A flood risk dynamic deduction simulation method based on digital twinning

Abstract

This invention provides a dynamic simulation method for flood risk based on digital twins. The method includes: acquiring basic multidimensional parameters of the flood control area; constructing a matching digital twin simulation base and performing hydrodynamic flow calculations within it to output the initial flood prediction situation; during the calculation process, creating dedicated retention state recording variables for each virtual drainage node within the base; accumulating accumulation count values ​​in the stack space based on the simulated inflow volume entering the node; performing downgrading and flow limiting operations on the preset drainage throughput bandwidth of the node accordingly; and continuing calculations according to the limited bandwidth; continuously monitoring the surface grid parameters corresponding to the node; triggering an asynchronous dredging update event when the set clearing and reset conditions are met, forcibly clearing the accumulation count values ​​in the stack space to zero and restoring the preset bandwidth; updating the prediction situation based on the restored drainage capacity and generating early warning commands. The method simulates the process of debris blockage and flow reset at drainage nodes, ensuring the continuity of the simulation calculation.

Description

Technical Field

[0001] This invention relates to the field of computer digital twin and flood control simulation technology, and in particular to a dynamic simulation method for flood risk based on digital twin. Background Technology

[0002] Currently, urban flood control and waterlogging risk simulation typically relies on fluid dynamics simulation technology and computer digital models. In the current flood risk simulation system, the computational model is mainly based on a static three-dimensional terrain elevation base and a fixed underground pipe network topology. It uses two-dimensional shallow water equations to solve for surface runoff and couples a one-dimensional pipe network model to solve for underground drainage processes. In real urban rainfall and waterlogging physical scenarios, surface runoff carries debris such as leaves and silt during its confluence and accumulates at drainage nodes due to interception by grids. This accumulation reduces the effective flow area of ​​the drainage nodes and increases local drainage resistance. With prolonged rainfall and changes in hydrological conditions, the accumulated debris will shift or disintegrate under the influence of buoyancy and dynamic shear forces, thus dynamically restoring the flow capacity of the drainage nodes.

[0003] Current fluid dynamics simulation systems offer various computational solutions to address the nonlinear changes in the state of drainage nodes. Some approaches directly introduce a solid-liquid multiphase flow calculation module into the underlying partial differential equations to simultaneously calculate the physical processes of debris movement and accumulation following the water flow. This approach can comprehensively recreate the details of solid-liquid multiphase mechanics, but it also increases the size of the computational matrix and the processor's floating-point computation load. Other methods update the flow capacity by pausing the computation process and reconstructing the mesh boundary conditions. This allows for intervention to adjust local flow boundaries at specific times, but it also divides the continuous simulation process into multiple discrete computation intervals. Still other simulation models, based on computational resource considerations, choose to set fixed drainage parameters for a unified solution throughout the entire cycle. The output of the flood evolution results reflects the flow field evolution under fixed drainage conditions. All these approaches have played a positive predictive role in specific flood control simulation scenarios.

[0004] Considering the dynamic alternation of drainage node states in real hydrological environments, how to simulate the evolution of drainage capacity caused by debris blockage and water scouring while maintaining the topological stability of the basic model's spatial grid and the unchanged structure of the hydrodynamic partial differential equations is a direction that those skilled in the art are focusing on and optimizing. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic simulation method for flood risk based on digital twins, in order to solve the problems pointed out in the background art.

[0006] This invention provides a method for dynamic simulation of flood risk based on digital twins, comprising the following steps: Obtain basic multidimensional parameters of the flood control area; A digital twin simulation platform matching the flood control area is constructed based on the aforementioned basic multidimensional parameters; Hydrodynamic flow calculations are performed in the digital twin simulation platform to output the initial flooding prediction situation; The flood risk dynamic simulation method based on digital twins also includes: During the hydrodynamic flow calculation, a dedicated variable for recording the state of stagnant matter is created for each virtual drainage node within the digital twin simulation base. Based on the simulated inflow volume entering each of the virtual drainage nodes, the accumulation count value is accumulated in the state record variable of the retained material; Based on the accumulation count value, the preset drainage throughput bandwidth of each virtual drainage node is downgraded and limited to obtain a limited drainage throughput bandwidth, and the hydrodynamic flow calculation is continued according to the limited drainage throughput bandwidth. Continuously monitor the surface grid parameters corresponding to each of the virtual drainage nodes; When it is determined that the surface grid parameters meet the set clearing and reset conditions, an asynchronous dredging and update event is triggered for the target virtual drainage node that meets the conditions; In response to the asynchronous dredging update event, the accumulation count value in the status record variable of the stagnant material corresponding to the target virtual drainage node is forcibly cleared to zero, the preset drainage throughput bandwidth is restored, and the initial flooding prediction status is updated based on the restored drainage capacity to generate a terminal early warning command.

[0007] Optionally, obtaining the basic multidimensional parameters of the flood control area includes: The surface elevation point cloud sequence and land feature material classification labels of the flood control area were extracted through the remote sensing mapping channel. Connect to the IoT hydrological sensing terminal to read rainfall time series and river boundary water level series; The surface elevation point cloud sequence, the land feature material classification label, the rainfall time series sequence, and the river boundary water level sequence are spatiotemporally aligned and combined to form the basic multidimensional parameters.

[0008] Optionally, the step of constructing a digital twin simulation base matching the flood control area based on the basic multidimensional parameters includes: The surface elevation point cloud sequence is used for grid subdivision to generate a spatial topological grid matrix; Based on the aforementioned land feature material classification labels, the basic roughness drag coefficient is mapped to each grid cell within the spatial topology grid matrix; In the spatial topology grid matrix, the specific coordinate positions of the corresponding underground pipe network inlets are marked, and the specific coordinate positions are instantiated into each of the virtual drainage nodes; The spatial topology grid matrix, the basic roughness resistance coefficient, and each of the virtual drainage nodes are integrated to solidify and generate the digital twin simulation base.

[0009] Optionally, the step of performing hydrodynamic flow calculations in the digital twin simulation base and outputting the initial flooding prediction situation includes: The rainfall time series and the river boundary water level series are used as driving sources and injected into the digital twin simulation base. Set the evolution time step, and drive the water mass exchange and momentum transfer calculations between grid cells according to the evolution time step; Extract the simulated water depth and simulated flow velocity of each of the grid cells in each evolution cycle, and screen out the set of grid cells whose simulated water depth exceeds the safety threshold; The initial flooding prediction situation is generated based on the distribution range of the over-limit grid set.

[0010] Optionally, the step of creating dedicated variables for recording the state of retained waste at each virtual drainage node within the digital twin simulation base includes: In the system memory, an independent memory data node is allocated for each of the aforementioned virtual drainage nodes; The independent memory data nodes are encapsulated as the status record variables of the retained material corresponding to each of the virtual drainage nodes; The variable for recording the state of the retained material is initially set to a blank state, and the initial base of the accumulation count value is set to zero.

[0011] Optionally, the step of accumulating the accumulation count value into the retention state record variable based on the simulated inflow volume entering each of the virtual drainage nodes includes: Within each evolution cycle, the simulated inflow volume input to each of the virtual drainage nodes is extracted from the hydrodynamic flow calculation. Retrieve the pre-set virtual impurity content ratio coefficient; Multiply the simulated inflow volume by the virtual impurity content ratio coefficient to obtain the calculated increment for the current period; The calculated increment is added to the previous recorded value of the residual state record variable to update the current accumulation count value.

