Digital elevation model (DEM)-based rainfall-induced landslide numerical simulation method

Abstract

The invention discloses a rainfall-induced landslide numerical simulation method based on a DEM, and the method comprises the steps: 1) constructing a fluid grid, and giving a grid initial boundary condition; 2) adopting a discrete element method to establish particles and walls in PFC3D, and establishing a DEM module; 3) establishing a CFD module by using a Python language, and calculating through the CFD module to obtain a pressure water head and a pressure gradient force; 4) establishing a coupling module to exchange data of the CFD module and the DEM module, and updating soil contact information, soil moisture content information and particle position information according to a pressure water head; 5) setting a specified iteration step number, and obtaining a simulation result of particle displacement under the rainfall effect.The method belongs to the technical field of geological disaster forecasting, introduces the V-G water holding model to represent the water-soil characteristic relation of unsaturated soil, can fully consider the decrease of the contact force between soil particles under the rainfall condition, and enables the simulation to be more accurate.

Description

Technical Field

[0001] The invention belongs to the technical field of geological disaster prediction and relates to a numerical simulation method of rainfall-induced landslide based on DEM. Background Art

[0002] Affected by global climate change and human engineering activities, various geological disasters such as earthquakes, landslides, and storms are becoming more frequent, seriously restricting people's production and life. Rainfall has been considered one of the main factors causing landslide disasters, but the mechanism of rainfall-induced landslides needs to be further understood.

[0003] Currently, the susceptibility of landslides can be predicted and evaluated through machine learning methods, but these methods cannot analyze the landslide-inducing mechanism at the particle scale.

[0004] Therefore, it is urgent to develop a new simulation method to numerically simulate and analyze rainfall-induced landslides at the particle scale in order to provide a basis for decision-making in a timely and accurate manner. Summary of the invention

[0005] The purpose of the present invention is to provide a numerical simulation method of rainfall-induced landslide based on DEM, which solves the problem that the existing simulation technology does not take into account the water-soil coupling characteristics sufficiently, resulting in large analysis errors.

[0006] The technical solution adopted by the present invention is a numerical simulation method of rainfall-induced landslide based on DEM, which is implemented according to the following steps: Step 1: Construct a fluid grid and set the initial boundary conditions of the grid; Step 2: Use the discrete element method in PFC3D to establish particles and walls and build a DEM module; Step 3: Use Python language to establish a CFD module, use the VG model to simulate the water-soil coupling behavior, and calculate the pressure head and pressure gradient force through the CFD module; Step 4: Establish a coupling module to exchange data between the CFD module and the DEM module. Specifically, use the program call command in the DEM module to call the CFD module, and update the soil contact information, soil moisture content information, and particle position information according to the pressure head. Step 5: Set the specified number of iteration steps to obtain the simulation results of particle displacement under rainfall.

[0007] The beneficial effects of the present invention are that the method is based on the coupling of computational fluid dynamics and discrete element method to construct a simulation method for rainfall-induced landslides. The data calculated by the CFD module and the DEM module are exchanged and transmitted with each other through the coupling module, and iteratively updated to calculate the interaction between particles and fluid-solid two phases, characterize the fluid-solid coupling behavior characteristics of rainfall-induced landslides, and further guide the assessment and prevention of rainfall-induced landslide disasters, which specifically includes the following aspects: (1) The method of the present invention can better consider the changes in soil parameters under the action of rainfall and effectively simulate the scene of soil instability and landslide under the action of rainfall.

[0008] (2) The method of the present invention regards solid particles as discrete phases, solves the particle motion control equation based on Newton's second law to simulate the movement of particles, regards liquid fluid as continuous phase, and solves and simulates the flow of pore fluid based on Richards equation. The Richards equation is linked to the VG water retention model through the pressure head h, which can be applied to soils of different properties. It is very convenient to use Gmsh to divide the grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a flow chart of the numerical simulation method of rainfall-induced landslide based on DEM of the present invention; Figure 2 is a schematic diagram of a fluid grid in Example 1 of the present invention; Figure 3 This is a pressure head distribution cloud diagram after three hours of rainfall in Example 1 of the present invention; Figure 4 It is a pressure head process line diagram in Example 1 of the present invention. DETAILED DESCRIPTION

[0010] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0011] Reference Figure 1 The DEM-based rainfall-induced landslide numerical simulation method of the present invention is implemented according to the following steps: Step 1: Construct the fluid grid and give the initial boundary conditions of the grid. The specific process is: use Gmsh to draw the tetrahedral grid of the landslide and save the grid as a .msh file; then import the .msh file into PFC3D to read and establish the fluid grid, give the pressure on the upper and lower surfaces of the fluid grid, and the side is an impermeable boundary.

[0012] Step 2: In PFC3D, the discrete element method is used to establish particles and walls. The control equations of moving solid particles, the normal contact force model between particles, and the tangential contact force model between particles are given to establish a DEM module. The specific process is: construct particles within the fluid grid range and give material parameters such as particle radius, density, stiffness emod, and friction coefficient fric.

[0013] The solid particles use the built-in contact bonding model of PFC3D, and the solid particles control equation is as follows: , , in, is the particle velocity vector; t It’s time; is the sum of the additional forces acting on the particle (externally applied forces and contact forces); is the pressure gradient force exerted by the fluid on the particle; is the particle mass; g is the acceleration due to gravity; is the particle angular velocity; is the particle moment of inertia; is the torque acting on the particle.

[0014] Step 3: Use Python language to establish a CFD module, use the VG model to simulate the water-soil coupling behavior, and calculate the pressure head and pressure gradient force through the CFD module. The specific process is to use fipy to solve the three-dimensional Richards equation to obtain the pressure head.

