiple disaster scenario analysis and disaster prevention and mitigation measures evaluation, and solves the technical problem that the existing technology often uses traditional two-dimensional GIS data as the basis for modeling, which cannot fully reflect the three-dimensional dynamic characteristics of the sea area and is difficult to better respond to and predict the situation of nearshore marine disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0098] Figure 1 is a flow chart of the present invention;

[0099] Figure 2 A three-dimensional real scene model diagram constructed in the embodiment;

[0100] Figure 3 This is a schematic diagram of the high-resolution storm surge model grid;

[0101] Figure 4 Schematic diagram of wave protection scheme. DETAILED DESCRIPTION

[0102] In order to make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0103] like Figure 1 As shown, it is a flow chart of a method for constructing a nearshore overbank disaster digital twin system provided by the present invention, and the method comprises the following steps:

[0104] S10, obtaining multi-angle photos and elevation data of the target area, specifically, using a drone with oblique photography and elevation measurement functions to perform aerial survey to obtain multi-angle photos and elevation data of the target area;

[0105] S20, constructing 3D real scene models based on multi-angle photos and elevation data, including image preprocessing, feature extraction and matching, aerial triangulation, dense matching and point cloud generation, 3D reconstruction, texture mapping and model optimization;

[0106] S30, using real-time 3D creation tools for wave and hydrodynamic simulation, where JONSWAP spectrum is used for wave simulation and simplified continuity equation and momentum equation are used for hydrodynamic simulation;

[0107] S40. Carry out real-time observation in the digital twin area or nearby areas, using buoys or bottom-mounted ADCP observation instruments to obtain real-time observation data, including wave height and water level;

[0108] S50, transferring the real-time observation data to the digital twin system, so that the twin scene and the observation data are linked in real time;

[0109] S60. Use high-resolution numerical models or artificial intelligence models to build wave models and storm surge models for the digital twin area;

[0110] S70. Use the ocean wave model and storm surge model to forecast ocean waves and storm surges, with a forecast validity period of no less than 72 hours and a forecast result interval of no more than 3 hours;

[0111] S80. Use wave models and storm surge models to simulate historical wave and storm surge scenarios of no less than 2 years, historical typical wave and storm surge disaster scenarios, wave and storm surge disaster scenarios with different return periods, and wave and storm surge scenarios under different disaster reduction measures;

[0112] S90. Build a disaster prevention and mitigation scenario library based on simulation results, and use marine environment simulation technology to demonstrate different disaster prevention and mitigation scenarios.

[0113] The specific implementation methods of the above steps are described in detail below:

[0114] The specific implementation method of step S10 is to first use a drone with oblique photography and elevation measurement functions to perform aerial photography and topographic measurement of the target area. The specific steps include: 1) Select a suitable drone platform, such as DJI Phantom 4RTK or Wingtra One, and install a high-precision camera sensor and RTK-GNSS positioning system. 2) Determine the appropriate route and altitude according to the actual situation of the target area, generally select a flight altitude of 50-100 meters to obtain clear oblique photography images and high-precision terrain data. 3) Accurately position and calibrate the drone to ensure the measurement accuracy of its position and attitude data. 4) Perform autonomous or remote-controlled aerial survey tasks according to the established route, and record the images and terrain data of each voyage. 5) Perform quality inspection and preprocessing on the acquired raw data to remove outliers and noise. By comprehensively utilizing these multi-angle photos and elevation data, a high-precision three-dimensional real-life model of the target area can be constructed. The main purpose of this step is to obtain the geometric information of the target area and lay the foundation for the subsequent establishment of a digital twin model.

[0115] The specific implementation method of step S20 is to construct a three-dimensional real scene model of the target area based on the multi-angle photos and elevation data obtained in step S10 using three-dimensional reconstruction technology. The specific steps include: 1) Image preprocessing, de-distortion, color correction and other processing of the original photos to improve the image quality. 2) Feature extraction and matching, using scale-invariant feature transform (SIFT), speeded up robust features (SURF) and other algorithms to extract and match key feature points from adjacent photos. 3) Aerial triangulation, based on the feature point matching results, using the bundle adjustment algorithm to estimate the pose parameters of each photo and obtain the camera's external orientation elements. 4) Dense matching and point cloud generation, using semi-global matching (SGM), deep learning and other methods to generate dense three-dimensional point cloud data. 5) Three-dimensional reconstruction, using point cloud data to construct a triangular mesh model, and perform geometric refinement and topology optimization. 6) Texture mapping, attaching the original photo to the surface of the three-dimensional model to obtain a real texture effect. 7) Model optimization, applying filtering and other technologies to improve the geometric accuracy and rendering effect of the three-dimensional model. The main purpose of this step is to use multi-source remote sensing data to build a high-fidelity three-dimensional real-scene model as the basis of the digital twin system.

