Hydro-junction flood control early warning rehearsal method based on digital twinning

Abstract

The invention relates to the technical field of flood control early warning, and discloses a hydro-junction flood control early warning rehearsal method based on digital twinning, and the method comprises the steps: constructing a hydro-junction-oriented sensing monitoring network to collect multi-dimensional sensing data; the method comprises the following steps: establishing a hydro-junction digital twin coupling model according to topographic surveying and mapping data, inputting gridding area rainfall driving data, driving each node to cooperatively calculate by using a distributed message queue technology to obtain a full-basin flow evolution process, and extracting a flood peak flow value and a flood peak arrival time; obtaining a key section dynamic water level sequence according to a full-basin flow evolution process; the flood peak flow value, the flood peak arrival time and the key section dynamic water level sequence are compared with flood control safety threshold values, and when any index exceeds the limit, early warning information and an engineering scheduling scheme are generated. According to the invention, the three-dimensional sensing and monitoring network and the digital twin distributed computing topology network are constructed, so that accurate fusion and intelligent reverse scheduling of multi-source data are realized, and the monitoring precision of the hydro-junction to cope with extreme flood is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of flood control early warning technology, and in particular to a flood control early warning simulation method for water conservancy hubs based on digital twins. Background Technology

[0002] As a key node in the flood control system of a river basin, the safe operation and scheduling decisions of water conservancy projects directly affect the safety of life and property in downstream areas. With the intensification of global climate change and the frequent occurrence of extreme rainfall events, river basin floods are exhibiting new characteristics such as suddenness, high peak volume, and short confluence time, posing severe challenges to traditional flood control early warning and scheduling of water conservancy projects.

[0003] Current flood monitoring and early warning technologies suffer from the following limitations: First, in terms of monitoring methods, traditional approaches rely heavily on sparsely distributed rain gauges and water level stations, lacking integrated air-space-ground sensing capabilities. This makes it difficult to obtain high-precision topographic and meteorological data in complex upstream mountainous areas or monitoring blind spots, hindering the accurate capture of spatial distribution characteristics of rainfall. Second, regarding forecasting models, existing hydrological forecasts often employ lumped models, which struggle to accurately simulate the spatial heterogeneity within watersheds. Furthermore, for mixed rain-snow floods unique to arid regions, there is a lack of effective co-firing mechanisms for runoff generation and concentration, limiting forecast accuracy and the extension of the lead time. In addition, traditional flood evolution simulations often separate hydrological and hydrodynamic processes, resulting in computational efficiency insufficient for real-time simulations. Moreover, there is a lack of intelligent scheduling platforms capable of responding to changes in engineering conditions in real time and performing reverse simulations.

[0004] In recent years, the rise of digital twin technology has provided a new path for the construction of smart water conservancy. However, most existing digital twin water conservancy applications focus on the three-dimensional visualization of static scenes, lacking in-depth mechanistic model coupling and high-concurrency real-time computing capabilities, making it difficult to truly achieve a closed-loop chain from data foundation to business scheduling. Therefore, there is an urgent need for flood control and early warning methods that can integrate multi-dimensional monitoring data, achieve refined simulation of complex runoff generation and confluence mechanisms, and possess real-time pre-simulation and intelligent scheduling decision-making capabilities, in order to improve the resilience and management level of water conservancy projects in the face of extreme floods. Summary of the Invention

[0005] This invention provides a flood early warning and simulation method for water conservancy hubs based on digital twins. By constructing a three-dimensional sensing monitoring network and a digital twin distributed computing topology network, it achieves accurate fusion of multi-source data, efficient simulation of rain and snow mixed floods, and intelligent reverse scheduling, significantly improving the monitoring accuracy, simulation efficiency, and emergency decision-making level of water conservancy hubs in response to extreme floods.

[0006] This invention provides a flood control early warning and simulation method for water conservancy projects based on digital twins, including:

[0007] S1. Construct an integrated three-dimensional sensing and monitoring network of "sky-air-ground-water" for water conservancy hubs to collect multi-dimensional sensing data; wherein, the multi-dimensional sensing data includes meteorological image data, topographic mapping data, and engineering and water condition monitoring data;

[0008] S2. Spatial discretization of the topographic mapping data is performed to obtain a hydrodynamic calculation grid. The meteorological image data is combined with the ground rain gauge data, and the optimal interpolation method is used to generate gridded surface rainfall driving data to form the digital twin calculation boundary conditions.

[0009] S3. Establish a distributed computing topology network as a digital twin coupling model of the water conservancy hub based on the topographic mapping data; wherein, the distributed computing topology network is composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes cascaded together, and each upstream small watershed node is equipped with an improved SCS runoff generation and runoff model and an improved SRM snowmelt runoff model.

[0010] S4. Input the gridded surface rainfall driving data into the digital twin coupling model of the water conservancy hub, use distributed message queue technology to drive the collaborative calculation of each node to obtain the whole basin flow evolution process and extract the flood peak flow value and flood peak arrival time, and load the whole basin flow evolution process into the hydrodynamic calculation grid to obtain the dynamic water level sequence of key sections.

[0011] S5. The flood peak flow rate, flood peak arrival time, and dynamic water level sequence of key sections are used as flood feature parameter vectors. The flood feature parameter vectors are compared with flood control safety thresholds in the knowledge base. When any indicator in the flood feature parameter vector exceeds the limit, early warning information and engineering scheduling plan for the water conservancy hub are generated.

[0012] Furthermore, S1 specifically includes:

[0013] S101. Use high-resolution remote sensing satellites and meteorological satellites to conduct large-scale scanning of water conservancy hubs and watersheds, analyze satellite remote sensing images to extract water body boundaries and snow cover range, and invert atmospheric cloud images and macro-rainfall trends to obtain satellite remote sensing images, snow cover information and atmospheric cloud image information as meteorological image data.

[0014] S102. Deploy drones and rotary-wing aircraft to conduct oblique photogrammetry of reservoir areas, dam hubs and key river sections to produce high-precision digital elevation models and digital orthophotos. In emergency situations, transmit back video of the disaster site to obtain digital elevation models, digital orthophotos and real-time disaster video information as topographic mapping data.

[0015] S103. GNSS positioning sensors and piezometers deployed on the dam body and bank slopes are used to monitor structural displacement and internal seepage pressure. Ground rain gauges and water level stations are used to monitor fixed-point precipitation and river water level. Displacement deformation, seepage pressure, rainfall and river water level information are used as the ground part of the engineering and water situation monitoring data.

[0016] S104. Use underwater robots or fixed underwater sensors to collect underwater flow velocity and water quality parameters at the dam toe and stilling basin, and detect the hidden engineering status of the dam toe sedimentation or scour depth, and obtain underwater flow velocity, water quality parameters and hidden engineering status information as the underwater part of the engineering and water condition monitoring data.

