A method, system, device and medium for simulating mountain torrent dynamics

Abstract

The present application relates to the technical field of water resource management, and in particular to a method, system, device and medium for simulating the dynamics of mountain torrents. The present application obtains soil data, extracts surface soil moisture data from the soil data, constructs a soil moisture inference model based on the surface soil moisture data, and infers soil infiltration data; obtains elevation data, simplifies the elevation data to obtain simplified terrain data; constructs a simulation grid, obtains geographic information data, and associates both the geographic information data and the simplified terrain data with the simulation grid; determines the flow production and flow confluence of each cell in the simulation grid based on the soil infiltration data and the geographic information data, and updates the cell water volume of the simulation grid based on the flow production and flow confluence, thereby realizing rapid and refined flood modeling in large-scale areas and improving the efficiency of simulating the dynamic mountain torrent evolution process.

Description

Technical Field

[0001] The present application relates to the technical field of water resource management, and in particular to a method, system, equipment and medium for simulating mountain torrent power. Background Art

[0002] Floods are one of the most common and destructive natural disasters in the world, causing extensive and far-reaching impacts on society. At present, the study of floods has become a global effort, and people have been working hard to understand, evaluate and predict flood events and their impacts. Especially flash floods, due to their strong suddenness, are facing the problem of frequent and increasing severity of flash floods all over the world. As a result, people's demand for flash flood simulation in large spaces and refined scales is also increasing. Therefore, fast and accurate flood simulation and prediction models are particularly important. With the development of science and technology, the hydrological and hydrodynamic simulation of floods has begun to develop from the early basic theories and empirical formulas to complex models. With the support of geographic information systems and remote sensing technology, more complex hydrological and hydrodynamic models have been developed.

[0003] Geospatial technology has been widely used in hydrological and hydrodynamic simulation. It not only integrates powerful functions such as spatial data management, analysis and visualization, but also greatly enhances the accuracy and operability of flood simulation. However, there are still some limitations in the existing technology in the process of hydrological and hydrodynamic simulation. Traditional hydrological and hydrodynamic models are often complex and computationally intensive. Especially in large-scale or high-resolution simulations, highly complex and sophisticated models require a lot of computing resources and time for processing and storage. Therefore, it is difficult to meet the needs of rapid simulation in large-scale areas, and there will be certain limitations in their widespread use. The above problems need to be solved. Summary of the invention

[0004] In order to achieve rapid and refined flood modeling in large-scale areas and improve the efficiency of dynamic flash flood evolution process simulation, this application provides a flash flood dynamic simulation method, system, device and medium, which adopts the following technical solutions:

[0005] In a first aspect, the present application provides a method for simulating flash flood dynamics, comprising:

[0006] Acquire soil data, extract surface soil moisture data from the soil data, build a soil moisture estimation model based on the surface soil moisture data, and infer soil infiltration data;

[0007] Acquire elevation data, and simplify the elevation data to obtain simplified terrain data;

[0008] Construct a simulation grid, obtain geographic information data, and associate both the geographic information data and simplified terrain data with the simulation grid;

[0009] According to the soil infiltration data and geographic information data, the flow production and sink of each cell in the simulation grid are determined, and the cellular water volume of the simulation grid is updated according to the flow production and sink.

[0010] Preferably, it also includes:

[0011] Obtain equipment resource conditions, split the soil infiltration data and geographic information data related to each cell according to the equipment resource conditions, and obtain the flow production and sink corresponding to each cell.

[0012] Preferably, constructing a soil moisture estimation model based on surface soil moisture data includes:

[0013]

[0014] Among them, S is the water storage of soil layer d, S0 is the water storage at the depth d0 of the soil surface, A, B, S c is a constant.

[0015] Preferably, the step of inferring the soil infiltration data includes inferring according to a soil moisture inference model to calculate the soil infiltration data, and the specific steps include:

[0016]

[0017] Where i0 is the initial infiltration rate, K s is the saturated hydraulic conductivity of the soil, which is obtained from the soil database according to the soil type, θ is the deep soil moisture content, and θ sat is the saturated moisture content of the soil.

[0018] Preferably, the specific steps of simplifying the elevation data to obtain simplified terrain data are:

[0019] Calculate the elevation data based on the current nearest river elevation to get the HAND value of the current point;

[0020] The HAND values ​​of several current points are combined into simplified terrain data.

[0021] Preferably, the specific steps of constructing the simulation grid are:

[0022] Get the geographic boundary of the study area, divide the range within the geographic boundary into several cells, and assign an identifier to each cell.

