WO2022007398A1 - Unstructured grid flood simulation system based on gpu acceleration thechnology - Google Patents

Unstructured grid flood simulation system based on gpu acceleration thechnology Download PDF

Info

Publication number
WO2022007398A1
WO2022007398A1 PCT/CN2021/075073 CN2021075073W WO2022007398A1 WO 2022007398 A1 WO2022007398 A1 WO 2022007398A1 CN 2021075073 W CN2021075073 W CN 2021075073W WO 2022007398 A1 WO2022007398 A1 WO 2022007398A1
Authority
WO
WIPO (PCT)
Prior art keywords
gpu
system based
grid
simulation system
source term
Prior art date
Application number
PCT/CN2021/075073
Other languages
French (fr)
Chinese (zh)
Inventor
侯精明
李轩
马利平
张兆安
王俊珲
汪煜
王兴桦
Original Assignee
西安理工大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 西安理工大学 filed Critical 西安理工大学
Publication of WO2022007398A1 publication Critical patent/WO2022007398A1/en
Priority to ZA2022/03769A priority Critical patent/ZA202203769B/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/05—Geographic models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/20—Processor architectures; Processor configuration, e.g. pipelining
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/60—Memory management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2200/00—Indexing scheme for image data processing or generation, in general
    • G06T2200/28—Indexing scheme for image data processing or generation, in general involving image processing hardware

Definitions

  • the invention belongs to the technical field of numerical simulation, and relates to an unstructured grid flood process simulation system based on a graphics processor (Graphics Processing Unit, referred to as GPU) acceleration technology.
  • a graphics processor Graphics Processing Unit, referred to as GPU
  • At least some embodiments of the present invention provide an unstructured grid flood process simulation system based on GPU acceleration technology to at least partially solve the problem that GPU acceleration technology in the related art is difficult to apply to unstructured grid for flood simulation .
  • an unstructured grid flood process simulation system based on GPU acceleration technology including:
  • Step 1 Use unstructured triangular meshes to spatially discretize the simulation area to obtain a triangular mesh terrain file
  • Step 2 Read the triangular mesh terrain file in Step 1, and save the node data and element data in the file into a program variable.
  • Step 3 According to the principle that adjacent triangles share the same point number on one side, the topological relationship is established on the triangular mesh terrain file;
  • Step 4 Calculate the relevant variables of each triangular mesh element according to the established topological relationship
  • Step 5 Convert the coordinate values in the node data into the unit data, and initialize the relevant variables
  • Step 6 Allocate memory for the relevant variables in the GPU, and copy the values of the above-mentioned relevant variables into the GPU variables;
  • Step 7 Use the finite volume method based on Godunov format to discretize the two-dimensional shallow water equation, calculate the flux term, bottom slope source term, friction source term and time step respectively on the GPU, and obtain the calculation results;
  • Step 8 Use the cudaMemcpy function to copy the calculation results obtained on the GPU to the host memory, and output the simulated water depth distribution map.
  • the topological relationship in step 3 includes the left mesh number of the current edge, the right mesh number of the current edge, and the properties of the current edge.
  • the relevant variables in step 4 include the area of the triangular mesh, the length of each side of the triangular mesh, the components of the normal unit vector of each side of the triangular mesh in the x-direction and the y-direction, and two adjacent Distance between triangle mesh centroids.
  • step 6 the cudaMalloc of the cuda library function is used to perform variable allocation, and the cudaMemcpy function is used to copy the variable value.
  • the calculation methods of the flux term, the bottom slope source term, the friction source term and the time step in step 7 correspond to the HLLC-based Riemann solver method, the bottom slope flux method, and the semi-implicit method, respectively. and the second-order Runge-Kutta method.
  • the two-dimensional shallow water equation in step 7 is as follows:
  • t is time, in s
  • i is the source term of rainfall and infiltration
  • q is a variable vector including water depth h, single-width flow q x and q y in the two directions of x and y;
  • u and v are the flow velocity in the x and y directions
  • F and G are the flux vectors in the x and y directions
  • S is the source term vector, including rainfall and infiltration source term i, bottom slope source term and frictional resistance source term;
  • z b is the elevation of the bottom of the river bed, in m;