[0017] S105. The collected meteorological image data, topographic mapping data, and engineering and water condition monitoring data are transmitted to the central server using the communication network. The data is then cleaned and correlated using a unified timestamp and spatial coordinate system to complete the data standardization process.

[0018] Furthermore, S2 specifically includes:

[0019] S201. Extract terrain feature constraint lines using the digital orthophoto, generate an unstructured grid based on the terrain feature constraint lines, and map the digital elevation model data to the grid nodes to generate a hydrodynamic calculation grid for solving the two-dimensional shallow water equation.

[0020] S202. Using GIS spatial analysis to analyze the topographic mapping data, the watershed is divided into multiple small watershed units. The average slope and watershed area of ​​each small watershed unit are calculated as geometric parameters for the improved SCS runoff generation and confluence model. The area-elevation curves of each small watershed unit are extracted to determine the area ratio of the elevation zone as geometric parameters for the improved SRM snowmelt runoff model.

[0021] S203. Extract the real-time upstream water level and gate opening from the engineering and water situation monitoring data, and set the initial water level and initial gate opening for the two-dimensional hydrodynamic model and the reservoir flood control calculation model respectively.

[0022] S204. The optimal interpolation method is used to fuse the initial radar precipitation estimate retrieved from the meteorological image data with the ground rain gauge observations to generate gridded areal rainfall driving data. The calculation formula is as follows:

[0023]

[0024] Among them, R″ i R' is the precipitation analysis value generated for the i-th grid point. i Let P be the initial radar precipitation estimate for the i-th grid point, n be the number of rain gauges involved in the calculation, and P be the initial estimate for the i-th grid point. k As a weighting factor, Let k be the observed value from the k-th rain gauge. The initial estimate of radar precipitation at the location of the kth rain gauge station;

[0025] S205. The hydrodynamic calculation grid, geometric parameters, initial water level, initial gate opening, and gridded surface rainfall driving data are standardized and encapsulated to form the digital twin calculation boundary conditions.

[0026] Further, in S204, the weighting factor P k The calculation is adaptive based on the distribution density of rain gauges; that is, when the number of rain gauges within the set search radius meets a preset threshold, a dense correlation function is used. Calculate the correlation function; otherwise, use a sparse correlation function. Calculate, where r ij r represents the distance between rain gauges. u This represents the distance between the grid point and the rain gauge, where 'a' is the relevant length parameter.

[0027] Furthermore, S3 specifically includes:

[0028] S301. Based on the topographic mapping data analysis of several small watershed units and their topological relationships, establish a distributed computing topology network composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes, and configure a message queue interface for each node to achieve decoupled data transmission.

[0029] S302. Configure the improved SCS runoff generation and confluence model and the improved SRM snowmelt runoff model in the upstream small watershed nodes, and construct a multi-source runoff coupling mechanism. That is, determine the precipitation pattern according to the real-time temperature. If it is a simple rainfall, only the SCS model is called. If it is a mixture of rain and snow, the rainfall runoff calculated by the SCS model and the snowmelt runoff calculated by the SRM model are linearly superimposed as the total runoff of the node.

[0030] The improved SCS runoff generation and runoff model uses the SCS-CN curve number method to calculate net rainfall and incorporates the previous soil moisture level to correct the CN value. It uses the triangular generalized unit line method for runoff calculation and sets the time base width of the unit line to 8 / 3 times the time of the flood peak.

[0031] The improved SRM snowmelt runoff model refines the calculation step to a time scale and introduces a time coefficient for precipitation-induced snowmelt intensity. Its calculation formula is as follows:

[0032]

[0033] Among them, Q n+1 For time-period forecast traffic flow, C s Here, is the snowmelt runoff coefficient, 'a' is the degree-day factor, and 'T' is the temperature nHere, ΔT is the temperature, and S is the temperature correction value. n C represents the snow cover rate. r Here, P is the snowmelt runoff coefficient, T0 is the critical snowmelt temperature, A is the catchment area, and k is the runoff coefficient. n+1 This is the flow attenuation coefficient;

[0034] S303. Initialize the model using the digital twin calculation boundary conditions, inject the average slope of the watershed, the watershed area and the elevation zone distribution parameters into the corresponding upstream small watershed nodes, and establish a binding relationship between the gridded surface rainfall driving data and the input variables of the upstream small watershed nodes.

[0035] Furthermore, S4 specifically includes:

[0036] S401. Using GIS spatial analysis, the gridded surface rainfall-driven data is mapped to the upstream small watershed nodes, time-series slices are performed according to the calculation step size, and the data is pushed to the input queues of each node in parallel through a message queue.

[0037] S402. Utilize distributed message queue technology to drive the various nodes in the topology network constructed in step S3 to run sequentially in order to perform distributed topology network collaborative computing;

[0038] S403. Using a cascaded calculation strategy, the overflow of the downstream flow process or the overflow of the river section node is mapped to the source and sink terms of the hydrodynamic calculation grid. The two-dimensional shallow water equation set is solved on the grid using the finite volume method to obtain the dynamic water level sequence of the key section.

[0039] S404. Extract the maximum flow value and its corresponding time from the discharge flow process or river evolution process as the flood peak flow value and flood peak arrival time, and combine them with the dynamic water level sequence to generate a flood characteristic parameter vector.

[0040] Furthermore, S402 specifically includes:

[0041] After receiving the data, the upstream small watershed node calls the improved SCS and SRM models to calculate rainfall and snowmelt runoff, linearly superimposes them to obtain the unit outlet flow, and encapsulates it into a message to send to the midstream river node.

[0042] The midstream river node receives confluence messages from the input queue and calculates the river channel evolution process using the Muskingan model; the calculation formula is as follows: In the formula Q out I represents outflow, I represents inflow, and C0, C1, and C2 are the bus parameters;

[0043] The downstream reservoir node receives the inflow, and based on the initial water level and initial gate opening, the water balance equation is used. With the discharge capacity curve Q=f(H,O) g The iterative solution is performed to obtain the discharge flow process and the reservoir water level change process, where V is the reservoir capacity, H is the reservoir water level, and O is the reservoir level. g This refers to the gate opening degree.

[0044] Furthermore, S5 specifically includes:

[0045] S501. Construct a flood control scheduling knowledge base, establish a knowledge graph with reservoirs, rivers and protected objects as entities, define the relationship between characteristic water levels, reservoir capacity curves, river safe discharge attributes and gate discharge capacity curves, and set graded early warning thresholds.

[0046] S502. Analyze the flood characteristic parameter vector, decompose it into peak flow value, peak arrival time and key section dynamic water level sequence, and extract the reservoir water level and gate opening from the engineering and water situation monitoring data at the current time as the starting boundary for scheduling calculation.

[0047] S503. Perform multi-dimensional risk assessment, compare the decomposed prediction indicators with the graded early warning thresholds, and issue a flood warning when the dynamic water level sequence of the key section exceeds the guaranteed water level or the flood peak flow value exceeds the safe discharge of the river channel. Issue an engineering safety warning when the predicted reservoir water level exceeds the design flood level or the dam monitoring data is abnormal.