[0116] The specific implementation method of step S30 is to use real-time 3D creation tools such as Unity or UnrealEngine to simulate the waves and hydrodynamic processes in the digital twin scene. The specific steps include: 1) Importing the 3D real scene model constructed in step S20 into the 3D creation tool, setting the basic properties of the scene such as terrain, material, etc. 2) Selecting a suitable wave simulation algorithm. The present invention adopts a wave model based on the JONSWAP spectrum. Among them, the wave energy spectrum The expression is: ;in, 3) Set the effective wave height according to the measured data or forecast results , spectral peak period and peak elevation factor 4) Use simplified forms of the continuity equation and momentum equation to simulate the hydrodynamic process. The continuity equation is: ;The momentum equation is: ; ;5) Visualize the dynamic changes of waves and hydrodynamics through real-time rendering technology. The main purpose of this step is to build a lifelike digital twin model of the marine environment, laying the foundation for subsequent disaster scenario simulation.

[0117] The specific implementation method of step S40 is to set up real-time observation equipment in or near the digital twin area to obtain measured data of ocean dynamic parameters. The specific steps include: 1) Deploy observation instruments such as buoys and bottom-mounted ADCPs in the target sea area to ensure their stability and data transmission reliability. 2) Use high-frequency (1-2Hz) sampling to record time series data such as water level, flow velocity, and flow direction. 3) Perform quality control on the original observation data, eliminate outliers and noise, and perform necessary corrections and interpolations. 4) Transmit the processed measured data to the digital twin system in real time and associate it with the simulation scene in real time. The main purpose of this step is to obtain measured data of ocean dynamic parameters to provide a basic basis for the calibration and verification of the digital twin model.

[0118] The specific implementation method of step S50 is to input the real-time observation data obtained in step S40 into the digital twin system so that the digital twin scene is linked with the actual ocean conditions in real time. The specific steps include: 1) Establishing a data communication interface to realize the real-time transmission and interaction between the observation equipment data and the digital twin system. 2) Designing a data assimilation algorithm, such as Kalman filtering or variational method, to assimilate the measured data into the simulation scene and correct the model prediction results. 3) Developing a visual dashboard to intuitively display the comparison between the digital twin scene and the measured data to provide support for management decisions. 4) Establishing an automatic calibration mechanism to dynamically adjust the parameters and boundary conditions of the digital twin model according to the feedback of the measured data to improve the prediction accuracy. The main purpose of this step is to tightly couple the digital twin scene with the actual ocean conditions and enhance the model prediction ability and application value.

[0119] The specific implementation method of step S60 is to use a high-resolution numerical model or an artificial intelligence model to construct a wave and storm surge model of the digital twin area. First, the implementation method of using a high-resolution numerical model is introduced: 1) Determine the scope of the calculation domain and the grid division, use an unstructured triangular grid, and encrypt the grid distribution in the nearshore area. 2) Set initial conditions and boundary conditions, use the global tidal model to provide astronomical tide levels, use the global wave model to provide wave spectrum parameters, and use a high-resolution meteorological model to provide wind field and pressure data. 3) Select a suitable numerical discretization scheme, such as the fourth-order Runge-Kutta time format, MUSCL space format, and TVD format for convection term discretization. 4) Perform sensitivity analysis and automatic calibration of model parameters to minimize the deviation between simulated values ​​and observed values. 5) Couple the wave model and storm surge model to consider the impact of waves on storm surges, such as radiation stress and wave-induced mean flow.

[0120] Next, the implementation method of using artificial intelligence models is introduced: 1) Collect long-term series of wave and tide observation data, as well as corresponding meteorological and marine environmental data. 2) Perform data preprocessing, including outlier removal, standardization, and feature engineering. 3) Select appropriate machine learning algorithms. For wave models, support vector regression (SVR) or gradient boosting decision tree (GBDT) can be used, and for storm surge models, long short-term memory network (LSTM) or temporal convolutional network (TCN) can be used. 4) Use training set data to train the model and optimize hyperparameters on the validation set. 5) Evaluate model performance on the test set, calculate indicators such as RMSE, MAE, and correlation coefficient, and consider ensemble learning methods to improve prediction stability. 6) Deploy the model to a cloud platform or edge device to achieve real-time data acquisition and prediction.

[0121] The main purpose of this step is to construct a high-precision ocean dynamics model to provide reliable numerical support for subsequent disaster scenario analysis.

[0122] The specific implementation method of step S70 is to use the wave and storm surge model constructed in step S60 to make real-time forecasts for more than 72 hours, with a time interval of no more than 3 hours. The specific steps include: 1) obtaining real-time meteorological and boundary condition data, such as wind field, pressure field and astronomical tide level. 2) inputting these data into the wave model and storm surge model to run numerical simulation. The wave model includes energy balance equation, source term equation, radiation stress equation and bottom friction equation, and the storm surge model includes continuity equation, momentum equation, pressure equation and Coriolis force equation. 3) output simulation results, such as wave height, wave period, tide level, flow velocity and other parameters. 4) display the forecast results through a visual interface and update them at intervals of 3 hours to provide decision support for management departments. The main purpose of this step is to achieve continuous monitoring and forecasting of ocean dynamic processes and provide timely information for disaster prevention and mitigation.