[0012] Optionally, the step of performing a degradation and rate limiting operation on the preset drainage throughput bandwidth of each of the virtual drainage nodes based on the accumulation count value to obtain a limited drainage throughput bandwidth includes: Set the maximum allowed stacking capacity and the minimum guaranteed bandwidth ratio; Calculate the saturation ratio of the stack count value to the maximum allowable stacking capacity limit; Subtract the saturation ratio from the full load baseline ratio to obtain the current available performance ratio; Compare the current available performance ratio with the minimum guaranteed bandwidth ratio, and select the larger value as the effective quota ratio; Multiply the effective quota ratio by the preset drainage throughput bandwidth to output the limited drainage throughput bandwidth.

[0013] Optionally, the continuous monitoring of surface grid parameters corresponding to each of the virtual drainage nodes; when it is determined that the surface grid parameters meet the set clearing and reset conditions, an asynchronous dredging update event is triggered for the target virtual drainage node that meets the conditions, including: Lock the surface flow-bearing grid directly above each of the virtual drainage nodes; The real-time simulated water depth and real-time water surface velocity of the surface flow-bearing grid are obtained and used as the parameters of the surface grid; Preset buoyancy drift depth threshold and turbulent scouring velocity threshold; When it is determined that the real-time simulated water depth of the corresponding target virtual drainage node exceeds the buoyancy drift depth threshold, or when it is determined that the real-time water surface velocity exceeds the turbulent scouring velocity threshold, the clearing and reset condition is confirmed to be met. A high-priority trigger command is sent to the main control scheduler, thereby triggering the asynchronous unblocking update event.

[0014] Optionally, the step of forcibly clearing the accumulation count value in the retention status record variable corresponding to the target virtual drainage node and restoring the preset drainage throughput bandwidth in response to the asynchronous dredging update event includes: Suspend the current downgrade flow restriction update command of the target virtual drainage node; Write a zeroing and reset flag to the independent memory data node corresponding to the target virtual drainage node to complete the forced zeroing action; Remove the parameter binding of the restricted drainage throughput bandwidth and redirect the overcurrent capacity parameter of the target virtual drainage node to the preset drainage throughput bandwidth.

[0015] Optionally, updating the initial flood prediction situation based on the restored drainage capacity and generating terminal early warning instructions includes: Using the restored preset drainage throughput bandwidth, the water accumulated in the surface runoff grid corresponding to the target virtual drainage node is guided to be discharged to the target virtual drainage node below. In subsequent evolution cycles, the water level fluctuation trajectory of the entire grid is re-tracked to generate a corrected dynamic flooding situation map; Extract the coordinates of target landmarks in a state of severe flooding from the dynamic flooding situation map; The target landmark coordinates are packaged into the terminal warning command and pushed to the flood control command and dispatch terminal via wireless communication link for high-brightness flashing display.

[0016] The present invention has achieved the following beneficial effects: This invention effectively maps the dynamic process of debris accumulation leading to reduced drainage capacity in real-world scenarios by establishing dedicated retention state record variables for each virtual drainage node during hydrodynamic flow calculations. Based on the accumulated and deposited count values ​​of simulated inflow entering the node, and using these count values ​​to downgrade and limit the preset drainage throughput bandwidth, the system continuously monitors surface grid parameters. When preset clearing and reset conditions are met, an asynchronous dredging update event is triggered, forcibly clearing the deposited count values ​​in the stack space and restoring the preset drainage throughput bandwidth. This recreates the physical discharge reset phenomenon of debris being washed away and dispersed under specific water flow conditions. By utilizing the dynamic interaction between asynchronous dredging update events and state record variables, this invention reduces the enormous computational overhead caused by introducing complex multiphase flow physical equations while maintaining the basic computational grid topology and partial differential equation structure. It also effectively reduces the interruption of calculations caused by grid boundary reconstruction during system downtime, resulting in output dynamic flooding situation maps that closely match real-world physical hydrological evolution patterns, providing objective and consistent data support for flood control command and dispatch.

[0017] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating a flood risk dynamic simulation method based on digital twins in an embodiment of the present invention. Figure 2 This is another flowchart illustrating a flood risk dynamic simulation method based on digital twins in an embodiment of the present invention. Detailed Implementation

[0020] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0021] Urban flood control and waterlogging risk simulation typically relies on fluid dynamics simulation techniques. In current flood risk simulation methods, the computational model is primarily based on a static three-dimensional terrain elevation base and a fixed underground pipe network topology. Two-dimensional shallow water equations are used to solve for surface runoff, coupled with a one-dimensional pipe network model to solve for underground drainage processes. In actual urban rainfall and waterlogging scenarios, surface runoff carries suspended solids and bedloads such as leaves, household waste, and silt during its runoff flow. When this runoff flows into drainage nodes such as storm drain grates, manholes, or pump station inlets, the solid debris carried in the runoff is intercepted by the grates and physically accumulates at the drainage nodes. As the rainfall duration increases, the accumulation leads to a dynamic decrease in the effective flow cross-sectional area of ​​the drainage nodes. Simultaneously, when the local water depth and surface velocity above the drainage node reach specific physical thresholds, the buoyancy and dynamic shear force generated by the water flow exceed the weight and frictional resistance of the accumulated material, causing displacement or collapse of the accumulated material, and the flow capacity of the drainage node subsequently recovers.

[0022] Existing fluid dynamics simulation systems, when dealing with the nonlinear changes in the state of drainage nodes, increase the degrees of freedom of the computational matrix and the floating-point operations and memory bandwidth usage of the central processing unit if they directly introduce solid-liquid multiphase flow calculations into the underlying partial differential equations to simulate the movement and stacking of debris following the water flow. If the technique of halting the system to reconstruct the grid boundary conditions and recalculate the entire watershed is used, the simulation process will be interrupted. If the flow restriction and reset processes caused by debris blockage are not considered in the simulation calculation, the simulated water depth output by the system will deviate from the actual water depth in the physical environment. This application provides a flood risk dynamic simulation method based on digital twins. Utilizing the computer's underlying memory stack allocation mechanism and asynchronous hardware interrupt mechanism, while maintaining the computational grid topology and partial differential equation structure, it simulates the blockage degradation and flow recovery states in the physical environment through memory address operations and pointer mapping.

[0023] Reference Figure 1 As shown in the figure, this application provides a method for dynamic simulation of flood risk based on digital twins, which includes the following steps: Step S110: Obtain the basic multidimensional parameters of the flood control area.

[0024] The fundamental multidimensional parameters of the flood control area are obtained to construct the hydrodynamic computation grid, assign physical resistance properties, and set the initial and boundary conditions for numerical solutions. These fundamental multidimensional parameters include geometrical elevation coordinates describing the surface topography of the flood control area, material classification features characterizing the frictional properties of the subsurface, and hydrological and meteorological sequence parameters serving as the input source for fluid computation.

[0025] Specifically, the remote sensing mapping channel extracts the surface elevation point cloud sequence and ground feature material classification labels of the flood control area. The flood control area is a geographic spatial range defined in the geographic information system using a polygon coordinate set. The remote sensing mapping channel includes an aircraft platform equipped with airborne lidar equipment and a data link interface for receiving optical remote sensing satellite imagery. The airborne laser emission controller emits laser pulses towards the ground surface according to a set pulse repetition frequency, and a photoelectric receiver receives the laser echo signals reflected from surface objects. The airborne data processing unit calculates the time difference between the emission and reception of the laser pulses, and combines the attitude angles and three-dimensional position data output by the inertial navigation system and the global positioning system to generate a raw set of reflected echo data containing a discrete three-dimensional coordinate matrix of the flood control area.