[0048] S504. Generate an intelligent scheduling scheme based on the reverse deduction algorithm, calculate the pre-discharge time window according to the arrival time of the flood peak, set the objective function to control the downstream flow within the safe discharge capacity, use the gate discharge capacity curve to solve the gate opening sequence required to satisfy the objective function in reverse, and use the engineering and water condition monitoring data to verify the gate opening and closing constraints.

[0049] S505. Perform closed-loop simulation, using the gate opening sequence as a new boundary condition to perform a secondary simulation in the distributed computing topology network of step S4, and evaluate the future reservoir water level and downstream flow process after implementing the scheme. If the evaluation result meets the flood control safety requirements, the recommended engineering scheduling scheme is output.

[0050] The present invention also provides a flood control early warning and simulation device for water conservancy projects based on digital twins, which, based on the flood control early warning and simulation method for water conservancy projects based on digital twins as described above, includes:

[0051] The data acquisition module is used to construct an integrated three-dimensional sensing and monitoring network of "sky-air-ground-water" for water conservancy hubs to collect multi-dimensional sensing data; wherein, the multi-dimensional sensing data includes meteorological image data, topographic mapping data, and engineering and water condition monitoring data;

[0052] The generation module is used to spatially discretize the topographic mapping data to obtain a hydrodynamic calculation grid, combine the meteorological image data with the ground rain gauge data, and use the optimal interpolation method to generate gridded surface rainfall driving data to form the digital twin calculation boundary conditions.

[0053] A construction module is used to establish a distributed computing topology network as a digital twin coupling model of the water conservancy hub based on the topographic mapping data; wherein, the distributed computing topology network is composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes cascaded together, and each upstream small watershed node is equipped with an improved SCS runoff generation and runoff model and an improved SRM snowmelt runoff model.

[0054] The pre-simulation module is used to input the gridded surface rainfall driving data into the digital twin coupling model of the water conservancy hub, use distributed message queue technology to drive the collaborative calculation of each node to obtain the whole basin flow evolution process and extract the flood peak flow value and flood peak arrival time, and load the whole basin flow evolution process into the hydrodynamic calculation grid to obtain the dynamic water level sequence of key sections.

[0055] The early warning module is used to take the flood peak flow value, the flood peak arrival time and the dynamic water level sequence of key sections as flood feature parameter vectors, and compare the flood feature parameter vectors with flood control safety thresholds in the knowledge base. When any indicator in the flood feature parameter vector exceeds the limit, early warning information and engineering scheduling plan for water conservancy projects are generated.

[0056] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0057] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0058] The beneficial effects of this invention are as follows:

[0059] This invention constructs a three-dimensional sensing and monitoring network encompassing "sky-air-ground-water" and a digital twin full-link closed-loop system for water conservancy projects. This addresses the blind spots and limited data sources inherent in traditional single-point monitoring in complex terrain, enabling all-weather acquisition and precise spatiotemporal fusion of multi-source heterogeneous data. It provides high-precision gridded driving data for the model. By constructing a distributed computing topology network including an improved SCS runoff generation model and an improved SRM snowmelt runoff model, and utilizing distributed message queue technology to drive parallel computing, it significantly improves the simulation accuracy for mixed rain and snow floods unique to arid and semi-arid regions. It also overcomes the limitation of lumped models in reflecting the spatial heterogeneity of watersheds, greatly increasing computational efficiency. Furthermore, through the cascaded solution of one-dimensional hydrological flow evolution and two-dimensional hydrodynamic water level extrapolation, combined with a knowledge graph-based reverse scheduling algorithm, it achieves a closed-loop business process from flood situation awareness and real-time dynamic simulation to intelligent engineering scheduling. This allows for the rapid generation of scientifically feasible scheduling plans, significantly shortening early warning response time and enhancing the flood control resilience and refined management level of water conservancy projects in the face of extreme flood disasters. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of a method flow according to an embodiment of the present invention.

[0061] Figure 2 This is a schematic diagram of the device structure according to an embodiment of the present invention.

[0062] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of the present invention.

[0063] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0064] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. Figure 1 As shown, this invention provides a flood control early warning and simulation method for water conservancy projects based on digital twins, including:

[0065] S1. Construct an integrated three-dimensional sensing and monitoring network encompassing "space-air-ground-water" for water conservancy projects to collect multi-dimensional sensing data; wherein, the multi-dimensional sensing data includes meteorological imagery data, topographic mapping data, and engineering and water condition monitoring data; specifically including:

[0066] S101. Implement "space-based" macroscopic remote sensing monitoring: Utilize high-resolution remote sensing satellites and meteorological satellites as high-altitude sensing nodes to conduct large-scale periodic scans of the entire watershed, analyze satellite remote sensing images to extract water body boundaries and snow cover ranges, identify snow-covered areas, and obtain macroscopic atmospheric cloud images and rainfall trend distribution data; the aforementioned satellite remote sensing images, snow cover information, and atmospheric cloud image information constitute the meteorological image data.

[0067] S102. Implement "airborne" high-precision local surveys: Deploy drones and rotary-wing aircraft as low-altitude sensing nodes to conduct high-frequency, high-precision aerial surveys of reservoir areas, dam hubs, and key river sections (using onboard high-definition cameras and lidar equipment) to obtain high-precision digital elevation models (DEMs) and digital orthophotos (DOMs), and transmit disaster site videos back in emergency situations; the aforementioned digital elevation models, digital orthophotos, and real-time disaster video information constitute the topographic mapping data.

[0068] S103. Implement ground-based engineering safety and hydrological monitoring: Utilize IoT sensors deployed at key locations in flood control projects as ground-based sensing nodes. Specifically, this involves:

[0069] Engineering monitoring: GNSS positioning sensors, piezometers and inclinometers are deployed on the dam surface, dam foundation and bank slope to collect data on the horizontal displacement, vertical settlement and internal seepage pressure of the dam in real time, in order to monitor the structural safety status of the dam.

[0070] Hydrological monitoring: Remote rain gauges and water level stations are deployed within the basin to collect real-time rainfall and river water level data at fixed points. The above-mentioned displacement deformation, seepage pressure, rainfall and river water level information constitute the ground part of the hydrological monitoring data.

[0071] S104. Conduct "water"-based concealed engineering detection: Utilize underwater robots (ROVs) or fixed underwater sensors as underwater sensing nodes to perceive the status of concealed engineering below the water surface. Through onboard multibeam sonar or pressure sensors, collect underwater flow velocity, water quality parameters, and the depth of sedimentation or scour at the dam toe and stilling basin (converted through dam toe pressure or sonar imaging) to assess the status of concealed engineering. The aforementioned underwater flow velocity, water quality parameters, and concealed engineering status information constitute the underwater portion of the engineering and water condition monitoring data.