  • the unstructured grid flood process simulation system based on the GPU acceleration technology provided by the embodiment of the present invention has stable calculation, high precision and high simulation efficiency, can realize accurate simulation of the flood process, and obtain a relatively detailed flood inundation map.
  • Emergency decision-making provides strong data support.
  • FIG. 1 is an unstructured triangular mesh discrete topographic map of a certain place of an unstructured mesh flood process simulation system based on GPU acceleration technology according to one embodiment of the present invention.
  • FIG. 2 is a structural flowchart of an unstructured grid flood process simulation system based on GPU acceleration technology according to one embodiment of the present invention.
  • FIG. 3 is a result diagram of simulated water depth distribution of an unstructured grid flood process simulation system based on GPU acceleration technology according to one embodiment of the present invention.
  • an embodiment of the present invention provides an unstructured grid flood process simulation system based on GPU acceleration technology, including the following processing steps:
  • Step 1 Use an unstructured triangular mesh to spatially discretize the simulation area to obtain a triangular mesh terrain file; compared with the structured mesh provided in the related art, the unstructured triangular mesh has a better expression effect on irregular terrain. good.
  • Step 2 Read the triangular mesh terrain file in step 1, and save the node (node) data and element (element) data in the triangular mesh terrain file into a program variable.
  • Step 3 According to the principle that the point numbers of two adjacent triangles sharing one side are the same, establish a topology relationship for the triangle mesh terrain file;
  • Step 4 Calculate the relevant variables of each triangular mesh according to the established topological relationship
  • Step 5 Since the coordinate values are stored in the node data, and the calculation is based on the element data as the research object, the coordinate values in the node data are converted into the element data, and the related data are initialized. variable;
  • Step 6 allocate memory for the relevant variables in step 5 in the GPU, and copy the values of the above-mentioned relevant variables into the GPU variables;
  • Step 7 Use the finite volume method of Godunov format to discretize the two-dimensional shallow water equation, and then calculate the flux term, bottom slope source term, friction source term and time step respectively on the GPU to obtain the calculation result;
  • Step 8 Use the cudaMemcpy function to copy the calculation results obtained on the GPU to the host memory, and output the simulated water depth distribution map, as shown in Figure 3.
  • the topological relationship in step 3 includes the left mesh number of the current edge, the right mesh number of the current edge, and the properties of the current edge.
  • the relevant variables in step 4 include the area of the triangular mesh, the length of each side of the triangular mesh, the components of the normal unit vector of each side of the triangular mesh in the x-direction and the y-direction, and two adjacent Distance between triangle mesh centroids.
  • step 6 the cudaMalloc function of the cuda library function is used for variable allocation, and the cudaMemcpy function is used for copying the variable value.
  • the calculation methods of the flux term, the bottom slope source term, the friction source term and the time step in step 7 correspond to the HLLC-based Riemann solver method, the bottom slope flux method, and the semi-implicit method, respectively. and the second-order Runge-Kutta method.
  • the two-dimensional shallow water equation in step 7 is as follows:
  • t is time, in s
  • i is the source term of rainfall and infiltration
  • q is a variable vector including water depth h, single-width flow q x and q y in the two directions of x and y;
  • u and v are the flow velocity in the x and y directions
  • F and G are the flux vectors in the x and y directions
  • S is the source term vector, including rainfall and infiltration source term i, bottom slope source term and frictional resistance source term;
  • z b is the elevation of the bottom of the river bed, in m;
  • the unstructured grid flood process simulation system based on the GPU acceleration technology provided by the embodiment of the present invention has stable calculation, high precision and high simulation efficiency, and can realize the accurate simulation of the flood process.
  • the GPU parallel acceleration technology is applied on the unstructured grid, The efficiency of model simulation and calculation is improved, and a more detailed flood inundation map is obtained, which provides strong data support for emergency decision-making of flood risk.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • Software Systems (AREA)
  • Computer Graphics (AREA)
  • Remote Sensing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Alarm Systems (AREA)