[0026] Furthermore, the original reflected echo data set includes reflection coordinate nodes of non-ground objects such as buildings, vegetation, and cables. The central processing unit calls the cloth simulation surface filtering algorithm module to filter the original reflected echo data set. The cloth simulation surface filtering algorithm module constructs a virtual 3D point cloud bounding box in memory and covers it with a virtual mass spring cloth mesh. It calculates the falling displacement of the cloth nodes under the action of set gravity parameters, as well as the collision response generated when they come into contact with the point cloud data nodes. After multiple iterative calculations, the set of points representing the exposed ground surface is extracted. After removing floating noise points, the set of exposed ground surface points is serialized according to a preset data structure, and the output is the ground elevation point cloud sequence. Each data node in the ground elevation point cloud sequence independently contains floating-point numbers of horizontal plane coordinates, floating-point numbers of vertical plane coordinates, and floating-point values ​​of altitude.

[0027] Synchronously, the system acquires multispectral orthophoto matrix data covering the flood control area via an optical data link interface. The multispectral orthophoto matrix data contains a two-dimensional array of spectral reflectance values ​​for the visible and near-infrared bands. The central processing unit loads the two-dimensional array of spectral reflectance values ​​into random access memory and calls a pre-configured support vector machine (SVM) classification algorithm model. The SVM classification algorithm model extracts multidimensional spectral feature vectors from the pixels in the multispectral orthophoto matrix and calculates the normalized vegetation index (NDI) and the normalized water index. Specifically, the formula for calculating the NDI is: The parameters in this formula are defined as follows: This represents the calculated normalized vegetation index (dimensionless). The near-infrared band reflectance value (dimensionless) is extracted from the multispectral orthophoto matrix. This represents the dimensionless reflectance value in the red light band. The formula for calculating the normalized water index is: The parameters in this formula are defined as follows: This represents the calculated normalized water index (dimensionless). Represents the reflectivity value in the green light band (dimensionless). This represents the near-infrared band reflectance value (dimensionless). Based on the extracted feature vectors, the support vector machine classification algorithm performs hyperplane decision boundary operations, dividing surface pixels into discrete category sets. Specific categories include impermeable paving, concrete buildings, permeable soil, vegetation, and natural water bodies. The system assigns an integer attribute code identifier to each category set, encapsulates it into a two-dimensional raster matrix structure with spatial reference features, and generates the land cover material classification label.

[0028] Specifically, the system connects to an IoT hydrological sensing terminal to read rainfall time-series and river boundary water level sequences. The IoT hydrological sensing terminal includes sensing and communication nodes deployed at water catchment nodes, drainage pumping station outlets, and confluence sections of main and tributary streams in the flood control area. The system server is configured with a transmission control protocol port to receive wireless data packets reported by the tipping bucket rain gauge terminal. The microcontroller unit of the tipping bucket rain gauge terminal records the number of pulse signals triggered by the mechanical tipping action and converts its analog-to-digital conversion into an instantaneous precipitation depth floating-point scalar. The system server collects data packets according to a set time period, parses the payload to extract the media access control address, timestamp parameters, and precipitation floating-point scalar. The extracted data is encapsulated in memory in ascending order of time into a one-dimensional dynamic array structure, establishing the rainfall time-series. Simultaneously, the system server receives data packets from the microwave radar water level gauge terminal. The microwave radar water level gauge terminal uses a digital signal processing chip to calculate the frequency difference between the echo signal and the transmitted signal of the frequency-modulated continuous wave, and calculates the elevation distance from the antenna to the water surface. The system analyzes the radar message and generates a sequence array that reflects the fluctuation of the hydrostatic pressure parameters at the model boundary over time, which is then established as the water level sequence at the river boundary.

[0029] Furthermore, the surface elevation point cloud sequence, the land cover material classification labels, the rainfall time series sequence, and the river boundary water level sequence are spatiotemporally aligned to form the basic multidimensional parameters. At the spatial data alignment processing level, the system uses the 2000 National Geodetic Coordinate System as a reference system and calls the coordinate transformation model matrix to perform matrix multiplication operations. The raster pixel coordinate array of the land cover material classification labels and the latitude and longitude coordinate data of each hydrological sensing terminal are translated, rotated, and scaled, projected onto the three-dimensional orthogonal coordinate system space where the surface elevation point cloud sequence resides. At the temporal data alignment processing level, the time starting floating-point number and discretization time step constant of the hydrodynamic numerical simulation calculation engine are set. The system iterates through and reads the timestamp nodes in the rainfall time series sequence and the river boundary water level sequence, calls the cubic spline function interpolation algorithm, and maps and interpolates the asynchronously sampled sensor data sequence to the discretization time step equally divided nodes of the calculation engine. After spatial coordinate transformation and time interpolation, the aligned data is serialized and encapsulated in the physical memory area into a multi-dimensional nested structure object with associated pointers, which is then combined to form the basic multi-dimensional parameters.

[0030] Step S120: Construct a digital twin simulation base that matches the flood control area based on the basic multidimensional parameters.

[0031] Specifically, the surface elevation point cloud sequence is used for mesh generation to produce a spatial topological mesh matrix. The mesh generation module of the central processing unit reads the surface elevation point cloud sequence structure residing in the memory segment. Using the three-dimensional coordinate discrete points in the structure as boundary constraints, the Delaunay triangle spatial partitioning algorithm component is invoked. The Delaunay triangle spatial partitioning algorithm component executes the empty circumscribed circle rule detection logic on the two-dimensional planar projection space, connecting the discrete coordinate points to form non-intersecting continuous triangular mesh polygon faces. After the initial mesh connection is generated, the system calls the gradient operation function to calculate the normal elevation difference gradient floating-point scalar between adjacent mesh coordinate control nodes. The formula for calculating the normal elevation difference gradient floating-point scalar is: The parameters in this formula are defined as follows: The output is a dimensionless floating-point scalar representing the normal elevation gradient. The absolute elevation value (in meters) of the first grid coordinate control node; The absolute elevation value (in meters) of the adjacent second grid coordinate control node; This represents the absolute value of the absolute elevation difference between the two nodes mentioned above; This represents the linear distance (in meters) between the two nodes projected onto the two-dimensional plane. For regions where the elevation gradient scalar is greater than the first preset gradient constant, the system executes a local mesh geometric subdivision instruction, inserting a calculation control node at the midpoint of the edge of the original triangular mesh polygon and re-performing the subdivision calculation. For regions where the elevation gradient scalar is less than the second preset gradient constant, the system executes a mesh edge fusion instruction. As an example, to balance terrain fidelity and computational power, the first preset gradient constant is set to a range of 0.15 to 0.30, and the second preset gradient constant is set to a range of 0.02 to 0.05. After mesh subdivision, the system assigns an unsigned long integer index number to each generated triangular mesh cell and calculates the three-dimensional coordinates of the centroid and the planar projected area constant of each mesh cell according to the geometric center formula. The system traverses the geometric edge data structure of all meshes, extracts the index codes of adjacent mesh cells sharing the same spatial edge, and establishes a directed graph of spatial topological adjacency relationships in memory based on a bidirectional address pointer linked list structure. The spatial topology grid matrix is ​​generated by formatting and storing the grid geometric centroid parameters, absolute elevation center values, and spatial topological adjacency relationship directed graph as a compressed sparse row data format matrix.