[0072] S105. Using 5G communication networks or satellite communication links, the meteorological image data, topographic mapping data, and engineering and water condition monitoring data collected from S101 to S104 are uniformly transmitted to the central server. The above data are cleaned, unified with timestamps and spatial coordinate systems, and data indexes are established to form a standardized multidimensional sensing dataset.

[0073] S2. The topographic mapping data is spatially discretized to obtain a hydrodynamic calculation grid. The meteorological image data is combined with ground rain gauge data, and the optimal interpolation method is used to generate gridded surface rainfall driving data, forming the digital twin calculation boundary conditions; specifically including:

[0074] S201. Construct a hydrodynamic computational grid (spatial discretization).

[0075] The construction of the hydrodynamic computational grid is based on the topographic mapping data collected by the UAV and rotary-wing aircraft in step S1, specifically including a high-precision digital elevation model (DEM) and digital orthophoto (DOM). The specific processing procedure is as follows:

[0076] Feature line extraction: The image processing algorithm automatically identifies the land-water boundary line, the top axis of the levee, and the deep channel line in the digital orthophoto (DOM) and extracts them as topographic feature constraint lines;

[0077] Mesh generation: Using the terrain feature constraint lines as boundaries, an unstructured mesh generation algorithm (such as Delaunay triangulation) is used to discretize the reservoir area, downstream river channel and floodplain of the water conservancy hub into a two-dimensional computational mesh composed of nodes and cells.

[0078] Elevation assignment: The terrain elevation data in the digital elevation model (DEM) is interpolated and mapped to each node of the two-dimensional computational grid, giving the grid three-dimensional spatial attributes. This grid serves as the geometric carrier for solving the two-dimensional shallow water equation (i.e., the hydrodynamic mathematical model) in the subsequent step S4.

[0079] S202, Analyze the hydrological geometric parameters of the watershed.

[0080] The hydrological model includes an improved SCS runoff generation and runoff model and an improved SRM snowmelt runoff model. To drive the hydrological model, the following geometric parameters are pre-analyzed from the topographic mapping data in step S1:

[0081] Sub-basin delineation: Using the DEM data after depression filling processing by the GIS hydrological analysis module, the flow direction and cumulative discharge are calculated, and the entire upstream basin is divided into several independent sub-basin units.

[0082] SCS model parameter extraction: For each small watershed unit, calculate its average watershed slope y and watershed area A, which are used to calculate the peak flood time in subsequent calculations;

[0083] SRM model parameter extraction: Statistical analysis of the area-elevation curves (Hypsometric Curve) of each small watershed unit, vertically dividing the watershed into multiple elevation zones, determining the area proportion of each elevation zone, which is used for subsequent zonal interpolation of temperature and snow cover.

[0084] S203. Set the initial running state of the mathematical model.

[0085] The mathematical models include a two-dimensional hydrodynamic model for simulating the flow of water in the reservoir area and the river channel, and a reservoir flood control calculation model for simulating reservoir operation.

[0086] The initial conditions of the above model are set using the engineering and hydrological monitoring data collected by the ground sensors in step S1. Specifically, the real-time readings of the upstream water level gauges are read to set the initial water level H0 of the reservoir area nodes in the two-dimensional hydrodynamic model grid; the real-time readings of the gate opening sensors are read to set the initial gate opening O in the reservoir flood control calculation model. g The system determines the startup mode. If it is a cold start, the initial velocity field is assumed to be zero. If it is a hot start, the flow field data from the previous simulation time is inherited as the initial conditions.

[0087] S204. Generate gridded surface rainfall driving data.

[0088] To address the issue of insufficient accuracy from a single data source, optimal interpolation (OI) is employed to fuse the meteorological imagery data retrieved from the meteorological radar in step S1 with rainfall data collected from ground rain gauges. The specific calculation process follows the formula below:

[0089]

[0090] Among them, R″ i R' represents the corrected precipitation analysis value of the i-th grid point, i.e., the final generated gridded surface rainfall driving data; i This represents the initial radar estimate of precipitation at the i-th grid point, which is directly obtained from meteorological radar data through the ZI relationship (relationship between radar reflectivity factor and rainfall intensity); n represents the total number of effective surface rain gauges involved in the calculation; k represents the index of the k-th surface rain gauge. This represents the actual rainfall value observed at the k-th ground rain gauge station; This represents the initial estimate of precipitation retrieved by radar at the location of the k-th ground rain gauge; P k This represents the weight factor of the k-th ground rain gauge for grid point i.

[0091] Weighting factor P k The method for determining the correlation function is as follows: The correlation function is adaptively selected based on the distribution density of ground rain gauges. A search radius centered on the grid points is set. If the number of rain gauges within the search radius (e.g., within a range of 5km to 30km) meets a preset threshold (determined as densely distributed), then a dense correlation function is used. If the preset threshold is not met (determined to be a sparse distribution), then a sparse correlation function is used: In the above formula, r ij r represents the distance between rain gauges. u This represents the distance between the grid point and the rain gauge, where 'a' is the relevant length parameter.

[0092] S205, Forming boundary conditions for digital twin computing

[0093] The hydrodynamic calculation grid generated in S201, the watershed hydrological geometric parameters analyzed in S202, the initial water level and gate opening set in S203, and the gridded surface rainfall driving data generated in S204 are standardized and encapsulated to form a complete digital twin calculation boundary condition, which is then transmitted to the digital twin coupled model constructed in step S3 via a data bus.

[0094] S3. Establish a distributed computing topology network as a digital twin coupling model for the water conservancy hub based on the topographic mapping data; wherein, the distributed computing topology network is composed of cascaded upstream small watershed nodes, midstream river section nodes, and downstream reservoir nodes, and each upstream small watershed node is equipped with an improved SCS runoff generation and confluence model and an improved SRM snowmelt runoff model; specifically including:

[0095] S301. Establish a distributed computing topology network.

[0096] To address the issue that traditional lumped-database models struggle to reflect spatial heterogeneity, a distributed computing topology network based on a directed acyclic graph (DAG) is constructed, using several small watershed units and their topological relationships analyzed in step S2. The specific construction process is as follows:

[0097] The entities in the physical watershed are abstracted into three types of computational node objects: runoff-generating nodes, evolution nodes, and regulation nodes. Runoff-generating nodes (upstream sub-basin nodes) correspond to each upstream sub-basin unit and are responsible for calculating the raw runoff generated by rainfall and snowmelt. Evolution nodes (midstream river section nodes) correspond to the midstream river section and are responsible for calculating the propagation and shoaling of water flow within the river channel. Regulation nodes (downstream reservoir nodes) correspond to the downstream water conservancy hubs (reservoirs) and are responsible for calculating reservoir regulation and gate discharge. Based on the water flow direction extracted from the Digital Elevation Model (DEM), input / output interfaces (I / O interfaces) between nodes are defined. For example, the output port of the upstream runoff-generating node connects to the input port of the corresponding downstream evolution node, forming a cascading logic. Finally, an independent RabbitMQ message queue is configured for each node. Runoff-generating nodes act as message producers, while evolution nodes and regulation nodes are both consumers and producers, thereby achieving decoupled transmission of hydrological process data.