[0023] Preferably, the method further includes acquiring historical data, obtaining the initial infiltration rate of each cell according to the historical data, and judging the infiltration capacity of the cell according to the time difference between the current moment and the initial state.

[0024] Preferably, the determining of the flow production and sinking of each cell in the simulation grid, and updating the water volume of the cells of the simulation grid according to the flow production and sinking comprises:

[0025] The water depth and water volume of each cell are calculated based on the calculated soil infiltration rate;

[0026] The cellular water flow direction, transferred water volume, transferred flow rate and transfer time are calculated, and the cellular water volume is updated according to the cellular water flow direction, transferred water volume, transferred flow rate and transfer time.

[0027] Preferably, the specific steps of calculating the soil infiltration rate of each cell at different times include:

[0028]

[0029] Where i is the soil infiltration rate at time t, i f is the stable infiltration rate, that is, the constant infiltration rate after soil saturation, i0 is the initial infiltration rate, β is the decay constant representing the rate at which the infiltration rate decreases over time, and t is the time.

[0030] Preferably, the specific steps of calculating the water depth and water volume of each cell according to the calculated soil infiltration rate include:

[0031]

[0032]

[0033] Where P is the total rainfall, h is the water depth of each cell formed by P, I0 is the initial loss, the loss before runoff, t c is the duration of superosmotic flow, P n is the rainfall during the non-overseepage period in the post-damage stage, A is the area of ​​each cell, V is the water volume of each cell, is the average infiltration rate after the damage.

[0034] Preferably, the water flow direction of the cell is determined by the hydraulic gradient, including:

[0035]

[0036] Where: S ij (t) is the hydraulic slope from cell i to cell j at time t, h i (t) is the water depth in cell i at time t, h j (t) is the water depth in cell j at time t, Z i is the ground elevation of cell i, Z j is the ground elevation of cell j, L ij is the horizontal distance between cell i and cell j.

[0037] In a second aspect, the present application provides a mountain flood dynamics simulation system, comprising:

[0038] The soil analysis module is used to obtain soil data, extract surface soil moisture data from the soil data, build a soil moisture estimation model based on the surface soil moisture data, and infer soil infiltration data;

[0039] The terrain simplification module is used to obtain elevation data and simplify the elevation data to obtain simplified terrain data;

[0040] A grid construction module is used to construct a simulation grid, obtain geographic information data, and associate both the geographic information data and simplified terrain data with the simulation grid;

[0041] The data processing module is used to determine the flow production and sink of each cell in the simulation grid according to precipitation, soil infiltration data and geographic information data, and update the cellular water volume of the simulation grid according to the flow production and sink.

[0042] Preferably, it also includes:

[0043] The resource allocation module is used to obtain the equipment resource status, split the soil infiltration data and geographic information data related to each cell according to the equipment resource status, and obtain the flow production and sink corresponding to each cell.

[0044] In a third aspect, the present application provides a flash flood dynamics simulation device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the flash flood dynamics simulation method as described above.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the mountain flood dynamics simulation method as described above when running.

[0046] In summary, compared with the prior art, the technical solution provided by this application has at least the following beneficial effects:

[0047] The present application constructs a soil moisture inference model by extracting surface soil moisture data from soil data. The soil moisture inference model infers soil infiltration data, obtains simplified terrain data by simplifying elevation data, associates geographic information data and simplified terrain data with the constructed simulation grid, and determines the flow production and flow confluence of each cell in the simulation grid based on the soil infiltration data and geographic information data. The cellular water volume of the simulation grid is updated according to the flow production and flow confluence, thereby realizing rapid and refined flood modeling in large-scale areas and improving the efficiency of simulating the dynamic flash flood evolution process. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a method for simulating flash flood dynamics described in an embodiment of the present application.

[0049] Figure 2 It is a module schematic diagram of a mountain flood dynamics simulation system described in an embodiment of the present application.

[0050] Description of reference numerals:

[0051] 1. Soil analysis module; 2. Terrain simplification module; 3. Grid construction module; 4. Data processing module. DETAILED DESCRIPTION

[0052] The following combination Figure 1-Figure 2 The present application is described in further detail. The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to be limiting.