[0032] Furthermore, based on the aforementioned land cover material classification labels, a basic roughness drag coefficient is mapped to each grid cell within the spatial topological grid matrix. The system process reads a preset Manning roughness coefficient lookup hash dictionary table from memory. The hash dictionary table stores key-value pair relationships between category attribute codes and empirical Manning constant floating-point values. For example, the empirical Manning constant floating-point value (whose physical unit dimensions are: The value ranges are set according to the material of the terrain: 0.013 to 0.015 for impermeable pavement, 0.012 to 0.014 for concrete buildings, 0.030 to 0.035 for permeable soil, 0.040 to 0.050 for vegetation, and 0.025 to 0.030 for natural water bodies. The system initiates a concurrent spatial query retrieval program, traversing each discrete grid cell within the spatial topological grid matrix. For the target grid cell, its geometric centroid coordinate parameters are extracted and substituted into the coordinate group of the two-dimensional layer of the terrain material classification label. The computational geometry library is then called to execute a ray intersection detection algorithm to determine the polygon physical attribute code to which the geometric centroid coordinates of the grid belong. The system uses the attribute code to perform an addressing and reading operation in the hash dictionary table to extract the mapped Manning constant floating-point value. Through a memory write instruction, the Manning constant floating-point value is assigned to the resistance attribute parameter address space within the target grid cell structure as the foundation roughness resistance coefficient.

[0033] Specifically, the specific coordinate positions of the corresponding underground pipe network inlets are marked in the spatial topology grid matrix, and these specific coordinate positions are instantiated as each of the virtual drainage nodes. The system input / output module reads the municipal drainage pipe network vector file and parses and extracts attribute fields as two-dimensional or three-dimensional geographic center coordinate point data columns for rainwater grates, manhole openings, and pump station inlets. The system central processing unit executes a spatial distance search algorithm to calculate the straight-line distance between each extracted point data and the geometric centroid coordinates of all grid cells in the spatial topology grid matrix. The target grid cell with the minimum calculated distance is marked as the specific coordinate position with boundary connectivity. Based on the preset drainage node class code template, the system operation kernel calls the constructor in the heap memory operation area to dynamically create corresponding logical operation structure object instances for each marked specific coordinate position, and instantiates them as each of the virtual drainage nodes. The system allocates independent member control variable spaces for each virtual drainage node, and combines this with the pipe diameter and flow section design parameters recorded in the associated pipe network vector file. Using hydraulic formulas, it calculates the allowable liquid volume capacity per unit calculation time for each node under unobstructed conditions, and configures this value as a preset drainage throughput bandwidth double-precision constant attribute field. Specifically, the hydraulic formula uses the orifice free outflow formula, whose algebraic expression is: The parameters in this formula are defined as follows: This is the scalar of the liquid volume capacity allowed to pass through each node per unit calculation time under unobstructed conditions, i.e., the preset drainage throughput bandwidth (unit: cubic meters / second). The empirical coefficient for pipe flow rate (a dimensionless constant, with a value range of 0.60 to 0.65). The physical opening area constant (unit: square meters) is extracted based on the pipe diameter and flow section design parameters. The acceleration due to gravity is constant (the standard value is 9.8 m / s²). The maximum operating head parameter for flood control design of the node is the maximum allowable water accumulation depth (unit: meters) of the drainage node.

[0034] Furthermore, the spatial topology mesh matrix, the basic roughness drag coefficients, and each of the virtual drainage nodes are integrated to solidify and generate the digital twin simulation base. The system memory management module executes memory block scheduling and rearrangement instructions. The memory bus address block copy and rearrangement operation is performed on the allocated mesh geometry topology structure array, the basic roughness drag coefficient list, and the set of memory blocks of independently instantiated virtual drainage node objects. After each core model data object is transferred and copied to a contiguous memory block space, the system control kernel modifies the virtual memory page access permission descriptor of that contiguous memory block space from read-write to read-only. The memory-combined data snapshot entity in read-only protection is encapsulated to form the digital twin simulation base object participating in the simulation calculation flow.

[0035] Step S130: Perform hydrodynamic flow calculations in the digital twin simulation base and output the initial flooding prediction situation.

[0036] Hydrodynamic flow calculation is a process in which the system uses the discrete integral numerical calculation algorithm of fluid mechanics to solve the two-dimensional shallow water continuity equation and momentum equation in finite volume. It is used to numerically simulate the diffusion, water accumulation and water receding evolution of non-steady flow liquids.

[0037] Specifically, the rainfall time series and the river boundary water level series are used as driving sources and injected into the digital twin simulation base. The system control process activates the evolution system timer. Before the start of the discrete integral calculation loop iteration cycle, the system command reads the time base value variable currently output by the timer. Based on this time base value variable, the system performs a search in the one-dimensional dynamic array queue of the rainfall time series to extract the instantaneous precipitation intensity double-precision floating-point scalar parameter value aligned with the current moment. The system calls the vectorized parallel matrix operation instruction set to convert the instantaneous precipitation intensity double-precision floating-point scalar value into a volume change rate and sets it as the volume source term coefficient parameter in the mass balance equation operation. Specifically, the formula for converting to a volume change rate is: The parameters in this formula are defined as follows: The volume source term coefficient parameters for the mass balance equation calculation, i.e. the calculated volume change rate (unit: cubic meters / second). The instantaneous precipitation intensity is converted to standard SI units as a double-precision floating-point scalar value (unit: m / s). This represents the constant of the planar projected area (unit: square meters) corresponding to the specific grid cell. Through parallel accumulation instructions, the calculated mass is injected into the water volume data recording unit of the exposed grid cell within the digital twin simulation base. Simultaneously, based on the current time reference value variable, the water level elevation scalar constant value under the associated time scale is extracted from the water level sequence at the river boundary. The system memory bus uses direct memory mapping overwrite instructions to overwrite the elevation scalar value into the water depth variable memory address area held within the boundary grid cell of the digital twin simulation base, setting it as an open water boundary condition.

[0038] Furthermore, an evolution time step is set, and the water mass exchange and momentum transfer calculations between grid cells are driven according to the evolution time step. Before the start of each iteration calculation differential cycle, the system main control process scans the minimum spatial span floating-point extreme value constant of the global computational grid nodes and the maximum water flow velocity vector amplitude variable calculated at the end of the previous calculation cycle. These parameters are substituted into the Courant-Friedrich-Lévy (CFL) dimensionless stability critical condition formula for algebraic calculation, and the current batch time interval value that satisfies the system convergence requirement is output. The Courant-Friedrich-Lévy (CFL) dimensionless stability critical condition formula that satisfies the system convergence requirement is specifically as follows: The parameters in this formula are defined as follows: This represents the current batch time interval value, i.e., the unified evolution time step across the entire domain (unit: seconds). The preset Courant number constant (its value is set between 0.5 and 0.9 to ensure that the explicit partial differential algorithm does not diverge). is the minimum spatial span floating-point extreme value constant (unit: meters) among the nodes of the global grid. This is the global maximum water flow velocity vector amplitude variable (unit: m / s) obtained at the end of the previous calculation period. The acceleration due to gravity is constant (the standard value is 9.8 m / s²). The instantaneous local water depth variable (in meters) corresponds to the maximum flow velocity grid cell. A global comparison extracts the minimum lower boundary control double-precision floating-point value, which is set as the unified evolution time step. The system applies the finite volume numerical discretization rule to perform discrete solutions for the two-dimensional planar flow field. For pairs of independent computational cells in adjacent triangular grids with a common boundary, the Riemann flux operator solver module is invoked to calculate the transient flow rate flux transport rate across the common boundary. The system algorithm module uses the combined force of the head pressure gradient formed by the difference in elevation between the substrate surface and the liquid surface in adjacent grids, retained from the previous calculation time, to calculate the momentum equation. The driving terms of the momentum equation's algebraic polynomial include the differential term of the head pressure drop gradient, the Coriolis force influence component, and the substrate friction drag dissipation term. The calculation parameters for the substrate friction dissipation term are extracted from the constant value of the foundation roughness drag coefficient bound in memory for specific grid cells, and substituted into the friction shear stress calculation formula to obtain the friction coefficient component result. Specifically, the formula for calculating frictional shear stress is based on Manning's law in two-dimensional fluid dynamics, and its specific expression is as follows: The physical meaning of each parameter in this formula is defined as follows: The vector of frictional shear stress calculated from the bottom layer (unit: Pascal or Newton / m²). The preset physical density constant of water (the system setting is to take the standard freshwater density value of 1000 kg / m³). The acceleration due to gravity is constant (the standard value is 9.8 m / s²). The basic roughness drag coefficient constant value (unit: ) is bound and mounted in memory for a specific mesh cell. ); The instantaneous two-dimensional water flow velocity vector (unit: m / s) calculated for the current grid cell; It is the absolute magnitude scalar of the instantaneous water flow velocity vector (unit: m / s); This represents the instantaneous simulated local water depth parameters (in meters) for the current grid cell. As the evolution time step progresses, the distribution of water mass and the water flow migration state within the digital twin three-dimensional spatial topological grid matrix are overwritten and updated.