[0098] S302. Configure the improved SCS runoff generation and confluence model and the improved SRM snowmelt runoff model in the runoff generation nodes, and construct a multi-source runoff coupling mechanism. Specifically:

[0099] ① Configure an improved SCS generation and merging model

[0100] An improved Soil Conservation Service (SCS) model is embedded within each runoff generation node to calculate surface runoff generated by rainfall. The traditional SCS model is specifically improved and configured as follows:

[0101] Net rainfall calculation: Net rainfall is calculated using the SCS-CN curve method, with anterior soil moisture content (AMC) dynamically correcting the CN value. The formula is as follows:

[0102]

[0103] Where R is net rainfall, P is total rainfall (from the driving data in step S2), and I... a The initial loss is S, and the potential maximum retention is S.

[0104] Flow calculation: The triangular generalized unit hydrograph method is used, and the main parameter of the unit hydrograph is defined as the peak flow rate q. p The time t from the rise of the flood peak to the appearance of the flood peak p To better reflect the rapid rise and fall of flood events in arid regions, the time base T of the unit line of the triangle was adjusted. b Strictly defined as the time t when the flood peak occurs p Multiples of, let the relationship be set as It exhibits a higher peak timing consistency in actual calibration.

[0105] ② Configure an improved SRM snowmelt runoff model

[0106] An improved Snowmelt Runoff Model (SRM) is implemented in parallel at runoff-generating nodes. To address the shortcomings of traditional SRM models in refined day-step forecasting, the following specific improvements and configurations are made:

[0107] Time step refinement: The calculation step is refined from "day" to "time period" (e.g., 1 hour) to capture the snow melt fluctuations caused by intraday temperature changes.

[0108] Introducing a time coefficient for snowmelt rate due to rainfall: To quantify the accelerating effect of heat carried by liquid rainfall on snowmelt, an improved formula for calculating snowmelt flow rate is constructed:

[0109]

[0110] Among them, Q n+1C represents the predicted snowmelt runoff for the (n+1)th time period; s C r These are the snowmelt runoff coefficient and the snowmelt precipitation runoff coefficient, respectively, reflecting the runoff generation capacity of different media; 'a' is the degree-day factor, representing the daily snowmelt depth per degree of positive temperature, dynamically adjusted with the seasons; T n S represents the measured temperature in the nth time period; ΔT represents the temperature lapse rate correction value, calculated using the elevation zone information analyzed in step S2; S n P represents the snow cover in time period n (provided by satellite remote sensing data); T represents the rainfall in time period n (from the driving data in step S2); T0 represents the critical temperature for snowmelt; k n+1 It represents the flow attenuation coefficient, reflecting the drainage characteristics of the watershed.

[0111] ③ Construct a multi-source flow coupling mechanism

[0112] Define the coupled computational logic within the runoff generation node to address the superposition problem of rain-snow mixed runoff. Specifically, within each computational step, first calculate based on the real-time temperature T... n The model determines the precipitation form (liquid rain or solid snow) based on the critical temperature. For simple rainfall, only the SCS model is used for calculation. For mixed rain and snow (i.e., snow cover with rainfall), both the SCS model (calculating rainfall runoff) and the SRM model (calculating snowmelt and rain-induced snowmelt) are used simultaneously. The runoff processes calculated by both models are linearly superimposed to generate the total outflow process for that node. This coupling mechanism ensures the model's versatility in scenarios such as summer torrential floods, spring snowmelt floods, and mixed floods.

[0113] S303. Initialize all nodes by injecting the geometric parameters (slope y, area A, elevation zone distribution) obtained in step S2 into the corresponding watershed nodes; and establish a data subscription relationship between the gridded areal rainfall driving data generated in step S2 and the input variable P of the runoff-producing nodes to complete the model startup preparation.

[0114] S4. Input the gridded areal rainfall-driven data into the digital twin coupled model of the water conservancy hub, and use distributed message queue technology to drive the collaborative calculation of each node to obtain the whole basin flow evolution process and extract the peak flow value and peak arrival time. Load the whole basin flow evolution process into the hydrodynamic calculation grid to obtain the dynamic water level sequence of key sections; specifically including:

[0115] S401. Using the GIS spatial analysis function, the gridded areal rainfall data is spatially overlaid with several upstream small watershed nodes defined in step S3. According to the calculation step size (e.g., 1 hour), the areal rainfall data is sliced ​​into time-series pulse signals. Through the publish / subscribe (Pub / Sub) mode of the RabbitMQ message bus, the corresponding rainfall data packets are pushed in parallel to the input queues of all upstream small watershed nodes, triggering the full network calculation.

[0116] S402. Utilize distributed message queue technology to drive the various nodes in the topology network constructed in step S3 to run sequentially, thereby performing distributed topology network collaborative computing. Specifically, this includes:

[0117] ① Upstream small watershed node (runoff generation and runoff calculation): The node listens to the input queue. Once it receives a rainfall signal, it calls the internally configured improved SCS and SRM models; calculates the rainfall runoff and snowmelt runoff for that period, and linearly superimposes them to obtain the unit outlet flow; encapsulates the flow data into a message with a timestamp and sends it to the input queue of the downstream river segment node logically connected to it.

[0118] ② Midstream node (evolution calculation): The node obtains one or more confluence messages from the upstream through the input queue; it calls the Muskingum model to perform river flood evolution calculation, using the following formula:

[0119]

[0120] Where I represents the inflow, Q out For outflow, C0, C1, and C2 are confluence parameters (determined by the river length and wave velocity); the calculated outflow process is written into the input queue of the downstream reservoir node.

[0121] ③ Downstream reservoir node (regulation calculation): The node aggregates all inflow information to obtain the total inflow process; it calls the reservoir flood regulation calculation model, based on the initial water level H0 and initial gate opening O set in step S2. g Using the water balance equation With the discharge capacity curve Q=f(H,O) g The solution is obtained through iterative steps, where V is the reservoir capacity, H is the reservoir water level, and O is the reservoir level. g This refers to the gate opening; the output shows the discharge flow rate after reservoir regulation and the reservoir water level change process.

[0122] S403, Solving the two-dimensional hydrodynamic equations (cascade mapping)

[0123] To obtain a detailed water level distribution in key inundation areas, a cascaded computation strategy is adopted, mapping the one-dimensional calculation results of S402 onto the hydrodynamic computation grid generated by S2. The specific implementation process is as follows:

[0124] The outflow from the reservoir nodes or the overflow from the river segment nodes in S402 are mapped to the source-sink term r(t) in the two-dimensional shallow water equations, or used as the inflow boundary condition of the grid boundary. The two-dimensional shallow water equations are discretized and solved on the grid using the finite volume method (FVM): the continuity equation is... The momentum equation is used to describe the motion of water flow under the action of gravity, friction, and inertial forces. The water level H and velocity vector (u,v) of the grid nodes are calculated iteratively at each time step to obtain the dynamic water level sequence of key sections (such as bridge sections and dangerous sections).