[0053] Reference Figure 1 The present application relates to a method for simulating flash flood dynamics, which specifically includes:

[0054] Step S1: Acquire soil data, extract surface soil moisture data from the soil data, construct a soil moisture estimation model based on the surface soil moisture data, and infer soil infiltration data;

[0055] Step S2: obtaining elevation data, and simplifying the elevation data to obtain simplified terrain data;

[0056] Step S3: construct a simulation grid, obtain geographic information data, and associate the geographic information data and simplified terrain data with the simulation grid;

[0057] Step S4: Determine the flow production and sink of each cell in the simulation grid according to the soil infiltration data and the geographic information data, and update the cellular water volume of the simulation grid according to the flow production and sink.

[0058] Specifically, the embodiment of the present application extracts the surface soil moisture data from the soil data to construct a soil moisture inference model. The soil data is obtained from a satellite remote sensing platform, and the key features of the surface soil, including soil moisture content, soil type, and surface temperature, are obtained by using advanced remote sensing products provided by satellites such as SMAP, Soil Moisture Active Passive, or Sentinel-1. Based on the vertical distribution and change characteristics of soil moisture, a model for inferring deep soil moisture based on surface soil moisture data is constructed. The dynamic change characteristics of soil moisture at different depths are comprehensively considered, so that the moisture state of deep soil can be effectively estimated, and the soil moisture inference model infers soil infiltration data. Simplified terrain data is then obtained by simplifying the elevation data, and the geographic information data and simplified terrain data are associated with the constructed simulation grid. Based on the soil infiltration data and the geographic information data, the flow production and flow confluence of each cell in the simulation grid are determined, and the cell water volume of the simulation grid is updated according to the flow production and flow confluence, so as to realize rapid and refined flood modeling in large-scale areas and improve the efficiency of dynamic flash flood evolution process simulation.

[0059] The characteristics of soil moisture changes include that the change between the soil moisture storage of a certain deep soil layer and the surface soil moisture storage is close to a linear relationship, that is, as the surface moisture storage increases, the moisture storage of the following deep layers also increases, but the speed of moisture increase in soil layers at different depths is different, and the deeper the soil layer, the slower the moisture increases. The change of soil moisture with depth is a curve, but as the surface soil moisture decreases, the change of soil moisture with depth is close to a linear relationship.

[0060] As one implementation method, constructing a soil moisture estimation model based on surface soil moisture data includes:

[0061]

[0062] Among them, S is the water storage of soil layer d, S0 is the water storage at the depth d0 of the soil surface, A, B, S c is a constant.

[0063] Specifically, the model is proposed based on the determination of surface soil moisture to infer deep soil moisture, which is transformed into the following formula:

[0064]

[0065] Through the multiple linear regression method, A, B, S in the formula can be determined c , thus the water storage of any soil layer can be obtained from the surface soil water storage and the soil layer depth d.

[0066] As one implementation method, inferring the soil infiltration data includes inferring according to a soil moisture inference model to calculate the soil infiltration data, and the specific steps include:

[0067]

[0068] Where i0 is the initial infiltration rate, K s is the saturated hydraulic conductivity of the soil, which is obtained from the soil database according to the soil type, θ is the deep soil moisture content, and θ sat is the saturated moisture content of the soil.

[0069] Specifically, the soil surface moisture content is obtained through remote sensing technology to infer the deep soil moisture content, which is then used to calculate the initial infiltration rate i0. The saturated hydraulic conductivity of the soil is obtained from the soil database according to the soil type, and the saturated moisture content of the soil is usually also obtained from the soil database according to the soil type.

[0070] As one of the implementation methods, the specific steps of simplifying the elevation data to obtain simplified terrain data are: calculating the elevation data according to the current altitude of the nearest river channel to obtain the HAND value of the current point; and combining the HAND values ​​of several current points into simplified terrain data.

[0071] Specifically, HAND stands for Height Above Nearest Drainage, which is a local relative terrain with the river system as a reference. It is mathematically defined as the difference in elevation between any position on the ground and the point where the water flows from the hillside into the river network, describing the relative height of any position on the ground compared to the local water system. The nearest neighbor river channel is located at the intersection of the slope confluence path and the river network confluence path, so HAND is equal to the height difference of the starting point of the slope confluence path, which is essentially the head of the slope confluence process. The formula is:

[0072]

[0073] Among them, H HAND is the HAND value of the current point, Z C is the altitude of the current point, Z ND It is the altitude of the river channel nearest to the current point. The nearest neighbor here refers to the closest along the confluence path, not the closest in straight line distance.

[0074] The HAND method simplifies complex terrain by calculating the relative elevation of each point in the terrain relative to the local lowest point, and provides an intuitive way to identify rainwater collection areas. This facilitates the construction of the two-dimensional cellular scene in this application.