[0039] Specifically, the simulated water depth and simulated flow velocity of each grid cell in each evolution cycle are extracted, and the set of grid cells whose simulated water depth exceeds the safety threshold is screened out. A monitoring thread component is configured in the system operating platform. When the calculation of the integral loop iteration reaches the configured snapshot extraction time point, the data reading thread is activated and executed by the scheduler. An independent data reading thread traverses the memory base address segment of the structure control block at the bottom layer of the global grid entity cell, extracting the two-dimensional simulated local water depth double-precision floating-point scalar value and the simulated flow velocity vector amplitude array stored in the memory data segment. A list of floating-point constant setting parameters representing the safe passage restriction thresholds for different road types resides in the system's main memory. As an example, for motor vehicle lanes, the safe passage restriction threshold floating-point constant is set to 0.15 meters to 0.20 meters; for non-motor vehicle lanes and sidewalks, the safe passage restriction threshold floating-point constant is set to 0.10 meters to 0.15 meters. The processor sends the extracted simulated water depth data scalars from each independent grid cell to the data comparison and judgment logic module, performing a digital Boolean comparison operation with a pre-set safety threshold floating-point constant. When the logic comparison module outputs a feedback result indicating that the judgment condition is met, i.e., the current simulated water depth value of the judgment grid is greater than the tolerance failure safety threshold constant, the system determines that the independent grid has entered a state of waterlogging blockage. The detection thread concurrently extracts the unique identification index number sequence, three-dimensional coordinate data, and current actual water depth value and flow velocity scalar parameter variables of the grid cell that has exceeded the boundary judgment, and packages them into an event object record. The system performs a full matrix scan and pushes the object record entities that meet the boundary conditions into a dynamically linked list queue allocated in the shared heap memory area, generating a total cluster of individual grids containing records that triggered the safety threshold, constituting the set of grid cells exceeding the limit.

[0040] The initial flood prediction situation is generated based on the distribution range of the over-limit grid set. The system extracts the three-dimensional coordinate projection data and simulated water depth parameters recorded for each object in the over-limit grid set. The system calls the graphics rendering module and allocates a two-dimensional frame buffer in the video memory operation area. The three-dimensional coordinate data is mapped to the corresponding pixel array in the two-dimensional frame buffer according to the spatial resolution, and different color channel values ​​are assigned to the pixels according to the simulated water depth scalar value. After smoothing processing, a two-dimensional color raster image layer representing the spatial distribution of flood depth is generated, superimposed on the basic geographic map of the flood control area, and output as the initial flood prediction situation.

[0041] Furthermore, referring to Figure 2 As shown, the method provided in this embodiment further includes: Step S140: During the hydrodynamic flow calculation, a dedicated variable for recording the state of stagnant matter is created for each virtual drainage node within the digital twin simulation base.

[0042] Specifically, during the system memory allocation cycle, the memory management unit detects the instantiation process of each of the virtual drainage nodes. The memory management unit calls the system-level interface to send an allocation instruction to the core stack management module of the operating system. The management module of the operating system allocates an independent memory data node for each of the virtual drainage nodes in the dynamic heap area of ​​the system main memory. This independent memory data node has a preset byte alignment format and contiguous physical space. The system logic processor then declares this memory data node in the data structure as the residual object status record variable corresponding to each of the virtual drainage nodes. The system writes a binary zero-level signal to the physical memory by performing a write operation on this memory data node, setting the residual object status record variable to an initial blank state. Simultaneously, the system internally declares a double-precision floating-point variable as the accumulation measurement base and assigns it a value of zero, completing the operation of setting the initial base of the accumulation count value to zero.

[0043] Step S150: Based on the simulated inflow volume entering each of the virtual drainage nodes, accumulate the accumulation count value into the retention status record variable.

[0044] Within each evolution cycle of the hydrodynamic flow calculation, the central processing unit (CPU) extracts the simulated inflow volume input to each of the virtual drainage nodes via the memory bus. As the internal system clock of the digital twin simulation base increments, the hydrodynamic integration module completes equation iteration upon reaching the computation node. The memory management unit instructs the processor to read the fluid pumping flux calculation results executed by each of the virtual drainage nodes. The arithmetic logic unit reads the double-precision floating-point value of the liquid fluid volume transferred from the topologically adjacent surface grid to the pipeline boundary within the current time step span by the fluid pumping flux operator. The controller stores this double-precision floating-point value in the system data cache in data format, declaring it as the simulated inflow volume.

[0045] Subsequently, the central processing unit's data scheduling module retrieves the pre-set virtual impurity content percentage coefficient. The storage controller initiates a data read request to the system's persistent storage area. The data segment stores a static model configuration file table, which stores floating-point constants representing the volume concentration of solid floating matter carried by surface runoff. Based on the geographic coordinate attributes of the specific virtual drainage node, the central processing unit executes a conditional matching instruction to extract the mapped percentage double-precision floating-point value from the configuration file table. The microprocessor pushes the extracted percentage value into the system's computational cache as the virtual impurity content percentage coefficient. For example, based on historical surface runoff impurity test statistics for the flood control area, the value range of the virtual impurity content percentage coefficient (dimensionless volume percentage) is set to 0.005% to 0.025%.

[0046] Next, the system multiplies the simulated inflow volume by the virtual impurity content ratio coefficient to obtain the calculated increment for the current period. The system extracts the simulated inflow volume and the virtual impurity content ratio coefficient, and executes a double-precision floating-point multiplication instruction. It calculates and outputs a real scalar value representing the virtual solid debris space volume that arrives at the boundary of the virtual drainage node with the water flow within the current evolution time step, and allocates it to the dynamic variable area, establishing it as the calculated increment. Then, the system adds the calculated increment to the previous recorded value of the retained material state record variable, updating the current accumulation count value. The system initiates an accumulation operation instruction to the retained material state record variable specifically mapped to the virtual drainage node. It reads the previous accumulated floating-point value currently residing in the retained material state record variable, and sums the previous accumulated floating-point value with the calculated increment. After outputting the double-precision floating-point summation result, it uses a variable assignment instruction to update the retained material state record variable with this value. By using variable accumulation and update operations, the equivalent mapping calculation was performed to determine the process by which the convergence of surface water flow leads to an increase in the volume of solid debris accumulation at the drainage outlet.

[0047] Step S160: Based on the accumulation count value, perform a downgrade and flow-limiting operation on the preset drainage throughput bandwidth of each virtual drainage node to obtain a limited drainage throughput bandwidth, and continue to perform the hydrodynamic flow calculation according to the limited drainage throughput bandwidth.