[0125] S404. Extract and generate a flood feature parameter vector. Perform data cleaning and feature extraction on the simulation results of S402 and S403, assembling them into a flood feature parameter vector for subsequent decision-making. Specifically:

[0126] Traverse the reservoir inflow or outflow process curves output by S402 and identify the maximum flow value as the flood peak flow value Q. max Record the timestamp corresponding to the maximum flow rate as the peak arrival time T. peak From the calculation results of S403, the set of water level values ​​of key flood control sections (such as the area in front of the dam and major urban sections) over time is extracted as the dynamic water level sequence H of key sections. seq The above indicators are standardized and combined to form a vector V = [Q]. max ,T peak H t1 H t2 [,...] serves as the flood characteristic parameter vector.

[0127] S5. The flood peak flow rate, flood peak arrival time, and dynamic water level sequence of key sections are used as flood characteristic parameter vectors. These vectors are then compared with flood control safety thresholds in a knowledge base. When any indicator in the flood characteristic parameter vector exceeds the limit, early warning information and engineering scheduling plans for the water conservancy project are generated. Specifically, this includes:

[0128] S501. To achieve intelligent decision-making, a flood control scheduling knowledge base containing engineering constraints and business rules is constructed. The specific construction process includes:

[0129] The paper-based reservoir operation regulations, dam safety management regulations, and watershed flood control plans were digitized and analyzed to establish a knowledge graph with "reservoir-river-protected object" as entities. Attributes and relationships were defined within the knowledge graph. The attribute definitions included characteristic water levels (flood control limit, flood control high water level, design flood level) and reservoir capacity curves for the reservoir entity; and safe discharge (e.g., 500m³) for the downstream river section. 3 / s); The relationship definition defines a nonlinear relationship curve Q = f(H,O) between the gate opening and the discharge flow. g Finally, set tiered early warning thresholds. For example, setting a level IV warning when the water level in front of the dam exceeds the flood limit by 0.5m, and a level III warning when it exceeds 1.0m.

[0130] S502, Analyzing the Input Flood Characteristic Parameter Vector and Engineering Status

[0131] The system receives the flood characteristic parameter vector V from step S4 and the real-time engineering and hydrological monitoring data from step S1 in real time, and decomposes V into the predicted peak flow value Q. max Predicting the arrival time of the flood peak T peak and predicting the dynamic water level sequence H of key cross sections seq Extract the current actual reservoir water level H. now and the current gate opening O now As the starting boundary for scheduling calculations, it ensures that the generated scheduling scheme is physically feasible to execute.

[0132] S503. Perform a multi-dimensional comparison between the deconstructed prediction metrics and the security thresholds in the knowledge base, triggering corresponding early warning signals. Specifically:

[0133] Engineering safety early warning: Check whether the predicted reservoir water level sequence exceeds the design flood level, or whether the dam seepage pressure collected in step S1 changes abruptly. If the limits are exceeded, issue a "red warning for engineering safety".

[0134] Flood control capacity early warning: Check the predicted water level H at key downstream sections. seq Does it exceed the guaranteed water level, or the predicted peak flow rate Q? max Does it exceed the safe discharge capacity of the downstream river channel? If it exceeds the limit, issue an "orange warning for flood control and dispatching".

[0135] Warning Issuance: Once a warning is triggered, a pop-up alert will be displayed via a digital twin visualization interface, and a text message will be sent to the relevant flood control personnel. Simultaneously, the critical time window posing the risk (i.e., the time T when the flood peak arrives) will be identified. peak (Before and after).

[0136] S504, Generate Intelligent Reverse Scheduling Scheme

[0137] When an early warning is triggered, an optimal scheduling scheme is automatically generated based on a reverse inference algorithm to eliminate or mitigate the risk. Specifically, the scheduling objective is set as "to control the downstream flow within the safe discharge range while ensuring dam safety." According to T... peak Calculate the available pre-discharge volume V from the current time to the time the flood peak arrives (e.g., 72 hours in advance). pre =(H now -H limit)×A res Based on the discharge capacity curve Q=f(H,O) g The required gate opening O is calculated in reverse. target Generate a sequence of gate operation instructions, for example, opening the No. 2 floodgate to 1.5 meters at time T1 and increasing it to 3.0 meters at time T2; finally, use the engineering data from step S1 to verify the feasibility of the scheme (such as gate opening and closing speed limits and downstream water level variability limits) and eliminate infeasible solutions.

[0138] S505. The generated scheduling scheme is not executed directly, but is first rehearsed in a digital twin scenario. The specific process is as follows:

[0139] The scheduling scheme (gate operation sequence) generated in S504 is re-substituted into the evolution model in step S4 for a second simulation. The simulation is then evaluated to determine whether the future reservoir water level and downstream flow process will return to a safe range after the scheme is implemented. If the simulation results meet the safety requirements, the scheme is marked as a recommended scheme, and a visualized flood control scheduling proposal is generated for the command decision-maker to confirm and issue with one click, thus achieving the final leap from digital space simulation to physical world execution.

[0140] like Figure 2 As shown, the present invention also provides a flood control early warning and simulation device for water conservancy projects based on digital twins. Based on the flood control early warning and simulation method for water conservancy projects based on digital twins as described above, the device includes:

[0141] The data acquisition module 1 is used to construct an integrated three-dimensional sensing and monitoring network of "sky-air-ground-water" for water conservancy hubs to collect multi-dimensional sensing data; wherein, the multi-dimensional sensing data includes meteorological image data, topographic mapping data, and engineering and water condition monitoring data;

[0142] Generation module 2 is used to spatially discretize the topographic mapping data to obtain a hydrodynamic calculation grid, combine the meteorological image data with the ground rain gauge data, and use the optimal interpolation method to generate gridded surface rainfall driving data to form the digital twin calculation boundary conditions.

[0143] Module 3 is used to build a distributed computing topology network as a digital twin coupling model of the water conservancy hub based on the topographic mapping data. The distributed computing topology network is composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes cascaded together. Each upstream small watershed node is equipped with an improved SCS runoff generation and runoff model and an improved SRM snowmelt runoff model.

[0144] Pre-simulation module 4 is used to input the gridded surface rainfall driving data into the digital twin coupling model of the water conservancy hub, use distributed message queue technology to drive the collaborative calculation of each node to obtain the whole basin flow evolution process and extract the flood peak flow value and flood peak arrival time, and load the whole basin flow evolution process into the hydrodynamic calculation grid to obtain the dynamic water level sequence of key sections.

[0145] The early warning module 5 is used to take the flood peak flow value, the flood peak arrival time and the dynamic water level sequence of key sections as flood feature parameter vectors, and compare the flood feature parameter vectors with the flood control safety thresholds in the knowledge base. When any indicator in the flood feature parameter vector exceeds the limit, early warning information and engineering scheduling plan for the water conservancy hub are generated.