[0075] As one implementation method, the specific steps of constructing a simulation grid are:

[0076] Get the geographic boundary of the study area, divide the range within the geographic boundary into several cells, and assign an identifier to each cell.

[0077] Specifically, construct a two-dimensional cellular geographic scene of the surface, obtain the basin geographic information of each cell according to the underlying surface type and simplified terrain data of the study area, and define the cell size, neighborhood relationship, and model boundary of the geographic scene model. Define the geographic boundaries and scope of the study area. Divide the study area into a two-dimensional grid, and each grid unit is called a cell. Determine the resolution of the grid, that is, the size of each cell, usually in meters. After creating the grid, assign a unique identifier or coordinate to each cell for simulation and data recording. Collect geographic information data of the study area, including underlying surface conditions, simplified terrain data, precipitation data, soil data monitored based on remote sensing technology, etc. Associate these data with the simulation grid to ensure that each cell has corresponding geographic attribute data.

[0078] As one implementation method, it also includes acquiring historical data, obtaining the initial infiltration rate of each cell according to the historical data, and judging the infiltration capacity of the cell according to the time difference between the current moment and the initial state.

[0079] Specifically, the initial state is defined for each cell, including information such as initial soil moisture content, initial outflow, and model boundaries. The initial state is generally determined based on historical data, measured data, or model prediction data. Boundary conditions are defined as "normal" cells that allow water flow, "open" outlet cells, representing drainage points or river outlets, and "closed" boundary cells, representing watershed boundaries or ridgelines, which determine whether water can flow into or out of a specific cell. The neighborhood relationship is determined to be a Moore-type eight-neighborhood, and the flow rule driven by elevation difference is used to simulate the flow of water from high to low.

[0080] As one implementation method, determining the flow generation and sinking of each cell in the simulation grid, and updating the water volume of the cell of the simulation grid according to the flow generation and sinking includes:

[0081] The water depth of each cell is calculated based on the calculated soil infiltration rate;

[0082] The cellular water flow direction, transferred water volume, transferred flow rate and transfer time are calculated, and the cellular water volume is updated according to the cellular water flow direction, transferred water volume, transferred flow rate and transfer time.

[0083] As one of the implementation methods, the rules for the state change of each cell are defined based on the physical characteristics of the watershed and the dynamics of the water flow. This includes determining key hydrological parameters such as flow direction, flow velocity, and water depth, as well as setting boundary conditions for watershed simulation. Through the integrated calculation and analysis of runoff generation and confluence, the state of each cell is dynamically updated, and the flow information is also calculated and updated.

[0084] As one implementation method, the specific steps of calculating the soil infiltration rate of each cell at different times include:

[0085]

[0086] Where i is the soil infiltration rate at time t, i f is the stable infiltration rate, that is, the constant infiltration rate after soil saturation, i0 is the initial infiltration rate, β is the decay constant representing the rate at which the infiltration rate decreases over time, and t is the time.

[0087] Specifically, the flow rate transferred from each cell to its neighboring cells is calculated. This involves considering the water flow connection and possible flow restrictions between cells. The water depth of each cell is obtained by using the initial loss and subsequent loss method to calculate the runoff: the initial loss and subsequent loss method simplifies the infiltration into two parts: initial loss and subsequent loss. The water loss from the beginning of rainfall to the occurrence of runoff is represented by the initial loss I0; the water loss after runoff is the subsequent loss, which is the average infiltration intensity during the runoff duration. , mm / min, and flow duration t c The product of and the amount of rainfall P that does not produce runoff in the later period n sum.

[0088] As one implementation method, the specific steps of calculating the water depth and water volume of each cell according to the calculated soil infiltration rate include:

[0089]

[0090]

[0091] Where P is the total rainfall, h is the water depth of each cell formed by P, I0 is the initial loss, the loss before runoff, t c is the duration of superosmotic flow, P n is the rainfall during the non-overseepage period in the post-damage stage, A is the area of ​​each cell, V is the water volume of each cell, is the average infiltration rate after the damage.

[0092] As one implementation method, the water flow direction of the cell is determined by the hydraulic gradient, including:

[0093]

[0094] Where: S ij(t) is the hydraulic slope from cell i to cell j at time t, h i(t) is the water depth in cell i at time t, h j(t) is the water depth in cell j at time t, Z i is the ground elevation of cell i, Zj is the ground elevation of cell j, L ij is the horizontal distance between cell i and cell j.