[0048] Specifically, a maximum allowable accumulation capacity and a minimum guaranteed bandwidth ratio are set. The central processing unit (CPU) allocates a byte address space within the system's constant memory physical declaration area, declaring and writing two double-precision floating-point configuration parameters. The first parameter is the upper limit of the volume of debris that can be accommodated when the drainage metal grating is filled with debris, defined as the maximum allowable accumulation capacity. The second parameter is a constant representing the flow rate ratio of water seepage under blockage conditions, defined as the minimum guaranteed bandwidth ratio. Both parameters are fixed in a read-only memory segment during the process loading phase. Specifically, the algebraic calculation formula for the maximum allowable accumulation capacity is: The parameters in this formula are defined as follows: This is the maximum allowable stacking capacity parameter (unit: cubic meters). Cross-sectional area of ​​the physical foundation of the sinking water collection well for the target virtual drainage node (unit: square meters); The preset thickness constant for grid debris interception (unit: meters). Furthermore, the minimum guaranteed bandwidth ratio characterizes the pore limit seepage capacity under clogging conditions, and its value is preferably a dimensionless scalar between 0.05 and 0.15, thereby forcibly preventing the discrete solution from diverging and crashing due to the bottom denominator flux of the computer reaching zero.

[0049] Subsequently, the arithmetic logic unit calculates the saturation ratio of the accumulated count value to the maximum allowable accumulated capacity limit. The instruction scheduling center retrieves the accumulated count value after completing the overwrite instruction. The accumulated count value is input into the calculation parameter buffer of the division operation logic module, and simultaneously, the maximum allowable accumulated capacity limit parameter is retrieved from the constant memory area and input into the calculation parameter buffer of the division operation logic module. The division operation logic module executes the number division operation instruction and calculates the dimensionless floating-point quotient variable whose value range is defined in the interval between zero and one. The dimensionless floating-point quotient variable quantifies the full load ratio of the current drainage inlet occupied by impurities, and the system assigns the dimensionless floating-point quotient variable the value of the saturation ratio.

[0050] Next, the microprocessor subtracts the saturation ratio from the full-load baseline ratio to obtain the current available performance ratio. The system's underlying logic pre-sets a double-precision floating-point constant of 1, representing the unobstructed state of the drainage nodes, as the full-load baseline ratio. The subtraction logic unit receives the full-load baseline ratio and the calculated saturation ratio, performing algebraic subtraction. A decay coefficient scalar that decreases over time is calculated, and the system memory allocation module stores this decay coefficient scalar in the dynamic cache stack, naming it the current available performance ratio.

[0051] Subsequently, the system compares the currently available performance ratio with the minimum guaranteed bandwidth ratio, selecting the larger value as the effective quota ratio. To avoid divergence in the solution of the partial differential equation system caused by calculating extremely small output flows, the central processing core sends the currently available performance ratio and the stored minimum guaranteed bandwidth ratio into the extreme value comparison logic unit to perform maximum value comparison logic operations. The digital comparator outputs the parameter variable with the larger value of the two parameters to the system bus. Based on the comparator feedback, the system control flow uses data transmission instructions to assign the parameter variable with the larger value to the dynamic variable, establishing it as the effective quota ratio.

[0052] Furthermore, the system multiplies the effective quota ratio by the preset drainage throughput bandwidth to output the limited drainage throughput bandwidth. The data processing and calculation core obtains the double-precision constant value of the preset drainage throughput bandwidth. The data processing and calculation core performs an algebraic multiplication operation on the preset drainage throughput bandwidth double-precision constant value and the dynamically calculated effective quota ratio. The data processing and calculation core outputs a flow discharge floating-point index after downgrade processing, which is defined as the limited drainage throughput bandwidth. Subsequently, the underlying control flow management process initiates a redirection overwrite instruction to the memory address block. The management process modifies and maps the boundary flow capacity constant pointer of the target virtual drainage node participating in the algorithm in the next evolution cycle from the preset drainage throughput bandwidth physical address to the storage address of the limited drainage throughput bandwidth. The discrete integral solving engine continues to execute the hydrodynamic flow calculation according to the replaced boundary flow condition parameters. The parameter replacement logic calculates and maps the flow boundary attributes that cause capacity degradation due to channel physical blockage while maintaining the macroscopic topology mesh unchanged.

[0053] Step S170: Continuously monitor the surface grid parameters corresponding to each of the virtual drainage nodes; when it is determined that the surface grid parameters meet the set clearing and reset conditions, trigger an asynchronous dredging update event for the target virtual drainage node that meets the conditions.

[0054] Specifically, the system's spatial indexing and addressing module locks onto the surface flow-bearing grid directly above each of the virtual drainage nodes. During the topology assembly phase of the digital twin simulation base, the system performs a query operation based on spatial vertical coordinate relationships. The system establishes a one-way physical address mapping link pointer between the structure of each virtual drainage node object responsible for pumping operations and the solid triangular mesh unit that coincides with the spatial normal and carries the surface water. The solid triangular mesh unit with the one-way physical address mapping link pointer is locked as the surface flow-bearing grid responsible for transporting water downwards.

[0055] Subsequently, the background asynchronous probe monitoring multi-thread acquires the real-time simulated water depth and real-time surface velocity of the surface runoff-bearing grid, which are used as the surface grid parameters. The background asynchronous probe monitoring multi-thread is activated at the moment when the fluid simulation model outputs a snapshot of the physical field data in each iteration cycle. The background monitoring multi-thread penetrates into the variable storage memory segment of the surface runoff-bearing grid through a memory access channel. It extracts the local inundation depth (double-precision floating-point scalar) and the horizontal water flow velocity vector magnitude written iteratively by the engine. The background monitoring multi-thread packages the local inundation depth (double-precision floating-point scalar) and the horizontal water flow velocity vector magnitude into a monitoring sample array, which is defined as the surface grid parameters.

[0056] Next, the system's safety constant declaration storage area presets a buoyancy drift depth threshold and a turbulent scour velocity threshold. The system writes two trigger floating-point constants. The first constant represents the depth scalar condition where the buoyancy displaced by the accumulated water causes the debris to suspend and displace algebraically, and is declared as the buoyancy drift depth threshold. The second constant represents the velocity modulus condition where the hydrodynamic shear stress generated by surface water flow causes the sediment to move numerically, and is declared as the turbulent scour velocity threshold. As an example, based on the physical equivalent density and gravitational stress characteristics of typical urban sedimentary layers, the buoyancy drift depth threshold is set to a floating-point constant between 0.20 meters and 0.35 meters; based on the empirical formula for bedload initiation velocity mechanics, the turbulent scour velocity threshold is set to a floating-point constant between 0.8 m / s and 1.5 m / s; thus, the scour and collapse physical phenomena have precise and quantifiable computer-determined trigger limits.

[0057] Furthermore, when the system determines that the real-time simulated water depth corresponding to the target virtual drainage node exceeds the buoyancy drift depth threshold, or determines that the real-time water surface velocity exceeds the turbulent scouring velocity threshold, it confirms that the clearing and reset conditions are met. The background monitoring thread extracts the surface grid parameters and performs logical comparison. It compares the obtained real-time simulated water depth variable with the buoyancy drift depth threshold constant, and simultaneously compares the obtained real-time water surface velocity variable with the turbulent scouring velocity threshold constant. When it determines that the real-time simulated water depth variable is greater than the threshold constant, or that the real-time water surface velocity variable is greater than the threshold constant, the system logic outputs true, confirming that the hydrological flow field value meets the critical limit and that the clearing and reset conditions are met. Finally, the system sends a high-priority trigger command to the main control scheduler, thereby triggering the asynchronous dredging update event. When the conditions are met, the system control flow is activated, and a high-priority trigger command is sent to the main control scheduler. The program execution unit suspends the current execution of the water flow smoothing integral operation context snapshot and calls the callback function of the corresponding event. The program execution pointer jumps to the entry address of the update service program code residing in the memory area, triggering the asynchronous unblocking update event execution logic for the specific node at the control logic level.