[0123] The specific implementation method of step S80 is to use the ocean dynamics model constructed in step S60 to simulate different historical wave and storm surge scenarios, as well as disaster scenarios with different recurrence periods. The specific steps include: 1) Collecting typical wave and storm surge events that have occurred in history and obtaining corresponding meteorological and boundary condition data. 2) Inputting these data into the wave model and storm surge model to reproduce the historical ocean dynamic process. 3) For extreme events that may occur in the future, such as a tsunami or tropical storm that occurs once in a hundred years, setting reasonable boundary conditions to simulate the wave and storm surge disasters caused by them. 4) Evaluating different disaster reduction measures, such as seawall reinforcement and coastal pipeline reconstruction, and simulating their impact on disaster scenarios. 5) Organizing the simulation results into a scenario library, including various scenarios under historical events, extreme situations and disaster reduction measures. The main purpose of this step is to fully understand the possible marine disaster risks in the region and provide a basis for the formulation of disaster prevention and reduction strategies.

[0124] The specific implementation method of step S90 is to build a targeted disaster prevention and mitigation scenario library based on the above simulation results, and use marine environment simulation technology for visual display. The specific steps include: 1) sorting out various wave and storm surge disaster scenarios obtained by simulation in step S80, and classifying and sorting them according to the probability of occurrence, degree of impact, etc. 2) using three-dimensional modeling and rendering technology, such as Unity or UnrealEngine, to make these scenarios into interactive simulation scenes. 3) adding elements such as beaches, buildings, and transportation facilities to the scene, and giving them real physical properties. 4) simulating disaster processes such as wave impact, tidal flooding, and storm damage, and showing their impact on the surrounding environment. 5) designing corresponding disaster prevention and mitigation measures for different scenarios, such as seawall reinforcement, evacuation channels, etc., and conducting simulation to verify their effects. 6) integrating the completed scenario library into the visual interface of the digital twin system to provide managers and the public with immersive disaster risk auxiliary decision support.

[0125] The step of acquiring the wave characteristic parameters specifically includes:

[0126] 1. Deploy wave observation equipment: Install wave buoys or fixed wave height meters in the target sea area to ensure equipment stability and data transmission reliability;

[0127] 2. Collect raw data: set appropriate sampling frequency (usually 1-2 Hz) and duration (at least 20 minutes) to record the water surface height time series;

[0128] 3. Data preprocessing: perform quality control on raw data, remove outliers and trends, and apply low-pass filters to eliminate high-frequency noise;

[0129] 4. Time domain analysis: calculation of effective wave height and average period ,in is the average value of the highest 1 / 3 wave height, is the average value of all wave periods;

[0130] 5. Frequency domain analysis: Perform fast Fourier transform (FFT) on the pre-processed data to obtain the wave energy spectrum ;

[0131] 6. Spectral parameter calculation: Calculate the spectrum peak period based on the energy spectrum , spectral width parameters and peak gain factor , the specific formula is as follows:

[0132] ,in is the spectrum peak frequency;

[0133] ,in is the nth order spectral moment, ;

[0134] ,in is the corresponding Pierson-Moskowitz spectrum;

[0135] 7. Direction parameter calculation: Use the direction wave meter data to calculate the average wave direction and direction widening :

[0136] ;

[0137] ;

[0138] in, and is the first-order directional coefficient, obtained by cross spectrum analysis.

[0139] The step of establishing the mapping relationship between the wave characteristic parameters and the environmental factors specifically includes:

[0140] 1. Data collection: Obtain long-term series of ocean wave characteristic parameters (such as , , ) and corresponding environmental factor data (such as wind speed ,wind direction , Wind-taking time , water depth );

[0141] 2. Data preprocessing: standardize the data, eliminate the dimension effect, and perform missing value interpolation and outlier processing;

[0142] 3. Correlation analysis: Calculate the correlation coefficient matrix between wave characteristic parameters and environmental factors to identify the main influencing factors;

[0143] 4. Establish empirical formulas: Based on physical mechanisms and statistical analysis, establish empirical formulas between wave characteristic parameters and environmental factors, such as the relationship between JONSWAP spectrum parameters and wind field parameters:

[0144] ;

[0145] ;

[0146] In the formula, is the effective wind zone length, in units of ; is the acceleration due to gravity, take 9.81 ;

[0147] 5. Nonlinear regression analysis: Use the nonlinear least squares method to fit the functional relationship between wave characteristic parameters and environmental factors, such as:

[0148] ;

[0149] ;

[0150] In the formula, , , , and , , , is the unknown coefficient, which is solved by minimizing the sum of squared errors;

[0151] 6. Machine learning model: Use algorithms such as support vector regression (SVR) or random forest (RF) to establish a nonlinear mapping relationship between wave characteristic parameters and environmental factors;

[0152] 7. Model validation and optimization: Use an independent validation dataset to evaluate model performance, calculate the root mean square error (RMSE) and correlation coefficient (R), and optimize model performance through cross-validation and parameter tuning.