[0095] Specifically, the process of runoff generation and confluence in the basin is simulated. Starting from the initial state, the state change of each cell is calculated step by step according to the defined state change rules. Consider factors such as rainfall, terrain influence, soil infiltration, surface runoff, etc. to simulate the flow and accumulation of water. The simulation process can be calculated and output according to the time step to observe the change of the basin state over time. For example, the confluence and accumulation of water are determined according to the set conversion rules. The direction of water flow between cells is determined by the dynamic net rainfall of each grid. By comparing the water level of the central cell with its neighboring cells, adjacent cells below the average water level are identified to determine the potential water flow transfer path. Subsequently, the flow direction and slope of each cell are calculated. The flow direction is based on the principle that water flows naturally from highlands to lowlands, and the slope is determined by the elevation difference of adjacent cells.

[0096] Based on the Moore-type eight-neighborhood model and combined with water level and elevation difference, the flow direction and flow distribution weight can be determined. The specific steps are as follows: Assume that the central cell elevation is H c =h c +Z c , the neighborhood cell height is H i =h i +Z i , if ΔH c→i =H c -H i >0, it means that the central cell C can flow to the neighboring cell i.

[0097] For a certain flow direction, calculate the amount of water flowing. If the potential energy difference between the central cell and the domain cell is 0 or the central cell has insufficient water supply, the water stops flowing. This can be determined by the following formula:

[0098]

[0099] It shows that the amount of water flowing to the neighboring cells is determined by the potential flow capacity between the central cell and the neighboring cells, that is, the elevation difference, and the actual flowable water volume of the central cell, that is, the smaller value of the water depth. This prevents excessive water from being distributed beyond physical limitations, such as the elevation difference and the water depth of the central cell.

[0100] Repeat the above steps until the water depth of the central cell no longer changes or the preset number of iterations is reached.

[0101] For all potential downstream cells, calculate their water transfer weights:

[0102]

[0103] Where N is the set of all potential downstream cells.

[0104] According to the calculated weight, the water volume of the central cell c is distributed To each downstream cell:

[0105]

[0106] Among them, h c is the ground elevation of the center cell c, Z c is the water depth of the central cell c, h i is the ground elevation of the domain cell i, Z i is the water depth of domain cell i, V c→i is the amount of water flowing from the central cell to the domain cell i, m 3 , is the amount of water m contained in the central cell c at time t 3 , is the weight of the amount of water transferred from the central cell c to the downstream cell i at time t+Δt.

[0107] The determined flow rate is the volume of water flowing through a certain cross section per unit time. The flow rate is calculated according to the Manning formula and then multiplied by the cross-sectional area of ​​the water flow to obtain the flow rate:

[0108]

[0109] Where n is the roughness coefficient, h i is the water depth, b is the width of a single cell, S i is the slope, S i ≥0;

[0110] Q c→i is the flow from the central cell c to the neighboring cell i, m 3 / s.

[0111] Transfer time: For each central cell, the time required to transfer the flow in each direction to the next cell is calculated based on the water volume and flow in eight directions. The time required for water to reach the adjacent area cells from the central cell can be expressed as follows:

[0112]

[0113] Among them, V c→i is the amount of water transferred to the i direction, Q c→i is the flow transmitted to direction i, m 3 / s;

[0114] For each central cell, find the maximum time in the eight directions, i.e., t max =max(t c→i), record the maximum time value t of each central cell max For a simulation time step, take the maximum time value t of all central cells max The median of is taken as the time step Δt.

[0115] Move the propagated water volume from the upstream cell to the downstream cell. For each time step, update the water volume in each cell. Calculate the water depth and flow velocity of the cell at time t+Δt, specifically:

[0116] For cell i, the amount of water exchanged in the next time step Δt at time t is the sum of the amount of water transferred as the central cell and the amount of water obtained as the downstream cell of other cells. Therefore, the water depth of cell i at time t+Δt is:

[0117]

[0118] in, is the water depth at time t+Δt, is the water depth at time t, is the total amount of water flowing into cell i within time Δt, is the total amount of water flowing out of cell i during time Δt.

[0119] In the next time step Δt at time t, the flow velocity of the central cell is the vector sum of the flow velocities from the central cell to its downstream cells, that is:

[0120]

[0121] in, is the flow velocity from cell i to its downstream cell j at time t, Z i , Z j is the surface elevation of cell i and its downstream cell j. The transfer factor is adjusted according to the updated water depth to reflect the impact of infiltration on water depth and flow. The above process is repeated at each time step, and the flow transfer is iterated until the entire simulation period is completed.