[0058] Step S180: In response to the asynchronous dredging update event, the accumulation count value in the status record variable of the stagnant material corresponding to the target virtual drainage node is forcibly cleared to zero, and the preset drainage throughput bandwidth is restored.

[0059] Specifically, the system's underlying task scheduler suspends the current degradation and rate limiting update instructions for the target virtual drainage node. After the central processing unit enters the service routine environment of the asynchronous drainage update event, it queries and retrieves the rate limiting and degradation method call pointers mounted within the target virtual drainage node structure. The system task scheduler updates the associated status flags in memory and sends a blocking and suspension signal to the called rate limiting and degradation control instruction stream. The periodic processes of reading stack data, calculating saturation, and calculating the bandwidth reduction ratio executed by the specific target drainage node are removed from the task queue, blocking the update operation of the limited drainage throughput bandwidth.

[0060] Subsequently, the system event handling mechanism writes a zeroing reset flag to the independent memory data node corresponding to the target virtual drainage node, completing the forced zeroing action. The system takes over data control of the target drainage node and locates the address of the allocated residual material status record variable. The system calls the variable reset function to zero out the data in the storage area storing the accumulation count value. After the variable reset operation, the residual data of the accumulated accumulation count value in the variable is returned to zero, completing the forced zeroing action. The operation simulates the reset state of debris detaching from the drainage orifice at the information level.

[0061] Next, the system management logic unbinds the parameter of the restricted drainage throughput bandwidth and redirects the overcurrent capacity parameter of the target virtual drainage node to the preset drainage throughput bandwidth. After the variable reset operation is completed, the system performs a parameter mapping modification operation. The system retrieves the call pointer in the target drainage node structure used to provide the baseline capacity for calculating boundary pumping parameters. The system disconnects the binding relationship between this pointer and the restricted drainage throughput bandwidth variable area and reconnects it to the preset drainage throughput bandwidth constant address recorded during the configuration phase. Through unbinding and pointer redirection, the pumping capacity parameter of the target virtual drainage node is restored to the design upper limit state. After the operation verification is completed, the system sends an event completion confirmation command to the main control scheduler, the processor restores the saved field snapshot data, releases the suspended state, and the calculation module resumes unidirectional calculation along the discrete integral time step.

[0062] The initial flooding prediction situation is updated based on the restored drainage capacity, and a terminal early warning command is generated.

[0063] Specifically, the fluid computing engine utilizes the restored preset drainage throughput bandwidth to guide the water accumulated within the surface runoff grid corresponding to the target virtual drainage node to flow downwards towards the target virtual drainage node. When the simulation system unblocks and enters the next discrete calculation iteration stage, the continuity equation solving engine extracts the water flow boundary capacity parameter of the corresponding node, restoring it from the limited flux to the double-precision constant value of the preset drainage throughput bandwidth. Based on the calculation rules of the gravity water pressure drop differential equation, the solving engine determines the mass of high-level water accumulated within the surface runoff grid. Under the operation of the partial differential equation momentum-driven term solution command, the water volume variable within the surface runoff grid is transferred and reduced towards the boundary of the target virtual drainage node below, presenting the data of water volume transfer downwards in the grid simulation output.

[0064] Next, in subsequent evolution cycles, the water level fluctuation trajectory of the entire grid is re-tracked to generate a corrected dynamic inundation situation map. As local discharge propagates through the grid calculations within the model's physical network, the calculated output of simulated water depth parameters in the associated regions decreases. The background situation monitoring thread extracts real-time simulated water depth and simulated flow velocity data sequences calibrated by the engine at subsequent equally divided time points. The system invokes the patch connectivity clustering algorithm to refresh and redraw the two-dimensional image matrix layer in the video memory rendering buffer. The system encapsulates and outputs the evolution layer containing the simulated local debris accumulation calculation phenomenon and the environmental threshold interruption logic-triggered water level drop calculation phenomenon as the corrected dynamic inundation situation map.

[0065] Subsequently, the spatial analysis module extracts the coordinates of target landmarks in a state of severe flooding from the dynamic flood situation map. The spatial analysis module performs a Boolean intersection search and comparison operation on the raster layer attached to the dynamic flood situation map. The system loads a severe flooding safety assessment warning threshold constant. As an example, based on the urban flood control emergency response disaster classification standard, the physical absolute depth value range of the severe flooding safety assessment warning threshold constant is set to 0.40 meters to 0.60 meters (for example, a single-precision floating-point constant of 0.50 meters is specifically fixed within the system). The spatial analysis module spatially locks the collection of connected polygonal structures with simulated water depth double-precision values ​​exceeding the threshold limit. The system retrieves the geographic coordinate resource table of points of interest and extracts the longitude and latitude floating-point data parameters of facility nodes such as transportation hubs and substations that have a spatial intersection relationship with the locked water polygons on the geographic projection plane. The system serializes the matched longitude and latitude floating-point data and stores it in a buffer array, extracting it as the target landmark coordinates.

[0066] Finally, the system packages the target landmark coordinates into the terminal warning command and pushes it to the flood control command and dispatch terminal for high-brightness flashing display via a wireless communication link. The underlying communication protocol module extracts the target landmark coordinate data structure array and performs serialization encoding and encapsulation operations according to the Transmission Control Protocol specification. The underlying communication protocol module constructs an information packet structure containing the positioning payload and outputs it as the terminal warning command data packet. The network controller uses the transmission protocol link to push the data packet to the command and dispatch terminal host through the communication link channel. After parsing the data packet payload, the dispatch terminal application code extracts the landing point coordinate group and calls the graphics processor rendering channel. The terminal display device, at the geographic coordinate projection plane position, modifies the internal attribute control parameters of the rendering command, sets the clock refresh cycle, and overlays and draws flashing geometric primitive components to complete the high-brightness flashing display.

[0067] It is understood that the variable node allocation, floating-point operations, state reset callback mechanism, memory data write operation, and pointer parameter remapping allocation mechanism involved in the embodiments rely on standard computer hardware architecture and instruction set platform for execution. The method computational architecture maps the local physical resistance evolution process through data variable operations, controls the computational flow operation of the finite difference hydrodynamic model under a fixed topological dimension matrix, and provides an equivalent transformation process for physical variables.

[0068] Furthermore, this application provides an electronic device for executing the aforementioned flood risk dynamic simulation method based on digital twins. The electronic device includes a processor chip, a memory module, a communication bus, a network transmission interface, and input / output control ports in its hardware structure. The processor chip embeds an arithmetic logic unit and a floating-point coprocessor for parsing and executing binary instruction streams in memory. The memory module includes static random access memory and dynamic random access memory, responsible for storing the operating system kernel code, flood control rainfall level matrix data, and various floating-point constant variables and stack pointer data structures involved in the method. The communication bus is responsible for establishing data transmission links between the processor, memory, and network interface. The network transmission interface is used to receive data reported by sensors and send early warning instruction data packets. When the processor chip reads and executes the program code stored in the memory module, through the processor's data reading, floating-point operations, logical comparisons, state resets, and variable assignments, the calculation process regarding the construction of the digital twin base, the degradation and current limiting calculation of debris accumulation, and the dredging, clearing, and reset response is realized.

[0069] Meanwhile, this application provides a computer-readable storage medium in which a sequence of computer-executable instructions is permanently stored. The physical form of the computer-readable storage medium includes solid-state drive (SSD) chips manufactured using insulated-gate semiconductor (IGDS) technology, mechanical hard disk drive (HDD) tracks and sectors processed using magnetic recording materials, or optical disc carriers prepared using laser phase-change technology. A program code stream capable of being read and decoded by a microprocessor is recorded in the physical storage structure. When the program code stream is extracted and executed by the central processing unit of the target terminal device, it drives the underlying hardware circuitry to implement the operational steps in the digital twin-based flood risk dynamic simulation method.