[0153] The steps of establishing the mapping relationship between the storm tide water level and the meteorological and ocean dynamic factors specifically include:

[0154] 1. Data collection: Obtain long-term series of storm surge water level observation data , and corresponding meteorological factors (such as air pressure , wind speed ,wind direction ) and ocean dynamic factors (such as astronomical tide levels , Wave parameters and );

[0155] 2. Data preprocessing: detrending and standardizing the data, and performing time synchronization and interpolation;

[0156] 3. Static water increase calculation: Based on the anti-air pressure effect, calculate the static water increase :

[0157] ;

[0158] In the formula, is the standard atmospheric pressure, take 101325 ; is the density of seawater, take 1025 ;

[0159] 4. Dynamic water increase analysis: Calculate dynamic water increase , that is, the total water increase minus the static water increase:

[0160] ;

[0161] 5. Wind stress parameterization: Calculation of wind stress , using the quadratic drag law:

[0162] ;

[0163] In the formula, is the air density, take 1.225 ; is the drag coefficient, the Garratt formula can be used:

[0164] ;

[0165] 6. Establish empirical formula: Based on physical mechanism, establish the relationship between dynamic water increase, wind stress and water depth:

[0166] ;

[0167] In the formula, is the empirical coefficient; For the coastline direction; is the average water depth;

[0168] 7. Nonlinear regression analysis: Use the nonlinear least squares method to fit the functional relationship between storm surge water level and meteorological and ocean dynamic factors:

[0169] ;

[0170] In the formula, , , , , is the unknown coefficient, which is solved by minimizing the sum of squared errors;

[0171] 8. Machine learning model: Use algorithms such as long short-term memory network (LSTM) or gradient boosting decision tree (GBDT) to establish a nonlinear time series mapping relationship between storm surge water level and meteorological and ocean dynamic factors;

[0172] 9. Model validation and optimization: Use an independent validation dataset to evaluate model performance, calculate the root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient (R), and optimize model performance through cross-validation and parameter tuning.

[0173] The specific implementation method of using high-resolution numerical models to construct the wave model and storm surge model of the digital twin area is:

[0174] 1. Determine the computational domain and mesh division:

[0175] (a) Determine the scope of the computational domain according to the research objectives, taking into account the influence of open boundary conditions;

[0176] (b) Use unstructured triangular meshes for spatial discretization and refine the mesh in coastal and complex terrain areas;

[0177] (c) The grid resolution gradually changes from several kilometers in the open sea to tens of meters near the coast;

[0178] 2. Set initial conditions and boundary conditions:

[0179] (a) Initial conditions: Use static start or hot start. In static start, the initial water level and flow rate are assumed to be zero. In hot start, the previous simulation results or reanalysis data are used.

[0180] (b) Open boundary conditions: using a global tidal model (such as FES2014) to provide astronomical tide levels and a global wave model (such as WaveWatchIII) to provide wave spectrum parameters;

[0181] (c) Land boundary conditions: no-slip boundary conditions are adopted, and the normal velocity is zero;

[0182] (d) Air-sea interface boundary conditions: using high-resolution numerical weather prediction models (such as WRF) to provide wind and pressure field data;

[0183] 3. Choose an appropriate numerical solution method:

[0184] (a) Time discretization: using explicit or semi-implicit schemes, such as the fourth-order Runge-Kutta method;

[0185] (b) Spatial discretization: using finite element or finite volume method, such as MUSCL format;

[0186] (c) Discretization of convection terms: using the upwind format or TVD format, such as the QUICKEST-ULTIMATE scheme;

[0187] 4. Perform model parameter debugging and verification:

[0188] (a) Sensitivity analysis: Through the single factor variation method, the influence of parameters such as bottom friction coefficient and horizontal eddy viscosity coefficient on the simulation results is analyzed;

[0189] (b) Parameter calibration: Use optimization algorithms (such as genetic algorithms) to automatically adjust model parameters to minimize the error between simulated and observed values;

[0190] (c) Model validation: select typical storm surge events, compare simulation results with observation data, and calculate statistical indicators (such as RMSE and correlation coefficient) to evaluate model performance;

[0191] 5. Coupled wave model and storm surge model:

[0192] (a) Using the segmented coupling method, the ocean wave model is run first, and then the wave parameters are input into the storm surge model;

[0193] (b) Consider the impact of waves on storm surge, including radiation stress, wave-induced bottom friction and wave-induced mean flow;

[0194] (c) Consider the impact of storm surge on ocean waves, including changes in water depth and background flow fields;

[0195] (d) Iteratively solve the coupled equations until convergence.

[0196] The specific implementation method of using artificial intelligence models to build the wave model and storm surge model of the digital twin area is:

[0197] 1. Collect historical observation data and numerical simulation results:

[0198] (a) Organize long-term wave and tide observation data, including , , and wait;

[0199] (b) Collect relevant meteorological and oceanographic environmental data, such as wind speed, wind direction, air pressure, water depth, etc.;

[0200] (c) Obtain simulation results of high-resolution numerical models to expand the training data set;

[0201] 2. Data preprocessing and feature engineering:

[0202] (a) Data cleaning: remove outliers and process missing data;

[0203] (b) Data standardization: normalize features of different dimensions;

[0204] (c) Time alignment: ensuring that all data have the same temporal resolution and synchronization;

[0205] (d) Feature construction: creating new features, such as wind speed squared, wind direction cosine component, etc.;

[0206] (e) Feature selection: Use correlation analysis or principal component analysis (PCA) to screen key features;

[0207] 3. Choose the appropriate machine learning algorithm:

[0208] (a) For the ocean wave model:

[0209] Support Vector Regression (SVR): Applicable to nonlinear relationships and has good generalization ability;

[0210] Random Forest (RF): can handle high-dimensional features and is robust to noise;

[0211] Gradient Boosted Decision Tree (GBDT): such as XGBoost or LightGBM, which has the characteristics of high accuracy and fast training;

[0212] (b) For storm surge model:

[0213] Long short-term memory network (LSTM): suitable for time series data and can capture long-term dependencies;