Abstract

Disclosed in the present invention is an unstructured grid flood simulation system based on GPU acceleration technology. The system comprises the steps: carrying out the spatial dispersion on a simulation region to obtain a triangular grid terrain file; reading the file, and storing nodes and unit data in the file in a program variable; establishing a topological relation for the file according to the principle that node numbers of two adjacent triangles sharing one side are the same; calculating a related variable of each triangular grid according to the established topological relation; converting coordinate values in the nodes into a unit, and initializing each related variable; allocating memories to all the variables in the GPU, and copying the values of the variables into GPU variables; respectively calculating a flux item, a bottom slope source item, a friction resistance source item and time stepping on the GPU; and copying a result value calculated on the GPU into a host memory, and outputting a simulated water depth distribution diagram. The unstructured grid flood simulation system based on GPU acceleration technology provided in the embodiments of the present invention is stable in calculation, high in precision and simulation efficiency, and can realize accurate simulation of the flood process.

Description

一种基于GPU加速技术的非结构网格洪水过程模拟系统An Unstructured Grid Flood Process Simulation System Based on GPU Acceleration Technology 技术领域technical field
本发明属于数值模拟技术领域，涉及一种基于图形处理器(Graphics Processing Unit，简称为GPU)加速技术的非结构网格洪水过程模拟系统。The invention belongs to the technical field of numerical simulation, and relates to an unstructured grid flood process simulation system based on a graphics processor (Graphics Processing Unit, referred to as GPU) acceleration technology.
背景技术Background technique
为应对频繁发生的水灾害问题，世界各国均建立了适应自己国情的洪水管理体系，发达国家由于起步较早，所建立的洪水管理体系较为完善，如美国、日本等；而发展中国家的洪水管理体系则相对落后一些。在洪水管理体系中，对洪水灾害的预警预报是至关重要的，而水动力学模型正是预报洪水灾害的有力技术手段。近年来，众多学者参与到水动力学模型的研究开发中，涌现出众多的一维、二维、三维水动力学模型，由于大范围洪水具有模拟范围广、洪水淹没面积大的特点，二维水动力学数值模型便脱颖而出。In order to deal with the frequent occurrence of water disasters, all countries in the world have established flood management systems adapted to their own national conditions. Developed countries, such as the United States and Japan, have established relatively complete flood management systems due to their early start. The management system is relatively backward. In the flood management system, the early warning and forecasting of flood disasters is very important, and the hydrodynamic model is a powerful technical means for forecasting flood disasters. In recent years, many scholars have participated in the research and development of hydrodynamic models, and many one-dimensional, two-dimensional and three-dimensional hydrodynamic models have emerged. The hydrodynamic numerical model stands out.
在二维水动力学模型中，常见的有结构网格和非结构网格。由于结构网格拓扑关系简单，操作处理方便，部分学者将GPU加速技术应用于结构网格。但是结构网格难以处理地形的复杂边界，而非结构网格便可很好地解决该问题，但非结构网格拓扑关系较为复杂。因此，建立一种基于GPU加速技术的非结构网格洪水过程模拟系统是十分必要的。In two-dimensional hydrodynamic models, there are common structured grids and unstructured grids. Due to the simple topology relationship and convenient operation and processing of structural grids, some scholars apply GPU acceleration technology to structural grids. However, the structured grid is difficult to deal with the complex boundary of the terrain, and the unstructured grid can solve this problem well, but the topology relationship of the unstructured grid is more complicated. Therefore, it is very necessary to establish an unstructured grid flood process simulation system based on GPU acceleration technology.
发明内容SUMMARY OF THE INVENTION
本发明至少部分实施例提供了一种基于GPU加速技术的非结构网格洪水过程模拟系统，以至少部分地解决相关技术中存在的GPU加速技术难以 应用于非结构网格来进行洪水模拟的问题。At least some embodiments of the present invention provide an unstructured grid flood process simulation system based on GPU acceleration technology to at least partially solve the problem that GPU acceleration technology in the related art is difficult to apply to unstructured grid for flood simulation .
在本发明其中一实施例中，提供了一种基于GPU加速技术的非结构网格洪水过程模拟系统，包括：In one embodiment of the present invention, an unstructured grid flood process simulation system based on GPU acceleration technology is provided, including:
步骤1、采用非结构三角形网格对模拟区域进行空间离散，得到三角形网格地形文件；Step 1. Use unstructured triangular meshes to spatially discretize the simulation area to obtain a triangular mesh terrain file;
步骤2、读取步骤1中的三角形网格地形文件，并将文件中的节点(node)数据和单元(element)数据保存至程序变量中。Step 2: Read the triangular mesh terrain file in Step 1, and save the node data and element data in the file into a program variable.
步骤3、依据相邻三角形共用一条边的点编号相同原则对三角形网格地形文件建立拓扑关系；Step 3. According to the principle that adjacent triangles share the same point number on one side, the topological relationship is established on the triangular mesh terrain file;
步骤4、根据建立的拓扑关系计算每个三角形网格单元的相关变量；Step 4. Calculate the relevant variables of each triangular mesh element according to the established topological relationship;