[0122] As one implementation method, the equipment resource situation is obtained, and the soil infiltration data and geographic information data related to each cell are split according to the equipment resource situation to obtain the flow production and flow confluence corresponding to each cell.

[0123] Specifically, the computational rules of cellular automata mainly focus on the interaction between a single cell and its neighboring cells. This mechanism realizes the parallelization of the entire computational process at the algorithmic level. Therefore, the embodiment of the present application will utilize the GPU parallel acceleration algorithm to significantly improve the computational efficiency of cellular automata and speed up the generation of simulation results. The GPU graphics processor is a high-performance parallel computing device with a large number of computing cores and high-bandwidth memory, which is suitable for processing large-scale parallel computing tasks. This embodiment will give full play to the computing power of the GPU and improve the efficiency and performance of cellular automaton simulation through reasonable data parallelization, memory optimization, core algorithm optimization, parallel scheduling and synchronization. The following are general steps to improve the computational efficiency of cellular automata using GPU acceleration:

[0124] Data parallelization decomposes the computational tasks of cellular automata into multiple independent subtasks, each of which corresponds to a separate computing core. For cellular automata simulation, the entire cellular grid is divided into multiple sub-cells. The computational processes between different cells can be assigned to different GPU cores for parallel processing. Multiple cores can simultaneously update the states of the cellular regions they are responsible for, thus achieving data parallel processing.

[0125] Hardware resource management, when conducting cellular automaton flood simulation, the key to hardware resource management is to reasonably manage and allocate computing resources according to the actual GPU device resource conditions, including the number of cores, memory capacity, and processing speed. This involves using multi-GPU parallel computing technology to reasonably allocate the integrated flow generation and convergence computing tasks to multiple GPU devices so that they can be processed simultaneously. This approach not only improves computing power, but also optimizes overall computing efficiency and performance.

[0126] Core algorithm optimization, when optimizing the core algorithm on the GPU, the key is to fine-tune the computationally intensive parts of the cellular automaton model to maximize the computational efficiency of the GPU. This embodiment uses efficient algorithms such as fast Fourier transform FFT to optimize the update process of the cellular state. FFT is suitable for processing pattern recognition and signal processing tasks in large data sets, and can significantly reduce the calculation time. In addition, for the parallel architecture of the GPU, this embodiment reconstructs the calculation process to reduce redundant calculations by reducing memory access and optimizing thread management. Including using shared memory to reduce the number of global memory accesses, and reasonably arranging threads to reduce data dependencies and waiting times between threads. Such optimization is intended to reduce invalid calculations and improve the overall computing utilization and performance of the GPU.

[0127] Parallel scheduling and synchronization, achieving efficient parallel computing scheduling and synchronization under the GPU architecture is crucial to ensuring data integrity and reducing computing conflicts. By adopting the CUDA, Compute Unified Device Architecture, programming framework, fine parallel task scheduling and inter-thread synchronization can be achieved. The CUDA architecture allows precise task allocation to GPU cores and organizes computing tasks through the concepts of thread blocks and grids. In addition, the synchronization primitives provided by CUDA, such as thread barriers and atomic operations, can be used to control data access and updates between threads to ensure the atomicity and consistency of operations. This parallel scheduling and synchronization mechanism not only optimizes computing efficiency in a highly parallel GPU environment, but also enhances the accuracy of data processing, which is particularly critical for complex data sets and intensive computing tasks.

[0128] Memory optimization. In GPU-accelerated cellular automaton simulation, memory access optimization is crucial to ensure the efficiency of the computation process. This requires strategic optimization in the GPU memory architecture, especially through the application of shared memory and texture memory. For each cell and its neighborhood in the cellular automaton, storing frequently accessed state data in shared memory can significantly reduce the number of accesses to the slow global memory. Shared memory provides faster data access speed when processing cell states within a thread block, optimizing the efficiency of cell state updates. At the same time, considering the spatial data access pattern in the cellular automaton model, texture memory can be used to store and access grid data, and its caching mechanism can be used to improve the data access efficiency of spatial and temporal locality. In addition, optimizing the access to global memory, such as through data alignment and reasonable layout, reduces memory fragmentation and bandwidth consumption, and further improves the computing performance of cellular automata on the GPU. Through this hierarchical memory management strategy, the execution of cellular automaton simulation on the GPU is not only more efficient, but also more adaptable to the processing needs of large-scale data sets.