[0214] Gated Recurrent Unit (GRU): A variant of LSTM with fewer parameters and faster training;

[0215] Temporal Convolutional Network (TCN): combines the advantages of convolutional neural networks and sequence modeling;

[0216] 4. Train the model and optimize hyperparameters:

[0217] (a) Dataset division: training set, validation set, and test set are divided into training set, validation set, and test set in a ratio of 8:1:1;

[0218] (b) Model training: Use the training set data to train the model, using optimization algorithms such as mini-batch gradient descent;

[0219] (c) Hyperparameter optimization: Use grid search or Bayesian optimization methods to find the optimal hyperparameter combination on the validation set;

[0220] (d) Regularization: Apply L1 / L2 regularization or Dropout techniques to prevent overfitting;

[0221] 5. Verify model performance and perform ensemble learning:

[0222] (a) Evaluate the model performance on the test set and calculate indicators such as RMSE, MAE and correlation coefficient;

[0223] (b) Analyze the performance of the model under different conditions, such as extreme events and normal conditions;

[0224] (c) Using ensemble learning methods, such as bagging or stacking, to combine multiple models to improve prediction stability and accuracy;

[0225] (d) Compare with traditional numerical models and analyze the advantages and limitations of artificial intelligence models;

[0226] 6. Model deployment and real-time prediction:

[0227] (a) Deploy the trained model to the cloud platform or edge device;

[0228] (b) Establish real-time data acquisition and preprocessing processes;

[0229] (c) Design a visualization interface for forecast results to facilitate decision makers’ understanding and use;

[0230] (d) Implement an online update mechanism for the model to continuously optimize model performance based on new data.

[0231] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are run in a computer, they are used to execute the above-mentioned method for constructing a digital twin system for nearshore overtopping disasters.

[0232] A third aspect of the present invention provides a system for constructing a digital twin system for nearshore overtopping disasters, which includes the above-mentioned computer-readable storage medium.

[0233] Specifically, the principle of the present invention is: first, the oblique photography and elevation measurement functions of the drone are used to obtain multi-angle photos and elevation data of the target area, providing high-quality basic data for subsequent three-dimensional reconstruction. Compared with traditional two-dimensional geographic information data, this method can more comprehensively depict the geographical environment of the sea area, laying a solid foundation for the construction of the digital twin model. In the three-dimensional reconstruction link, the present invention uses advanced algorithms such as feature extraction and matching, dense matching, and point cloud generation, which can extract key features from multi-source image data, and finally construct a realistic three-dimensional real-scene model with the help of aerial triangulation and geometric optimization. This series of technical means ensures the high fidelity of the digital twin model, which is conducive to the subsequent simulation of ocean dynamic processes.

[0234] Secondly, for the two main types of nearshore disasters, sea waves and storm surges, the present invention constructs a high-precision numerical model based on physical mechanisms. Among them, the sea wave model uses energy balance equations, source term equations, and radiation stress equations, etc., which can accurately reflect the propagation and change process of waves; the storm surge model uses continuity equations, momentum equations, and pressure equations, etc., which can simulate the dynamic process of surges and water level changes. These models not only contain the basic laws of ocean dynamics, but also integrate a large amount of historical observation data and numerical simulation results to ensure the reliability and accuracy of the simulation results. At the same time, the present invention also proposes two construction methods using high-resolution numerical models and artificial intelligence models. The former is based on physical mechanisms, and the latter is based on data-driven. The two complement each other and are conducive to adapting to different application scenario requirements.

[0235] Thirdly, the present invention realizes the dynamic fusion of the digital twin system and real-time observation data. By deploying observation equipment such as buoys and ADCP, the measured data of ocean dynamic parameters are obtained and transmitted to the digital twin system in real time. Using data assimilation algorithms such as Kalman filtering and variational methods, the measured data are fed back to the numerical model to correct the simulation results. This method not only improves the accuracy of model prediction, but also enhances the digital twin system's perception of actual ocean conditions, so that the virtual scene is highly consistent with the physical world.

[0236] Finally, based on the constructed digital twin model, the present invention carried out simulations of ocean dynamic processes at different time scales, including historical event reproduction, extreme situation prediction, and disaster reduction measures evaluation. These simulation results were organized into a rich disaster prevention and reduction scenario library and displayed using three-dimensional visualization technology. This method can not only help managers fully understand the disaster risks in the region, but also provide a basis for the formulation of disaster prevention and reduction strategies.

[0237] In order to better understand and implement the present invention, an embodiment of a specific application scenario of the present invention is provided below: In this embodiment, the coastal sea and land of Macau Road in Shibei District, Qingdao City are selected as the target area to construct a digital twin system for nearshore overtopping disasters. The area is located on the east coast of Qingdao Bay, including approximately 3 square kilometers of sea area and 2 square kilometers of land. It is an important tourist attraction and residential area in Qingdao. The area has been hit by typhoons and storm surges many times in history, such as the backflow of seawater and damage to coastal facilities caused by Typhoon "Wimbia" in 2018 and Typhoon "Fireworks" in 2021, which has had a serious impact on the lives of local residents and the tourism industry. In order to improve the disaster prevention and mitigation capabilities of the area, the Qingdao Municipal Ocean Development Bureau decided to adopt the method proposed in the present invention to conduct comprehensive digital twin modeling and disaster scenario analysis of the target area.