步骤5、将节点数据中的坐标值转换到单元数据中，并初始化相关变量；Step 5. Convert the coordinate values in the node data into the unit data, and initialize the relevant variables;
步骤6、在GPU中为相关变量分配内存，并将上述相关变量的值复制到GPU变量中；Step 6. Allocate memory for the relevant variables in the GPU, and copy the values of the above-mentioned relevant variables into the GPU variables;
步骤7、采用基于Godunov格式的有限体积法离散二维浅水方程，在GPU上分别计算通量项、底坡源项、摩阻源项和时间步进，并得到计算结果；Step 7. Use the finite volume method based on Godunov format to discretize the two-dimensional shallow water equation, calculate the flux term, bottom slope source term, friction source term and time step respectively on the GPU, and obtain the calculation results;
步骤8、采用cudaMemcpy函数将GPU上得到的计算结果复制到主机内存中，并输出模拟的水深分布图。Step 8. Use the cudaMemcpy function to copy the calculation results obtained on the GPU to the host memory, and output the simulated water depth distribution map.
在一些实施例中，步骤3中拓扑关系包括当前边的左侧网格编号，当前边的右侧网格编号，以及当前边的属性。In some embodiments, the topological relationship in step 3 includes the left mesh number of the current edge, the right mesh number of the current edge, and the properties of the current edge.
在一些实施例中，步骤4中相关变量包括三角形网格的面积、三角形网格各边的长度，三角形网格各边的法线单位向量在x方向及y方向的分量，以及相邻两个三角形网格形心之间的距离。In some embodiments, the relevant variables in step 4 include the area of the triangular mesh, the length of each side of the triangular mesh, the components of the normal unit vector of each side of the triangular mesh in the x-direction and the y-direction, and two adjacent Distance between triangle mesh centroids.
在一些实施例中，步骤6中采用cuda库函数的cudaMalloc进行变量分 配，采用cudaMemcpy函数进行变量值的复制。In some embodiments, in step 6, the cudaMalloc of the cuda library function is used to perform variable allocation, and the cudaMemcpy function is used to copy the variable value.
在一些实施例中，步骤7中通量项、底坡源项、摩阻源项和时间步进的计算方法分别对应基于HLLC的黎曼求解器法、底坡通量法、半隐式法和二阶龙格库塔法。In some embodiments, the calculation methods of the flux term, the bottom slope source term, the friction source term and the time step in step 7 correspond to the HLLC-based Riemann solver method, the bottom slope flux method, and the semi-implicit method, respectively. and the second-order Runge-Kutta method.
在一些实施例中，步骤7中二维浅水方程如下：In some embodiments, the two-dimensional shallow water equation in step 7 is as follows:
Figure PCTCN2021075073-appb-000001
Figure PCTCN2021075073-appb-000001
Figure PCTCN2021075073-appb-000002
Figure PCTCN2021075073-appb-000002
式中：where:
t为时间，单位s；t is time, in s;
i为降雨及下渗源项；i is the source term of rainfall and infiltration;
q为变量矢量包括水深h，x、y两个方向上的单宽流量q x和q y； q is a variable vector including water depth h, single-width flow q x and q y in the two directions of x and y;
u、v为x、y方向上的流速；u and v are the flow velocity in the x and y directions;
F、G为x、y方向上的通量矢量；F and G are the flux vectors in the x and y directions;
S为源项矢量，包括降雨及下渗源项i、底坡源项及摩阻力源项；S is the source term vector, including rainfall and infiltration source term i, bottom slope source term and frictional resistance source term;
z b为河床底面高程，单位m； z b is the elevation of the bottom of the river bed, in m;
C f为床面摩擦系数，C f＝gn 2/h 1/3，其中n为曼宁系数。 C f is the bed surface friction coefficient, C f =gn 2 /h 1/3 , where n is the Manning coefficient.
本发明实施例的有益效果如下：The beneficial effects of the embodiments of the present invention are as follows:
本发明实施例提供的基于GPU加速技术的非结构网格洪水过程模拟系统计算稳定、精度和模拟效率较高，可实现对洪水过程的精确模拟，得到较为详细的洪水淹没图，为洪水风险的应急决策提供有力的数据支撑。The unstructured grid flood process simulation system based on the GPU acceleration technology provided by the embodiment of the present invention has stable calculation, high precision and high simulation efficiency, can realize accurate simulation of the flood process, and obtain a relatively detailed flood inundation map. Emergency decision-making provides strong data support.
附图说明Description of drawings
图1是根据本发明其中一实施例的基于GPU加速技术的非结构网格洪水过程模拟系统的某处非结构三角形网格离散地形图。FIG. 1 is an unstructured triangular mesh discrete topographic map of a certain place of an unstructured mesh flood process simulation system based on GPU acceleration technology according to one embodiment of the present invention.
图2是根据本发明其中一实施例的基于GPU加速技术的非结构网格洪水过程模拟系统的结构流程图。FIG. 2 is a structural flowchart of an unstructured grid flood process simulation system based on GPU acceleration technology according to one embodiment of the present invention.
图3是根据本发明其中一实施例的基于GPU加速技术的非结构网格洪水过程模拟系统的模拟水深分布结果图。FIG. 3 is a result diagram of simulated water depth distribution of an unstructured grid flood process simulation system based on GPU acceleration technology according to one embodiment of the present invention.
具体实施方式detailed description
下面结合附图和具体实施方式对本发明进行详细说明。The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
如图1和图2所示，本发明其中一实施例提供了一种基于GPU加速技术的非结构网格洪水过程模拟系统，包括以下处理步骤：As shown in FIG. 1 and FIG. 2 , an embodiment of the present invention provides an unstructured grid flood process simulation system based on GPU acceleration technology, including the following processing steps:
步骤1、采用非结构三角形网格对模拟区域进行空间离散，得到三角形网格地形文件；与相关技术中所提供的结构网格相比，该非结构三角形网格对不规则地形的表达效果更佳。Step 1. Use an unstructured triangular mesh to spatially discretize the simulation area to obtain a triangular mesh terrain file; compared with the structured mesh provided in the related art, the unstructured triangular mesh has a better expression effect on irregular terrain. good.
步骤2、读取步骤1中的三角形网格地形文件，并将三角形网格地形文件中的节点(node)数据和单元(element)数据保存至程序变量中。Step 2: Read the triangular mesh terrain file in step 1, and save the node (node) data and element (element) data in the triangular mesh terrain file into a program variable.
步骤3、依据两个相邻三角形共用一条边的点编号相同原则对三角形网格地形文件建立拓扑关系；Step 3. According to the principle that the point numbers of two adjacent triangles sharing one side are the same, establish a topology relationship for the triangle mesh terrain file;
步骤4、根据建立的拓扑关系计算每个三角形网格的相关变量；Step 4. Calculate the relevant variables of each triangular mesh according to the established topological relationship;
步骤5、由于坐标值保存在节点(node)数据中，而计算时是以单元(element)数据为研究对象进行计算的，因此，将节点数据中的坐标值转换到单元数据中，并初始化相关变量；Step 5. Since the coordinate values are stored in the node data, and the calculation is based on the element data as the research object, the coordinate values in the node data are converted into the element data, and the related data are initialized. variable;
步骤6、在GPU中为步骤5中的相关变量分配内存，并将上述相关变量的值复制到GPU变量中；Step 6, allocate memory for the relevant variables in step 5 in the GPU, and copy the values of the above-mentioned relevant variables into the GPU variables;
步骤7、采用Godunov格式的有限体积法离散二维浅水方程，然后在GPU上分别计算通量项、底坡源项、摩阻源项和时间步进，得到计算结果；Step 7. Use the finite volume method of Godunov format to discretize the two-dimensional shallow water equation, and then calculate the flux term, bottom slope source term, friction source term and time step respectively on the GPU to obtain the calculation result;
步骤8、采用cudaMemcpy函数将GPU上得到的计算结果复制到主机内存中，并输出模拟的水深分布图，如图3所示。Step 8. Use the cudaMemcpy function to copy the calculation results obtained on the GPU to the host memory, and output the simulated water depth distribution map, as shown in Figure 3.
在一些实施例中，步骤3中拓扑关系包括当前边的左侧网格编号，当前边的右侧网格编号，以及当前边的属性。In some embodiments, the topological relationship in step 3 includes the left mesh number of the current edge, the right mesh number of the current edge, and the properties of the current edge.
在一些实施例中，步骤4中相关变量包括三角形网格的面积、三角形网格各边的长度，三角形网格各边的法线单位向量在x方向及y方向的分量，以及相邻两个三角形网格形心之间的距离。In some embodiments, the relevant variables in step 4 include the area of the triangular mesh, the length of each side of the triangular mesh, the components of the normal unit vector of each side of the triangular mesh in the x-direction and the y-direction, and two adjacent Distance between triangle mesh centroids.
在一些实施例中，步骤6中采用cuda库函数的cudaMalloc进行变量分配，采用cudaMemcpy函数进行变量值的复制。In some embodiments, in step 6, the cudaMalloc function of the cuda library function is used for variable allocation, and the cudaMemcpy function is used for copying the variable value.
在一些实施例中，步骤7中通量项、底坡源项、摩阻源项和时间步进的计算方法分别对应基于HLLC的黎曼求解器法、底坡通量法、半隐式法和二阶龙格库塔法。In some embodiments, the calculation methods of the flux term, the bottom slope source term, the friction source term and the time step in step 7 correspond to the HLLC-based Riemann solver method, the bottom slope flux method, and the semi-implicit method, respectively. and the second-order Runge-Kutta method.
在一些实施例中，步骤7中二维浅水方程如下：In some embodiments, the two-dimensional shallow water equation in step 7 is as follows:
Figure PCTCN2021075073-appb-000003
Figure PCTCN2021075073-appb-000003
Figure PCTCN2021075073-appb-000004
Figure PCTCN2021075073-appb-000004
式中：where:
t为时间，单位s；t is time, in s;
i为降雨及下渗源项；i is the source term of rainfall and infiltration;
q为变量矢量包括水深h，x、y两个方向上的单宽流量q x和q y； q is a variable vector including water depth h, single-width flow q x and q y in the two directions of x and y;
u、v为x、y方向上的流速；u and v are the flow velocity in the x and y directions;
F、G为x、y方向上的通量矢量；F and G are the flux vectors in the x and y directions;
S为源项矢量，包括降雨及下渗源项i、底坡源项及摩阻力源项；S is the source term vector, including rainfall and infiltration source term i, bottom slope source term and frictional resistance source term;
z b为河床底面高程，单位m； z b is the elevation of the bottom of the river bed, in m;
C f为床面摩擦系数，C f＝gn 2/h 1/3，其中n为曼宁系数。 C f is the bed surface friction coefficient, C f =gn 2 /h 1/3 , where n is the Manning coefficient.
本发明实施例提供的基于GPU加速技术的非结构网格洪水过程模拟系统计算稳定、精度和模拟效率较高，可实现对洪水过程的精确模拟，在非结构网格上应用GPU并行加速技术，提高了模型模拟计算效率，得到较为详细的洪水淹没图，为洪水风险的应急决策提供有力的数据支撑。The unstructured grid flood process simulation system based on the GPU acceleration technology provided by the embodiment of the present invention has stable calculation, high precision and high simulation efficiency, and can realize the accurate simulation of the flood process. The GPU parallel acceleration technology is applied on the unstructured grid, The efficiency of model simulation and calculation is improved, and a more detailed flood inundation map is obtained, which provides strong data support for emergency decision-making of flood risk.