[0129] Parameters are set and calibrated according to the specific problem and data reliability, and attention is paid to the impact of the simulation time step and the numerical approximation method used in the model on the results, and the simulation results are dynamically visualized. For example, the total runoff of the basin, the runoff distribution at each time, and the flooded depth are calculated. This step specifically includes:

[0130] Parameter setting and preliminary calibration, depending on the simulation purpose and available data, determine the model parameters, such as terrain characteristics, soil type, etc. Perform preliminary calibration to ensure that the model reflects the actual hydrological processes.

[0131] The model is calibrated and validated using actual observed data. Parameters are adjusted until the model output is as close as possible to the actual observed data.

[0132] Visual dynamic display,In the implementation of the hydrological model, it is crucial to visualize the dynamic display of simulation results.,This process involves collecting simulation output data, including overall runoff,,time series distribution of runoff, and flooding depth in different spatial areas.,These data are then used for in-depth analysis to identify key hydrological features and potential problems,,as well as to conduct comparative analysis of results under different hydrological scenarios.

[0133] In order to effectively display complex data sets, this embodiment implements a visualization strategy that integrates spatiotemporal scales, aiming to construct a timeline by constructing a dynamic frame sequence corresponding to the simulation time point. On this basis, the inundation map of each time step and the relevant hydrological parameters are integrated into the dynamic display of the time series. Furthermore, this embodiment finely configures the animation parameters, including frame duration, total animation duration and frame rate, to optimize the visual display effect. Finally, using multiple media such as maps, charts and animations, the spatial distribution of the inundation range, inundation depth and flow velocity in the entire basin, as well as the dynamic change process of the inundation depth at certain key nodes are dynamically presented, achieving a highly intuitive and data-rich visual presentation, effectively supporting the analysis and decision-making process of the hydrological model.

[0134] Reference Figure 2 , a mountain flood dynamics simulation system is provided for an embodiment of the present application, the system comprising:

[0135] Soil analysis module 1 is used to obtain soil data, extract surface soil moisture data from the soil data, build a soil moisture estimation model based on the surface soil moisture data, and infer soil infiltration data;

[0136] The terrain simplification module 2 is used to obtain elevation data and simplify the elevation data to obtain simplified terrain data;

[0137] A grid construction module 3 is used to construct a simulation grid, obtain geographic information data, and associate both the geographic information data and simplified terrain data with the simulation grid;

[0138] The data processing module 4 is used to determine the flow production and sink of each cell in the simulation grid according to the precipitation, soil infiltration data and geographic information data, and update the cellular water volume of the simulation grid according to the flow production and sink.

[0139] The resource allocation module is also used to obtain the equipment resource status, and split the soil infiltration data and geographic information data related to each cell according to the equipment resource status to obtain the flow production and sink corresponding to each cell.

[0140] An embodiment of the present application provides a flash flood dynamics simulation device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the flash flood dynamics simulation method as described above.

[0141] An embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned flash flood dynamics simulation method when running.

[0142] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device and product can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0143] In several embodiments provided in this application, it should be understood that the disclosed methods, systems, devices and program products may be implemented in other ways.

[0144] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0145] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for simulating flash flood dynamics, characterized in that: include: Acquire soil data, extract surface soil moisture data from the soil data, build a soil moisture estimation model based on the surface soil moisture data, and infer soil infiltration data; Acquire elevation data, simplify the elevation data to obtain simplified terrain data; construct a simulation grid, acquire geographic information data, and associate both the geographic information data and the simplified terrain data with the simulation grid; According to the soil infiltration data and geographic information data, the flow production and sink of each cell in the simulation grid are determined, the cellular water volume of the simulation grid is updated according to the flow production and sink, and the historical data is obtained. The initial infiltration rate of each cell is obtained according to the historical data, and the infiltration capacity of the cell is determined according to the time difference between the current time and the initial state; Obtain equipment resource conditions, split soil infiltration data and geographic information data related to each cell according to the equipment resource conditions, and obtain the flow production and sink corresponding to each cell; Cellular automata use GPU parallel acceleration algorithms, including reasonable data parallelization, memory optimization, core algorithm optimization, parallel scheduling and synchronization; Data parallelization, decomposing the computational tasks of cellular automata into multiple independent subtasks, each of which corresponds to a separate computing core, and assigning the computational processes between different cells to different GPU cores for parallel processing; Hardware resource management: manage and allocate computing resources based on the actual GPU device resources, including the number of cores, memory capacity, and processing speed; Core algorithm optimization, fine-tuning of computationally intensive parts of the cellular automaton model; Parallel scheduling and synchronization: efficient parallel computing scheduling and synchronization under GPU architecture; Memory optimization, storing frequently accessed state data in shared memory, reducing the number of accesses to slow global memory.