[0238] Step S10: Obtain multi-angle photos and elevation data of the target area. The research team used the DJI Matrice300RTK drone to conduct aerial survey missions in the target area from April 15 to 17, 2023. The drone flew at an altitude of 120 meters, with a route overlap of 80% front and back and 60% lateral. A total of 18 flights were completed, and 7,500 high-resolution oblique photography images were obtained, covering a total area of ​​approximately 5 square kilometers. The P1 camera carried by the drone has 45 million pixels and a resolution of 2.5 cm. At the same time, the airborne lidar module L1 obtained high-precision three-dimensional point cloud data with a point cloud density of 200 points / square meter. After data preprocessing and quality control, the orthophoto map and digital surface model (DSM) of the target area were finally formed.

[0239] Table 1: Main parameters of UAV aerial survey missions

[0240] parameter Numeric Aerial survey time April 15-17, 2023 Flight altitude 120m Route overlap 80% front and rear, 60% sideways Number of voyages 18 times Get the number of images 7,500 Camera resolution 45 million pixels Ground resolution 2.5 cm Point cloud density 200 dots / square meter

[0241] Step S20: Construct a three-dimensional real-scene model based on multi-angle photos and elevation data. The research team used Bentley Context Capture software to process the acquired images and point cloud data. First, through feature extraction and matching algorithms, about 1.5 million feature points were extracted and matched from adjacent photos to achieve aerial triangulation of the image. Then, a multi-view stereo matching algorithm was used to generate a high-density three-dimensional point cloud with a point cloud density of 500 points / square meter. Finally, through steps such as Poisson reconstruction and texture mapping, a three-dimensional real-scene model with a resolution of 10 cm was constructed. The model includes buildings along Macau Road, seaside boardwalks, breakwaters and other facilities, as well as underwater terrain in nearshore waters. The absolute position accuracy of the model reaches 5 cm, and the relative accuracy is better than 2 cm, such as Figure 2 shown.

[0242] Step S30: Use real-time 3D creation tools to simulate waves and hydrodynamics. Based on the 3D real-life model built in step S20, the research team used the Unity2022.3 engine to model and simulate the ocean dynamic process. First, the 3D model was imported into the Unity scene, and the terrain, material, and lighting properties that were consistent with the actual situation were set. Then, the wave simulation algorithm based on the JONSWAP spectrum was used to set the wave parameters according to the observation data of Qingdao Ocean Station in the past five years: effective wave height , spectral peak period and peak elevation factor For the hydrodynamic simulation, the team used simplified continuity and momentum equations, where the water depth , flow rate and Parameters such as are derived from the measured data obtained in step S40. Through Unity's physics engine and custom shaders, realistic wave propagation, wave breaking and surge effects are achieved. The simulation results show that under normal sea conditions, the maximum wave height along the coast of Macau Road is about 1.8 meters, and the surge can reach the bottom of the seaside boardwalk.

[0243] Step S40: Carry out real-time observation in the digital twin area. In order to obtain the measured data of ocean dynamic parameters, the research team deployed multiple sets of observation equipment in the target area. Specifically, it includes: 1. A buoy wave monitoring system (WaveSentry-9) is deployed about 1 km from the coastline to continuously record the water level time series with a sampling frequency of 2Hz. 2. Two bottom-mounted ADCPs (RDI Workhorse Sentinel 600kHz) are deployed in nearshore waters to measure the flow velocity and flow direction data of different water layers with a sampling frequency of 1Hz and a vertical resolution of 0.5 meters. 3. Three tide stations are set up along Macau Road to continuously monitor water level changes with a sampling frequency of 1Hz and an accuracy of better than 1 cm. 4. An automatic weather station is installed in the seaside park to monitor meteorological elements such as wind speed, wind direction, and air pressure with a sampling frequency of 1 minute.

[0244] Table 2: Main parameters of real-time observation system

[0245] Observation equipment quantity Location Main parameters Buoy wave monitoring system 1 1 km from the shore Sampling frequency 2Hz ADCP 2 Nearshore waters Sampling frequency 1Hz, vertical resolution 0.5m Tide Station 3 Along Macau Road Sampling frequency 1Hz, accuracy <1cm Automatic Weather Station 1 Seaside Park Sampling frequency 1 minute

[0246] Step S50: The real-time observation data is transmitted to the digital twin system so that the twin scene and the observation data are linked in real time. The research team developed a data transmission system based on the 4G network to transmit the data collected by the observation equipment to the digital twin platform in real time. The platform uses ApacheKafka message queue technology to process high-frequency data streams to ensure the real-time and reliability of the data. In order to achieve the dynamic fusion of the digital twin model and the measured data, the team used the ensemble Kalman filter (EnKF) algorithm to assimilate the measured water level, flow rate and other data into the ocean dynamics model. At the same time, a web-based visualization dashboard was developed to intuitively display the comparison between the digital twin scene and the measured data. After multiple iterations of optimization, the wave height forecast error of the digital twin model was reduced to within 10 cm, and the tide level forecast error was controlled within 5 cm, reaching a practical level.