[0146] Each of the above modules is used to perform the corresponding steps in the above-mentioned digital twin-based flood control early warning and simulation method for water conservancy hubs. The specific implementation methods are as described in the above-mentioned method embodiments, and will not be repeated here.

[0147] like Figure 3 As shown, the present invention also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores all data required for the process of the digital twin-based flood control early warning and simulation method for water conservancy projects. The network interface is used for communication with external terminals via a network connection. The computer program is executed by the processor to implement the digital twin-based flood control early warning and simulation method for water conservancy projects.

[0148] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment on which the present application is applied.

[0149] An embodiment of this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements any of the above-described methods for flood early warning and simulation of water conservancy hubs based on digital twins.

[0150] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by hardware related to computer program instructions. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. Any references to memory, storage, databases, or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM), such as dynamic RAM (used as main storage) or static RAM (commonly used as cache memory). By way of illustration and not limitation, RAM has various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), and Rambus DRAM (RDRAM).

[0151] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0152] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A flood control early warning and simulation method for water conservancy projects based on digital twins, characterized in that, include: S1. Construct an integrated three-dimensional sensing and monitoring network of "sky-air-ground-water" for water conservancy hubs to collect multi-dimensional sensing data; wherein, the multi-dimensional sensing data includes meteorological image data, topographic mapping data, and engineering and water condition monitoring data; S2. Spatial discretization of the topographic mapping data is performed to obtain a hydrodynamic calculation grid. The meteorological image data is combined with the ground rain gauge data, and the optimal interpolation method is used to generate gridded surface rainfall driving data to form the digital twin calculation boundary conditions. S3. Establish a distributed computing topology network as a digital twin coupling model of the water conservancy hub based on the topographic mapping data; wherein, the distributed computing topology network is composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes cascaded together, and each upstream small watershed node is equipped with an improved SCS runoff generation and runoff model and an improved SRM snowmelt runoff model. S4. Input the gridded surface rainfall driving data into the digital twin coupling model of the water conservancy hub, use distributed message queue technology to drive the collaborative calculation of each node to obtain the whole basin flow evolution process and extract the flood peak flow value and flood peak arrival time, and load the whole basin flow evolution process into the hydrodynamic calculation grid to obtain the dynamic water level sequence of key sections. S5. The flood peak flow rate, flood peak arrival time, and dynamic water level sequence of key sections are used as flood feature parameter vectors. The flood feature parameter vectors are compared with flood control safety thresholds in the knowledge base. When any indicator in the flood feature parameter vector exceeds the limit, early warning information and engineering scheduling plan for the water conservancy hub are generated.

2. The flood control early warning and simulation method for water conservancy hubs based on digital twins according to claim 1, characterized in that, S1 specifically includes: S101. Use high-resolution remote sensing satellites and meteorological satellites to conduct large-scale scanning of water conservancy hubs and watersheds, analyze satellite remote sensing images to extract water body boundaries and snow cover range, and invert atmospheric cloud images and macro-rainfall trends to obtain satellite remote sensing images, snow cover information and atmospheric cloud image information as meteorological image data. S102. Deploy drones and rotary-wing aircraft to conduct oblique photogrammetry of reservoir areas, dam hubs and key river sections to produce high-precision digital elevation models and digital orthophotos. In emergency situations, transmit back video of the disaster site to obtain digital elevation models, digital orthophotos and real-time disaster video information as topographic mapping data. S103. GNSS positioning sensors and piezometers deployed on the dam body and bank slopes are used to monitor structural displacement and internal seepage pressure. Ground rain gauges and water level stations are used to monitor fixed-point precipitation and river water level. Displacement deformation, seepage pressure, rainfall and river water level information are used as the ground part of the engineering and water situation monitoring data. S104. Use underwater robots or fixed underwater sensors to collect underwater flow velocity and water quality parameters at the dam toe and stilling basin, and detect the hidden engineering status of the dam toe sedimentation or scour depth, and obtain underwater flow velocity, water quality parameters and hidden engineering status information as the underwater part of the engineering and water condition monitoring data. S105. The collected meteorological image data, topographic mapping data, and engineering and water condition monitoring data are transmitted to the central server using the communication network. The data is then cleaned and correlated using a unified timestamp and spatial coordinate system to complete the data standardization process.

3. The flood control early warning and simulation method for water conservancy projects based on digital twins according to claim 2, characterized in that, S2 specifically includes: S201. Extract terrain feature constraint lines using the digital orthophoto, generate an unstructured grid based on the terrain feature constraint lines, and map the digital elevation model data to the grid nodes to generate a hydrodynamic calculation grid for solving the two-dimensional shallow water equation. S202. Using GIS spatial analysis to analyze the topographic mapping data, the watershed is divided into multiple small watershed units. The average slope and watershed area of ​​each small watershed unit are calculated as geometric parameters for the improved SCS runoff generation and confluence model. The area-elevation curves of each small watershed unit are extracted to determine the area ratio of the elevation zone as geometric parameters for the improved SRM snowmelt runoff model. S203. Extract the real-time upstream water level and gate opening from the engineering and water situation monitoring data, and set the initial water level and initial gate opening for the two-dimensional hydrodynamic model and the reservoir flood control calculation model respectively. S204. The optimal interpolation method is used to fuse the initial radar precipitation estimate retrieved from the meteorological image data with the ground rain gauge observations to generate gridded areal rainfall driving data. The calculation formula is as follows: Among them, R″ i R′ is the precipitation analysis value generated for the i-th grid point. i Let P be the initial radar precipitation estimate for the i-th grid point, n be the number of rain gauges involved in the calculation, and P be the initial estimate for the i-th grid point. k As a weighting factor, Let k be the observed value from the k-th rain gauge. The initial estimate of radar precipitation at the location of the kth rain gauge station; S205. The hydrodynamic calculation grid, geometric parameters, initial water level, initial gate opening, and gridded surface rainfall driving data are standardized and encapsulated to form the digital twin calculation boundary conditions.

4. The flood control early warning and simulation method for water conservancy projects based on digital twins according to claim 3, characterized in that, In S204, the weighting factor P k The calculation is adaptive based on the distribution density of rain gauges; that is, when the number of rain gauges within the set search radius meets a preset threshold, a dense correlation function is used. Calculate, otherwise use a sparse correlation function. Calculate, where r ij r represents the distance between rain gauges. u This represents the distance between the grid point and the rain gauge, where 'a' is the relevant length parameter.