2. The method for simulating mountain flood dynamics according to claim 1, characterized in that: The soil moisture estimation model constructed according to the surface soil moisture data comprises: S-S0=S c +A×(d-d0)+S0×B×(d-d0) 2 Among them, S is the water storage of soil layer d, S0 is the water storage at the depth d0 of the soil surface, A, B, S c is a constant.

3. The method for simulating mountain torrent dynamics according to claim 2, characterized in that: The step of inferring the soil infiltration data includes inferring according to the soil moisture inference model to calculate the soil infiltration data, and the specific steps include: Where i0 is the initial infiltration rate, K s is the saturated hydraulic conductivity of the soil, which is obtained from the soil database according to the soil type, θ is the deep soil moisture content, and θ sat is the saturated moisture content of the soil.

4. The method for simulating mountain flood dynamics according to claim 1, characterized in that: The specific steps of simplifying the elevation data to obtain simplified terrain data are: Calculate the elevation data based on the current nearest river elevation to get the HAND value of the current point; The HAND values ​​of several current points are combined into simplified terrain data.

5. The method for simulating mountain flood dynamics according to claim 1, characterized in that: The specific steps of constructing the simulation grid are: Get the geographic boundary of the study area, divide the range within the geographic boundary into several cells, and assign an identifier to each cell.

6. The method for simulating mountain flood dynamics according to claim 1, characterized in that: Determining the flow production and sink flow of each cell in the simulation grid, and updating the water volume of the cell in the simulation grid according to the flow production and sink flow includes: The water depth and water volume of each cell are calculated based on the calculated soil infiltration rate; The cellular water flow direction, transferred water volume, transferred flow rate and transfer time are calculated, and the cellular water volume is updated according to the cellular water flow direction, transferred water volume, transferred flow rate and transfer time.

7. The method for simulating mountain flood dynamics according to claim 1, characterized in that: The specific steps for calculating the soil infiltration rate of each cell at different times include: i=i f +(i0-i f )e (-βt) Where i is the soil infiltration rate at time t, i f is the stable infiltration rate, that is, the constant infiltration rate after soil saturation, i0 is the initial infiltration rate, β is the decay constant representing the rate at which the infiltration rate decreases over time, and t is the time.

8. The method for simulating mountain torrent dynamics according to claim 6, characterized in that: The specific steps of calculating the water depth and water volume of each cell according to the calculated soil infiltration rate include: V=0.001×h×A Where P is the total rainfall, h is the water depth of each cell formed by P, I0 is the initial loss, the loss before runoff, t c is the duration of superosmotic flow, P n is the rainfall during the non-overseepage period in the post-damage stage, A is the area of ​​each cell, V is the water volume of each cell, is the average infiltration rate after the damage.

9. The method for simulating mountain torrent dynamics according to claim 6, characterized in that: The water flow direction of the cell is determined by the hydraulic gradient, including: Where: S ij (t) is the hydraulic slope from cell i to cell j at time t, h i (t) is the water depth in cell i at time t, h j (t) is the water depth in cell j at time t, z i is the ground elevation of cell i, z j is the ground elevation of cell j, L ij is the horizontal distance between cell i and cell j.

10. A mountain flood dynamics simulation system based on the mountain flood dynamics simulation method according to claim 1, characterized in that: include: The soil analysis module is used to obtain soil data, extract surface soil moisture data from the soil data, build a soil moisture estimation model based on the surface soil moisture data, and infer soil infiltration data; The terrain simplification module is used to obtain elevation data and simplify the elevation data to obtain simplified terrain data; A grid construction module is used to construct a simulation grid, obtain geographic information data, and associate both the geographic information data and simplified terrain data with the simulation grid; The data processing module is used to determine the flow production and sink of each cell in the simulation grid according to the precipitation, soil infiltration data and geographic information data, update the cellular water volume of the simulation grid according to the flow production and sink, and obtain historical data, obtain the initial infiltration rate of each cell according to the historical data, and judge the infiltration capacity of the cell according to the time difference between the current time and the initial state; The resource allocation module is used to obtain the equipment resource status, split the soil infiltration data and geographic information data related to each cell according to the equipment resource status, and obtain the flow production and sink corresponding to each cell.

11. A flash flood dynamics simulation device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the mountain flood dynamics simulation method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the mountain flood dynamics simulation method according to any one of claims 1 to 9 when running.

Images