[0247] Step S60: Use high-resolution numerical models to construct the wave model and storm surge model of the digital twin area. The research team selected the SWAN and ADCIRC modeling systems and used unstructured grids to discretize the target sea area. The computational domain covers Qingdao Bay and its surrounding waters, with a total area of ​​approximately 200 square kilometers. The grid resolution in the nearshore area is 10 meters, and gradually transitions to 100 meters in the offshore area. The team uses the global tidal model NAO.99b to provide open-boundary astronomical tide levels, uses the WAVEWATCHIII model to provide wave spectrum parameter boundaries, and uses the WRF meteorological model to obtain wind field and sea level pressure field data at a height of 10 meters, such as Figure 3 shown.

[0248] In terms of the wave model, the Spectral Wave Model based on unstructured grids is used, taking into account wind-generated waves, swells, wave breaking, bottom friction and other processes. The expression of is based on the JONSWAP spectrum:

[0249] ;

[0250] in, is the Phillips constant, is the peak frequency, is the peak enhancement factor, is the directional distribution function. By fitting the historical observation data, we determined , And other parameters.

[0251] For the storm surge model, a two-dimensional hydrodynamic module (HydrodynamicModule) is used, including the continuity equation and momentum equation. The calculation formulas for wind stress and bottom friction stress are:

[0252] ;

[0253] By analyzing historical storm surge events, the appropriate drag coefficient was determined and Manning coefficient The calculation time step of the model is 6 seconds, and the output interval is 10 minutes. After repeated debugging and parameter calibration, the forecast accuracy of this high-resolution numerical model has been significantly improved. In the verification during Typhoon "Miaobai" in 2022, the correlation coefficients of the simulation results of wave height and water level with the observed data reached 0.92 and 0.95 respectively, and the root mean square errors were 0.22 meters and 0.15 meters respectively.

[0254] Step S70: Use the constructed wave model and storm surge model for real-time forecasting. Based on the digital twin model constructed above, the research team has achieved continuous monitoring and forecasting of the dynamic process in the coastal waters of Macau Road. The specific process is as follows: 1. Obtain the wind field and pressure field forecast data of the WRF model from the team's business system once every hour, with a time resolution of 1 hour and a spatial resolution of 3 kilometers. 2. Obtain the open boundary condition data of the global tidal model and wave model from the team's business system once every 6 hours. 3. Input these data into the SWAN and ADCIRC models, run numerical simulations, and obtain forecast results such as wave height, period, wave direction, tide level and flow rate in the next 72 hours, with a time resolution of 10 minutes and a spatial resolution of 10 meters. 4. Display the forecast results through the WebGIS platform, including wave height distribution maps, tide level time series, flow field vector maps, etc., and provide early warning information for key locations.

[0255] The forecast results show that under normal weather conditions, the effective wave height along the coast of Macau Road is usually between 0.5 and 1.5 meters, and the maximum tidal change is about 2.2 meters. When strong winds and waves occur, the system will automatically issue an early warning message to remind relevant departments to take protective measures.

[0256] Step S80: Use the constructed model to simulate different historical wave and storm surge scenarios, extreme event scenarios, and scenarios under different disaster reduction measures. First, the research team reproduced two typical storm surge nearshore wave events, Typhoon "Fireworks" in 2021 and Typhoon "Meihua" in 2022. The simulation results show that during "Meihua", the maximum effective wave height along the coast of Macau Road reached 4.0 meters, and the maximum water increase reached 0.8 meters, causing seawater to overflow the breakwater and cause damage to coastal facilities. During the "Fireworks" period, although the effective wave height was relatively small (maximum 2.9 meters), it caused serious seawater backflow due to the superposition with astronomical tides. For this, the research team designed a plan to add flood control sandbags to the existing coast, which can temporarily increase the height of the square wave embankment by 1 meter. Figure 4 shown.

[0257] Step S90: construct a disaster prevention and mitigation scenario library based on the simulation results, and use marine environment simulation technology to display different disaster prevention and mitigation scenarios.

[0258] The research team sorted and classified the various wave and storm surge disaster scenarios simulated in step S80, and ranked them according to the probability of occurrence, impact degree and other characteristics. A total of 50 typical scenarios were constructed, including 15 historical event re-enactments, 10 extreme situation predictions and 25 situations under different disaster reduction measures. These scenarios cover all possibilities from minor impacts to catastrophic consequences.

[0259] In order to intuitively display these scenarios, the team used the UnrealEngine engine to build an interactive 3D simulation scene. In the scene, the topography, buildings, roads, vegetation and other elements along the Macau Road were carefully restored and given real physical properties. For different scenarios, the process of wave impact, tidal flooding, storm damage and other processes were simulated, and their impact on the surrounding environment was demonstrated. The specific implementation method is as follows:

[0260] 1. Ocean wave simulation: The FFT-based sea surface generation algorithm is combined with the Jonswap wave model to achieve realistic ocean wave effects. By adjusting parameters such as wave height, wavelength, and propagation direction, various states from calm sea surfaces to violent waves can be simulated.

[0261] 2. Tidal flooding: Using fluid dynamics simulation technology combined with high-precision terrain data, the process of water intrusion into land is realized. By adjusting the water level and flow rate, different degrees of seawater backflow and waterlogging can be simulated.