Claims (6)

  1. 一种基于GPU加速技术的非结构网格洪水过程模拟系统，包括：An unstructured grid flood process simulation system based on GPU acceleration technology, including:
    采用非结构三角形网格对模拟区域进行空间离散，得到三角形网格地形文件；Use unstructured triangular mesh to discretize the simulation area in space, and obtain the triangular mesh terrain file;
    读取所述三角形网格地形文件，并将所述三角形网格地形文件中的节点数据和单元数据保存至程序变量中。The triangular mesh terrain file is read, and the node data and unit data in the triangular mesh terrain file are saved into program variables.
    依据两个相邻三角形共用一条边的点编号相同原则对所述三角形网格地形文件建立拓扑关系；According to the principle that the point numbers of two adjacent triangles sharing one side are the same, the topological relationship is established on the triangular mesh terrain file;
    根据所述拓扑关系计算每个三角形网格的相关变量；Calculate the relevant variables of each triangular mesh according to the topological relationship;
    将节点数据中的坐标值转换到单元数据中，并初始化所述相关变量；Convert the coordinate values in the node data into the unit data, and initialize the relevant variables;
    在GPU中为所述相关变量分配内存，并将所述相关变量的值复制到GPU变量中；Allocate memory for the relevant variable in the GPU, and copy the value of the relevant variable into the GPU variable;
    步骤7、采用Godunov格式的有限体积法离散二维浅水方程，然后在GPU上分别计算通量项、底坡源项、摩阻源项和时间步进，得到计算结果；Step 7. Use the finite volume method of Godunov format to discretize the two-dimensional shallow water equation, and then calculate the flux term, bottom slope source term, friction source term and time step respectively on the GPU to obtain the calculation result;
    步骤8、采用cudaMemcpy函数将GPU上得到的计算结果复制到主机内存中，并输出模拟的水深分布图。Step 8. Use the cudaMemcpy function to copy the calculation results obtained on the GPU to the host memory, and output the simulated water depth distribution map.
  2. 根据权利要求1所述的基于GPU加速技术的非结构网格洪水过程模拟系统，其中，所述拓扑关系包括当前边的左侧网格编号，所述当前边的右侧网格编号，以及所述当前边的属性。The unstructured grid flood process simulation system based on GPU acceleration technology according to claim 1, wherein the topological relationship includes the left grid number of the current edge, the right grid number of the current edge, and the Describe the properties of the current edge.
  3. 根据权利要求1所述的基于GPU加速技术的非结构网格洪水过程模拟系统，其中，所述相关变量包括三角形网格的面积、三角形网格各边的长度，三角形网格各边的法线单位向量在x方向及y方向的分量，以及相邻两个三角形网格形心之间的距离。The unstructured grid flood process simulation system based on GPU acceleration technology according to claim 1, wherein the relevant variables include the area of the triangular grid, the length of each side of the triangular grid, the normal of each side of the triangular grid The components of the unit vector in the x and y directions, and the distance between the centroids of two adjacent triangle meshes.
  4. 根据权利要求1所述的基于GPU加速技术的非结构网格洪水过程模拟系统，其中，采用cuda库函数的cudaMalloc进行变量分配，采用cudaMemcpy函数进行变量值的复制。The unstructured grid flood process simulation system based on GPU acceleration technology according to claim 1, wherein the variable allocation is performed using cudaMalloc of the cuda library function, and the variable value is copied using the cudaMemcpy function.
  5. 根据权利要求1所述的基于GPU加速技术的非结构网格洪水过程模拟系统，其中，通量项、底坡源项、摩阻源项和时间步进的计算方法分别对应基于HLLC的黎曼求解器法、底坡通量法、半隐式法和二阶龙格库塔法。The unstructured grid flood process simulation system based on GPU acceleration technology according to claim 1, wherein the calculation methods of flux term, bottom slope source term, friction source term and time step respectively correspond to HLLC-based Riemann Solver method, bottom slope flux method, semi-implicit method, and second-order Runge-Kutta method.
  6. 根据权利要求1所述的基于GPU加速技术的非结构网格洪水过程模拟系统，其中，所述二维浅水方程如下：The unstructured grid flood process simulation system based on GPU acceleration technology according to claim 1, wherein, the two-dimensional shallow water equation is as follows:
    Figure PCTCN2021075073-appb-100001
    Figure PCTCN2021075073-appb-100001
    Figure PCTCN2021075073-appb-100002
    Figure PCTCN2021075073-appb-100002
    式中：where:
    t为时间，单位s；t is time, in s;
    i为降雨及下渗源项；i is the source term of rainfall and infiltration;
    q为变量矢量包括水深h，x、y两个方向上的单宽流量q x和q y； q is a variable vector including water depth h, single-width flow q x and q y in the two directions of x and y;
    u、v为x、y方向上的流速；u and v are the flow velocity in the x and y directions;
    F、G为x、y方向上的通量矢量；F and G are the flux vectors in the x and y directions;
    S为源项矢量，包括降雨及下渗源项i、底坡源项及摩阻力源项；S is the source term vector, including rainfall and infiltration source term i, bottom slope source term and frictional resistance source term;
    z b为河床底面高程，单位m； z b is the elevation of the bottom of the river bed, in m;
    C f为床面摩擦系数，C f＝gn 2/h 1/3，其中n为曼宁系数。 C f is the bed surface friction coefficient, C f =gn 2 /h 1/3 , where n is the Manning coefficient.
PCT/CN2021/075073 2020-07-08 2021-02-03 Unstructured grid flood simulation system based on gpu acceleration thechnology WO2022007398A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
ZA2022/03769A ZA202203769B (en) 2020-07-08 2022-04-01 A flood process simulation system based on unstructured grid and gpu acceleration technique