[0070] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for dynamic simulation of flood risk based on digital twins, comprising the following steps: Obtain basic multidimensional parameters of the flood control area; A digital twin simulation platform matching the flood control area is constructed based on the aforementioned basic multidimensional parameters; Hydrodynamic flow calculations are performed in the digital twin simulation platform to output the initial flooding prediction situation; The characteristic is that, The flood risk dynamic simulation method based on digital twins also includes: During the hydrodynamic flow calculation, a dedicated variable for recording the state of stagnant matter is created for each virtual drainage node within the digital twin simulation base. Based on the simulated inflow volume entering each of the virtual drainage nodes, the accumulation count value is accumulated in the state record variable of the retained material; Based on the accumulation count value, the preset drainage throughput bandwidth of each virtual drainage node is downgraded and limited to obtain a limited drainage throughput bandwidth, and the hydrodynamic flow calculation is continued according to the limited drainage throughput bandwidth. Continuously monitor the surface grid parameters corresponding to each of the virtual drainage nodes; When it is determined that the surface grid parameters meet the set clearing and reset conditions, an asynchronous dredging and update event is triggered for the target virtual drainage node that meets the conditions; In response to the asynchronous dredging update event, the accumulation count value in the status record variable of the stagnant material corresponding to the target virtual drainage node is forcibly cleared to zero, the preset drainage throughput bandwidth is restored, and the initial flooding prediction status is updated based on the restored drainage capacity to generate a terminal early warning command.

2. The flood risk dynamic simulation method based on digital twins according to claim 1, characterized in that, the acquisition of basic multidimensional parameters of the flood control area includes: The surface elevation point cloud sequence and land feature material classification labels of the flood control area were extracted through the remote sensing mapping channel. Connect to the IoT hydrological sensing terminal to read rainfall time series and river boundary water level series; The surface elevation point cloud sequence, the land feature material classification label, the rainfall time series sequence, and the river boundary water level sequence are spatiotemporally aligned and combined to form the basic multidimensional parameters.

3. The flood risk dynamic simulation method based on digital twins according to claim 2, characterized in that, the step of constructing a digital twin simulation base matching the flood control area based on the basic multidimensional parameters includes: The surface elevation point cloud sequence is used for grid subdivision to generate a spatial topological grid matrix; Based on the aforementioned land feature material classification labels, the basic roughness drag coefficient is mapped to each grid cell within the spatial topology grid matrix; In the spatial topology grid matrix, the specific coordinate positions of the corresponding underground pipe network inlets are marked, and the specific coordinate positions are instantiated into each of the virtual drainage nodes; The spatial topology grid matrix, the basic roughness resistance coefficient, and each of the virtual drainage nodes are integrated to solidify and generate the digital twin simulation base.

4. The flood risk dynamic simulation method based on digital twins according to claim 3, characterized in that, The process of performing hydrodynamic flow calculations in the digital twin simulation platform and outputting the initial flood prediction situation includes: The rainfall time series and the river boundary water level series are used as driving sources and injected into the digital twin simulation base. Set the evolution time step, and drive the water mass exchange and momentum transfer calculations between grid cells according to the evolution time step; Extract the simulated water depth and simulated flow velocity of each of the grid cells in each evolution cycle, and screen out the set of grid cells whose simulated water depth exceeds the safety threshold; The initial flooding prediction situation is generated based on the distribution range of the over-limit grid set.

5. The flood risk dynamic simulation method based on digital twins according to claim 4, characterized in that, The provision of dedicated variables for recording the state of retained objects at each virtual drainage node within the digital twin simulation base includes: In the system memory, an independent memory data node is allocated for each of the aforementioned virtual drainage nodes; The independent memory data nodes are encapsulated as the status record variables of the retained material corresponding to each of the virtual drainage nodes; The variable for recording the state of the retained material is initially set to a blank state, and the initial base of the accumulation count value is set to zero.

6. The flood risk dynamic simulation method based on digital twins according to claim 5, characterized in that, The step of accumulating a buildup count value into the retention state record variable based on the simulated inflow volume entering each of the virtual drainage nodes includes: Within each evolution cycle, the simulated inflow volume input to each of the virtual drainage nodes is extracted from the hydrodynamic flow calculation. Retrieve the pre-set virtual impurity content ratio coefficient; Multiply the simulated inflow volume by the virtual impurity content ratio coefficient to obtain the calculated increment for the current period; The calculated increment is added to the previous recorded value of the residual state record variable to update the current accumulation count value.

7. The flood risk dynamic simulation method based on digital twins according to claim 6, characterized in that, the step of performing a degradation and flow limiting operation on the preset drainage throughput bandwidth of each of the virtual drainage nodes based on the accumulation count value to obtain the limited drainage throughput bandwidth includes: Set the maximum allowed stacking capacity and the minimum guaranteed bandwidth ratio; Calculate the saturation ratio of the stack count value to the maximum allowable stacking capacity limit; Subtract the saturation ratio from the full load baseline ratio to obtain the current available performance ratio; Compare the current available performance ratio with the minimum guaranteed bandwidth ratio, and select the larger value as the effective quota ratio; Multiply the effective quota ratio by the preset drainage throughput bandwidth to output the limited drainage throughput bandwidth.

8. The flood risk dynamic simulation method based on digital twins according to claim 7, characterized in that the continuous monitoring of the surface grid parameters corresponding to each of the virtual drainage nodes; When the surface grid parameters are determined to meet the set clearing and reset conditions, an asynchronous dredging update event is triggered for the target virtual drainage node that meets the conditions, including: Lock the surface flow-bearing grid directly above each of the virtual drainage nodes; The real-time simulated water depth and real-time water surface velocity of the surface flow-bearing grid are obtained and used as the parameters of the surface grid; Preset buoyancy drift depth threshold and turbulent scouring velocity threshold; When it is determined that the real-time simulated water depth of the corresponding target virtual drainage node exceeds the buoyancy drift depth threshold, or when it is determined that the real-time water surface velocity exceeds the turbulent scouring velocity threshold, the clearing and reset condition is confirmed to be met. A high-priority trigger command is sent to the main control scheduler, thereby triggering the asynchronous unblocking update event.

9. The flood risk dynamic simulation method based on digital twins according to claim 8, characterized in that, in response to the asynchronous dredging update event, forcibly clearing the accumulation count value in the state record variable of the retained material corresponding to the target virtual drainage node to zero, and restoring the preset drainage throughput bandwidth, includes: Suspend the current downgrade flow restriction update command of the target virtual drainage node; Write a zeroing and reset flag to the independent memory data node corresponding to the target virtual drainage node to complete the forced zeroing action; Remove the parameter binding of the restricted drainage throughput bandwidth and redirect the overcurrent capacity parameter of the target virtual drainage node to the preset drainage throughput bandwidth.

10. The flood risk dynamic simulation method based on digital twins according to claim 9, characterized in that, The process of updating the initial flood prediction situation based on the restored drainage capacity and generating terminal early warning instructions includes: Using the restored preset drainage throughput bandwidth, the water accumulated in the surface runoff grid corresponding to the target virtual drainage node is guided to be discharged to the target virtual drainage node below. In subsequent evolution cycles, the water level fluctuation trajectory of the entire grid is re-tracked to generate a corrected dynamic flooding situation map; Extract the coordinates of target landmarks in a state of severe flooding from the dynamic flooding situation map; The target landmark coordinates are packaged into the terminal warning command and pushed to the flood control command and dispatch terminal via wireless communication link for high-brightness flashing display.

Images