[0262] 3. Storm effect: The particle system simulates effects such as heavy rain and flying droplets, and the dynamic weather system adjusts parameters such as lighting and visibility to create a realistic typhoon environment.

[0263] 4. Facility destruction: Multiple levels of damage status are set for key facilities (such as breakwaters and buildings), and the destruction process can be dynamically displayed according to the stress conditions.

[0264] 5. Emergency response: The startup and operation process of emergency measures such as alarm systems, evacuation routes, and temporary shelters were simulated.

[0265] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for constructing a digital twin system for nearshore overtopping disasters, characterized in that: The following steps are involved: S10, obtaining multi-angle photos and elevation data of the target area, specifically, using a drone with oblique photography and elevation measurement functions to perform aerial survey to obtain multi-angle photos and elevation data of the target area; S20, constructing a three-dimensional real scene model based on the multi-angle photos and elevation data, including image preprocessing, feature extraction and matching, aerial triangulation, dense matching and point cloud generation, three-dimensional reconstruction, texture mapping and model optimization; S30, using real-time 3D creation tools for wave and hydrodynamic simulation, where JONSWAP spectrum is used for wave simulation and simplified continuity equation and momentum equation are used for hydrodynamic simulation; S40. Carry out real-time observation in the digital twin area or nearby areas, using buoys or bottom-mounted ADCP observation instruments to obtain real-time observation data, including wave height and water level; S50, transferring the real-time observation data to the digital twin system, so that the twin scene and the observation data are linked in real time; S60. Use high-resolution numerical models or artificial intelligence models to build wave models and storm surge models for the digital twin area; S70, using the ocean wave model and storm surge model to forecast ocean waves and storm surges, with a forecast validity period of no less than 72 hours and a forecast result time interval of no more than 3 hours; S80. Using the wave model and storm surge model, simulate historical wave and storm surge scenarios of no less than 2 years, historical typical wave and storm surge disaster scenarios, wave and storm surge disaster scenarios with different return periods, and wave and storm surge scenarios under different disaster reduction measures; S90, constructing a disaster prevention and mitigation scenario library based on the simulation results of the wave model and the storm surge model, and displaying disaster prevention and mitigation scenarios using marine environment simulation technology; Among them, the specific steps of using high-resolution numerical models to construct the wave model and storm surge model of the digital twin area include: determining the calculation domain and grid division, setting initial conditions and boundary conditions, debugging and verifying model parameters, and coupling the wave model and storm surge model; Among them, the specific steps of using artificial intelligence models to build wave models and storm surge models in digital twin areas include: collecting historical observation data and numerical simulation results, data preprocessing and feature engineering, training models and optimizing hyperparameters, verifying model performance and performing integrated learning.

2. The method for constructing a digital twin system for nearshore overtopping disasters according to claim 1, characterized in that: The ocean wave model includes an energy balance equation, a source term equation, a radiation stress equation and a bottom friction equation.

3. The method for constructing a digital twin system for nearshore overtopping disasters according to claim 2, characterized in that: The energy balance equation is used to describe the propagation and change of wave energy. The input variables include the wave energy spectrum and the wave propagation speed, and the output variable is the spatiotemporal change of the wave energy spectrum. The source term equation is used to describe the growth and dissipation of wave energy, the input variables include wind speed, water depth and wave energy spectrum, and the output variable is the growth and dissipation rate of wave energy; The radiation stress equation is used to calculate the horizontal force caused by waves, the input variables include wave energy spectrum and water depth, and the output variable is radiation stress; The bottom friction equation is used to calculate the energy dissipation of waves on the seabed, the input variables include wave parameters and bottom roughness, and the output variable is the bottom friction coefficient.

4. The method for constructing a digital twin system for nearshore overtopping disasters according to claim 2, characterized in that: The storm surge model includes a continuity equation, a momentum equation, a pressure equation, a Coriolis force equation and a boundary condition equation.

5. The method for constructing a digital twin system for nearshore overtopping disasters according to claim 4, characterized in that: The continuity equation is used to describe the conservation of water mass, the input variables include flow velocity and water depth, and the output variable is the rate of change of water level; The momentum equation is used to describe the movement of water bodies, the input variables include flow velocity, water depth and external force, and the output variable is the rate of change of flow velocity; The pressure equation is used to calculate the pressure distribution inside the water body, the input variables include water depth and density, and the output variable is pressure; The Coriolis force equation is used to describe the effect of the earth's rotation on the movement of water bodies. The input variables include flow velocity and latitude, and the output variable is the Coriolis force. The boundary condition equation is used to describe the hydrodynamic characteristics of the model boundary. The input variables include the boundary type and external forcing, and the output variable is the water level or flow velocity on the boundary.

6. The method for constructing a digital twin system for nearshore overtopping disasters according to claim 5, characterized in that: The high-resolution numerical model is specifically a data mapping relationship constructed based on a large amount of historical observation data and numerical simulation results, including the mapping relationship between wave characteristic parameters and environmental factors, and the mapping relationship between storm surge water levels and meteorological and ocean dynamic factors.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions, and when the program instructions are executed in a computer, they are used to execute the method for constructing a nearshore overtopping disaster digital twin system according to any one of claims 1 to 6.

8. A system for constructing a digital twin system for nearshore overtopping disasters, characterized in that: A computer-readable storage medium comprising the method of claim 7.

Images