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010650584.2A CN111768502A (en) 2020-07-08 2020-07-08 Non-structural grid two-dimensional flood simulation system based on GPU acceleration technology
CN202010650584.2 2020-07-08

Publications (1)

Publication Number Publication Date
WO2022007398A1 true WO2022007398A1 (en) 2022-01-13

Family

ID=72725165

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/075073 WO2022007398A1 (en) 2020-07-08 2021-02-03 Unstructured grid flood simulation system based on gpu acceleration thechnology

Country Status (3)

Country Link
CN (1) CN111768502A (en)
WO (1) WO2022007398A1 (en)
ZA (1) ZA202203769B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115564238A (en) * 2022-10-09 2023-01-03 中国水利水电科学研究院 Small reservoir flood forecasting and flood regulating analysis method based on hydrodynamic mechanism
CN116484688A (en) * 2023-04-26 2023-07-25 中国水利水电科学研究院 Urban inland inundation numerical value experiment method
CN117010232A (en) * 2023-06-27 2023-11-07 西安理工大学 Urban non-point source pollution whole process high-resolution simulation method based on GPU (graphic processing Unit) acceleration technology
CN117521404A (en) * 2023-11-30 2024-02-06 中国水利水电科学研究院 Dam break flood dangerous grade classification method for plain reservoir

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111767684B (en) * 2020-06-30 2024-04-19 西安理工大学 Modeling method of optimized friction source item implicit format two-dimensional shallow water equation
CN111768502A (en) * 2020-07-08 2020-10-13 西安理工大学 Non-structural grid two-dimensional flood simulation system based on GPU acceleration technology
CN112233234B (en) * 2020-10-15 2024-04-30 西安理工大学 Reconstruction method of river terrain in flood simulation prediction
CN112257313B (en) * 2020-10-21 2024-05-14 西安理工大学 GPU acceleration-based high-resolution numerical simulation method for pollutant transportation
CN112836872B (en) * 2021-01-29 2023-08-18 西安理工大学 Multi-GPU-based high-performance numerical solution method for pollutant convection diffusion equation
CN113191054A (en) * 2021-04-30 2021-07-30 西安理工大学 High-precision urban rainfall flood simulation method based on display card accelerated coupling pipe network
CN114282403B (en) * 2021-11-22 2024-07-23 西安理工大学 Efficient high-precision habitat simulation method for coupled habitat suitable model
CN115408955B (en) * 2022-08-25 2024-04-30 西安理工大学 Urban rainfall flood tracing numerical simulation method based on rainfall tracing
CN115795939B (en) * 2022-11-03 2023-06-13 天津大学 Two-dimensional flood process multi-GPU simulation method based on unstructured grid

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106599457A (en) * 2016-12-13 2017-04-26 中国水利水电科学研究院 Godunov format one-dimensional and two-dimensional coupling technology-based mountain flood numerical simulation method
CN107133427A (en) * 2017-06-07 2017-09-05 中国水利水电科学研究院 A kind of construction method of the flood risk analysis model based on 2DGIS platforms
CN111768502A (en) * 2020-07-08 2020-10-13 西安理工大学 Non-structural grid two-dimensional flood simulation system based on GPU acceleration technology

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107451372B (en) * 2017-08-09 2018-04-13 中国水利水电科学研究院 A kind of flood of a mountain area numerical simulation method that kinematic wave is combined with dynamic wave
CN109190261A (en) * 2018-09-07 2019-01-11 中国水利水电科学研究院 A kind of flood risk analysis method that one-dimensional river network generalization is coupled with one two dimension of high-performance
CN110929443B (en) * 2019-12-02 2021-11-30 中国水利水电科学研究院 Two-dimensional flood simulation method based on high-precision terrain generalization

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106599457A (en) * 2016-12-13 2017-04-26 中国水利水电科学研究院 Godunov format one-dimensional and two-dimensional coupling technology-based mountain flood numerical simulation method
CN107133427A (en) * 2017-06-07 2017-09-05 中国水利水电科学研究院 A kind of construction method of the flood risk analysis model based on 2DGIS platforms
CN111768502A (en) * 2020-07-08 2020-10-13 西安理工大学 Non-structural grid two-dimensional flood simulation system based on GPU acceleration technology