5. The flood control early warning and simulation method for water conservancy projects based on digital twins according to claim 4, characterized in that, S3 specifically includes: S301. Based on the topographic mapping data analysis of several small watershed units and their topological relationships, establish a distributed computing topology network composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes, and configure a message queue interface for each node to achieve decoupled data transmission. S302. Configure the improved SCS runoff generation and confluence model and the improved SRM snowmelt runoff model in the upstream small watershed nodes, and construct a multi-source runoff coupling mechanism. That is, determine the precipitation pattern according to the real-time temperature. If it is a simple rainfall, only the SCS model is called. If it is a mixture of rain and snow, the rainfall runoff calculated by the SCS model and the snowmelt runoff calculated by the SRM model are linearly superimposed as the total runoff of the node. The improved SCS runoff generation and runoff model uses the SCS-CN curve number method to calculate net rainfall and incorporates the previous soil moisture level to correct the CN value. It uses the triangular generalized unit line method for runoff calculation and sets the time base width of the unit line to 8 / 3 times the time of the flood peak. The improved SRM snowmelt runoff model refines the calculation step to a time scale and introduces a time coefficient for precipitation-induced snowmelt intensity. Its calculation formula is as follows: Among them, Q n+1 For time-period forecast traffic flow, C s Here, is the snowmelt runoff coefficient, 'a' is the degree-day factor, and 'T' is the temperature n Here, ΔT is the temperature, and S is the temperature correction value. n C represents the snow cover rate. r Here, P is the snowmelt runoff coefficient, T0 is the critical snowmelt temperature, A is the catchment area, and k is the runoff coefficient. n+1 This is the flow attenuation coefficient; S303. Initialize the model using the digital twin calculation boundary conditions, inject the average slope of the watershed, the watershed area and the elevation zone distribution parameters into the corresponding upstream small watershed nodes, and establish a binding relationship between the gridded surface rainfall driving data and the input variables of the upstream small watershed nodes.

6. The flood control early warning and simulation method for water conservancy projects based on digital twins according to claim 5, characterized in that, S4 specifically includes: S401. Using GIS spatial analysis, the gridded surface rainfall-driven data is mapped to the upstream small watershed nodes, time-series slices are performed according to the calculation step size, and the data is pushed to the input queues of each node in parallel through a message queue. S402. Utilize distributed message queue technology to drive the various nodes in the topology network constructed in step S3 to run sequentially in order to perform distributed topology network collaborative computing; S403. Using a cascaded calculation strategy, the overflow of the downstream flow process or the overflow of the river section node is mapped to the source and sink terms of the hydrodynamic calculation grid. The two-dimensional shallow water equation set is solved on the grid using the finite volume method to obtain the dynamic water level sequence of the key section. S404. Extract the maximum flow value and its corresponding time from the discharge flow process or river evolution process as the flood peak flow value and flood peak arrival time, and combine them with the dynamic water level sequence to generate a flood characteristic parameter vector.

7. The flood control early warning and simulation method for water conservancy projects based on digital twins according to claim 6, characterized in that, Specifically, S402 includes: After receiving the data, the upstream small watershed node calls the improved SCS and SRM models to calculate rainfall and snowmelt runoff, linearly superimposes them to obtain the unit outlet flow, and encapsulates it into a message to send to the midstream river node. The midstream river node receives confluence messages from the input queue and calculates the river channel evolution process using the Muskingan model; the calculation formula is as follows: In the formula Q out I represents the outflow, C0 represents the inflow, and C1, C2 are the bus parameters. The downstream reservoir node receives the inflow, and based on the initial water level and initial gate opening, the water balance equation is used. With the discharge capacity curve Q=f(H,O) g The iterative solution is performed to obtain the discharge flow process and the reservoir water level change process, where V is the reservoir capacity, H is the reservoir water level, and O is the reservoir level. g This refers to the gate opening degree.

8. The flood control early warning and simulation method for water conservancy hubs based on digital twins according to claim 7, characterized in that, S5 specifically includes: S501. Construct a flood control scheduling knowledge base, establish a knowledge graph with reservoirs, rivers and protected objects as entities, define the relationship between characteristic water levels, reservoir capacity curves, river safe discharge attributes and gate discharge capacity curves, and set graded early warning thresholds. S502. Analyze the flood characteristic parameter vector, decompose it into peak flow value, peak arrival time and key section dynamic water level sequence, and extract the reservoir water level and gate opening from the engineering and water situation monitoring data at the current time as the starting boundary for scheduling calculation. S503. Perform multi-dimensional risk assessment, compare the decomposed prediction indicators with the graded early warning thresholds, and issue a flood warning when the dynamic water level sequence of the key section exceeds the guaranteed water level or the flood peak flow value exceeds the safe discharge of the river channel. Issue an engineering safety warning when the predicted reservoir water level exceeds the design flood level or the dam monitoring data is abnormal. S504. Generate an intelligent scheduling scheme based on the reverse deduction algorithm, calculate the pre-discharge time window according to the arrival time of the flood peak, set the objective function to control the downstream flow within the safe discharge capacity, use the gate discharge capacity curve to solve the gate opening sequence required to satisfy the objective function in reverse, and use the engineering and water condition monitoring data to verify the gate opening and closing constraints. S505. Perform closed-loop simulation, using the gate opening sequence as a new boundary condition to perform a secondary simulation in the distributed computing topology network of step S4, and evaluate the future reservoir water level and downstream flow process after implementing the scheme. If the evaluation result meets the flood control safety requirements, the recommended engineering scheduling scheme is output.

9. A flood control early warning and simulation device for a water conservancy hub based on digital twins, based on the flood control early warning and simulation method for a water conservancy hub based on digital twins as described in any one of claims 1-8, characterized in that, The device includes: The data acquisition module is used to construct an integrated three-dimensional sensing and monitoring network of "sky-air-ground-water" for water conservancy hubs to collect multi-dimensional sensing data; wherein, the multi-dimensional sensing data includes meteorological image data, topographic mapping data, and engineering and water condition monitoring data; The generation module is used to spatially discretize the topographic mapping data to obtain a hydrodynamic calculation grid, combine the meteorological image data with the ground rain gauge data, and use the optimal interpolation method to generate gridded surface rainfall driving data to form the digital twin calculation boundary conditions. A construction module is used to establish a distributed computing topology network as a digital twin coupling model of the water conservancy hub based on the topographic mapping data; wherein, the distributed computing topology network is composed of upstream small watershed nodes, midstream river section nodes and downstream reservoir nodes cascaded together, and each upstream small watershed node is equipped with an improved SCS runoff generation and runoff model and an improved SRM snowmelt runoff model. The pre-simulation module is used to input the gridded surface rainfall driving data into the digital twin coupling model of the water conservancy hub, use distributed message queue technology to drive the collaborative calculation of each node to obtain the whole basin flow evolution process and extract the flood peak flow value and flood peak arrival time, and load the whole basin flow evolution process into the hydrodynamic calculation grid to obtain the dynamic water level sequence of key sections. The early warning module is used to take the flood peak flow value, the flood peak arrival time and the dynamic water level sequence of key sections as flood feature parameter vectors, and compare the flood feature parameter vectors with flood control safety thresholds in the knowledge base. When any indicator in the flood feature parameter vector exceeds the limit, early warning information and engineering scheduling plan for water conservancy projects are generated.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images