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HOU JINGMING, RUN WANG, GUODONG LI, GUIYI LI: "High-performance numerical model for high-resolution urban rainfall-runoff process based on dynamic wave method", JOURNAL OF HYDROELECTRIC ENGINEERING, vol. 37, no. 3, 31 March 2018 (2018-03-31), pages 40 - 49, XP055886181, ISSN: 1003-1243, DOI: 10.11660/slfdxb.20180305 *
MA LIPING, HOU JINGMING;ZHANG DAWEI;XIA JUNQIANG;LI BINGYAO;NING LIZHONG: "Study on 2-D Numerical Simulation Coupling With Breach Evolution in Flood Propagation", JOURNAL OF HYDRAULIC ENGINEERING, vol. 50, no. 10, 31 October 2019 (2019-10-31), pages 1253 - 1267, XP055886174, ISSN: 0559-9350, DOI: 10.13243/j.cnki.slxb.20190462 *
QI WENCHAO, WANG RUN, LIU LI, LI PENG, JING HAIXIAO, WANG WEN, HOU JINGMING: "Modelling of Dam-Breach Flooding Evolution Based on GPU-Accelerated Computing", PROCEEDINGS OF THE 2016 ANNUAL MEETING OF THE CHINESE NATIONAL COMMITTEE ON LARGE DAMS, 4 April 2019 (2019-04-04), XP055886178 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115564238A (en) * 2022-10-09 2023-01-03 中国水利水电科学研究院 Small reservoir flood forecasting and flood regulating analysis method based on hydrodynamic mechanism
CN116484688A (en) * 2023-04-26 2023-07-25 中国水利水电科学研究院 Urban inland inundation numerical value experiment method
CN116484688B (en) * 2023-04-26 2023-10-13 中国水利水电科学研究院 Urban inland inundation numerical value experiment method
US12056427B1 (en) 2023-04-26 2024-08-06 China Institute Of Water Resources And Hydropower Research Numerical experimental method for urban waterlogging
CN117010232A (en) * 2023-06-27 2023-11-07 西安理工大学 Urban non-point source pollution whole process high-resolution simulation method based on GPU (graphic processing Unit) acceleration technology
CN117521404A (en) * 2023-11-30 2024-02-06 中国水利水电科学研究院 Dam break flood dangerous grade classification method for plain reservoir
CN117521404B (en) * 2023-11-30 2024-05-07 中国水利水电科学研究院 Dam break flood dangerous grade classification method for plain reservoir

Also Published As

Publication number Publication date
CN111768502A (en) 2020-10-13
ZA202203769B (en) 2022-07-27

Similar Documents

Publication Publication Date Title
WO2022007398A1 (en) Unstructured grid flood simulation system based on gpu acceleration thechnology
US9465581B2 (en) Automated method and system for predicting high resolution tidal heights and currents in coastal zones
CN108629135A (en) Non- unified high-precision curved grid flow simulation of water quality and method for visualizing and system
CN102708227A (en) SPH (smoothed particle hydrodynamics) algorithm-based simulation method and simulation system of process of breaking dam by flood
Chen et al. Local curvature entropy-based 3D terrain representation using a comprehensive Quadtree
CN114547498A (en) Water flow two-dimensional flow field simulation visualization method and system based on B/S framework
Winkler et al. Virtual reality in urban water management: communicating urban flooding with particle-based CFD simulations
Ginting et al. Artificial viscosity technique: A Riemann-solver-free method for 2D urban flood modelling on complex topography
JP2006091811A (en) Three-dimensional moving image generating method for liquid and program thereof
KR101909383B1 (en) Modelling method of solid wall boundary condition in compressible boltzmann method
Roy et al. A MATLAB-Based Grid Generation Tool for Hydrodynamic Modeling
CN108256266A (en) A kind of one-dimension mathematical model and two-dimentional hydrodynamic model coupling process and system
Thakur et al. Exploring CCHE2D and its sediment modelling capabilities
Fries Higher-order accurate integration for cut elements with Chen-Babuška nodes
Pons et al. Adaptive mesh refinement method applied to shallow water model: a mass conservative projection
Wong et al. Integrating computational fluid dynamics (CFD) models with GIS: An evaluation on data conversion formats
Guo et al. Research on adaptive tide numerical simulation based on steering dynamic monitoring
Zhang et al. Conformal adaptive hexahedral-dominant mesh generation for CFD simulation in architectural design applications
CN117893712B (en) Surface structure grid generation method and device
Soomro et al. Discretization of laplacian operator on 19 points stencil using cylindrical mesh system with the help of explicit finite difference scheme
Li et al. Application of three-dimensional GIS to water resources
CN117576339B (en) Unstructured grid generation method with highest resolution of 1km
CN109242937B (en) Method and device for converting Mike flow field result file into shape vector file
Lu et al. Visualization and High-Performance Computing for City-Scale Nonlinear Time-History Analyses
Yang et al. 3D modeling of riverbeds based on NURBS algorithm

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21837718

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21837718

Country of ref document: EP

Kind code of ref document: A1