[0015] The relationship between the VG water-soil characteristic curve is as follows: , , , , , The three-dimensional Richards equation is given by the following equation: , in, Indicates volumetric water content in cm 3 cm -3 ; Indicates saturated moisture content, in cm 3 cm -3 ; is the residual moisture content, in cm 3 cm -3 ; , a, n, and m are model-related parameters; Indicates saturation; It indicates the pressure head, the unit is cmH2O, and it takes positive value; , They represent total head and position head respectively; represents the fluid force, r represents the particle radius, Represents the change in pressure gradient force; represents the hydraulic conductivity along the z direction, , Same reason.

[0016] The preferred parameters of the rainfall-induced landslide model are shown in Table 1.

[0017] Step 4: Establish a coupling module to exchange data between the CFD module and the DEM module. Specifically, use the program call command in the DEM module to call the CFD module, and update the soil contact information, soil moisture information, and particle position information according to the pressure head (the soil contact information is mainly the soil contact strength, and the soil moisture information is mainly the saturation). Soil contact strength includes normal contact strength and tangential contact strength , the expressions are as follows: , .

[0018] The relationship between soil contact strength and saturation is defined as follows: When the saturation is 0, the normal contact strength The value of 0; when the saturation is 1, the normal contact strength The value of 1; Similarly, when the saturation is 0, the tangential contact strength The value of 0; when the saturation is 1, the tangential contact strength The value of 1.

[0019] Step 5: Iterate the specified number of steps to 400,000 to obtain the simulation results of particle displacement under rainfall.

[0020] The design principle of the method of the present invention is: the landslide disaster induced by rainfall is simulated by combining discrete element and computational fluid dynamics, the Van-Genuchten (VG) water retention model is introduced to characterize the water-soil characteristic relationship of unsaturated soil, and the open source Fipy solver is used to calculate the three-dimensional Richards equation. The Richards equation is linked to the VG model, and the contact and displacement information of particles is obtained through continuous iterative updates to simulate the process of landslide occurrence under rainfall.

[0021] Example 1 The DEM-based rainfall-induced landslide numerical simulation method of Example 1 is implemented according to the following steps: Step 1: Construct the fluid grid and give the initial boundary conditions of the grid. The specific process is: use Gmsh to draw the tetrahedral grid of the landslide and save the grid as a .msh file; then import the .msh file into PFC3D to read and establish the fluid grid, give the pressure on the upper and lower surfaces of the fluid grid, and the side is an impermeable boundary.

[0022] Step 2: In PFC3D, the discrete element method is used to establish particles and walls. The control equations of moving solid particles, the normal contact force model between particles, and the tangential contact force model between particles are given, so as to establish a DEM module. The specific process is: construct particles within the fluid grid range, and give material parameters such as particle radius, density, stiffness emod, and friction coefficient fric.

[0023] The solid particles use the built-in contact bonding model of PFC3D, and the solid particles control equation is as follows: , , in, is the particle velocity vector; t It’s time; is the sum of the additional forces acting on the particle (externally applied forces and contact forces); is the pressure gradient force exerted by the fluid on the particle; is the particle mass; g is the acceleration due to gravity; is the particle angular velocity; is the particle moment of inertia; is the torque acting on the particle.

[0024] Step 3: Use Python language to build a CFD module, use the VG model to simulate the water-soil coupling behavior, and calculate the pressure head and pressure gradient force through the CFD module. The specific process is to use fipy to solve the three-dimensional Richards equation to obtain the pressure head.

[0025] The relationship between the VG water-soil characteristic curve is as follows: , , , , , The three-dimensional Richards equation is given by the following equation: , in, Indicates volumetric water content in cm 3 cm -3 ; Indicates saturated moisture content, in cm 3 cm -3 ; is the residual moisture content, in cm 3 cm -3 ; , a, n, and m are model-related parameters; Indicates saturation; It indicates the pressure head, the unit is cmH2O, and it takes positive value; , They represent total head and position head respectively; represents the fluid force, r represents the particle radius, Represents the change in pressure gradient force; represents the hydraulic conductivity along the z direction, , Same reason.

[0026] The preferred parameters of the rainfall-induced landslide model are shown in Table 1.

[0027] Table 1. Parameters of rainfall-induced landslide model

[0028] Step 4: Establish a coupling module to exchange data between the CFD module and the DEM module. Specifically, use the program call command in the DEM module to call the CFD module, and update the soil contact information, soil moisture information and particle position information according to the pressure head; the soil contact strength includes the normal contact strength and tangential contact strength , the expressions are as follows:

[0029]

[0030] The relationship between soil contact strength and saturation is defined as follows: When the saturation is 0, the normal contact strength The value of 0; when the saturation is 1, the normal contact strength The value of 1; Similarly, when the saturation is 0, the tangential contact strength The value of 0; when the saturation is 1, the tangential contact strength The value of 1.

[0031] Step 5: Iterate the specified number of steps to 400,000 to obtain the simulation results of particle displacement under rainfall.

[0032] The specific parameter settings and simulation results are: Reference Figure 2 , use Gmsh to draw the tetrahedral mesh of the landslide, and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish a fluid mesh, giving the pressures of the upper and lower surfaces of the fluid mesh as 1 and -1 MPa respectively, and the side as an impermeable boundary. In PFC3D, the discrete element method is used to establish a particle radius of 0.3, a density of 1100, a friction coefficient of 0.3, a wall stiffness of 1e9, and a solid-phase particle motion control equation given by Newton's laws of motion. The contact model between particles is a contact bonding model, thereby establishing a DEM module; use python for simulation, the variable is the pressure head, and the time step dt is set to 0.02, and the pressure head and water content of each grid are solved. Reference Figure 3 and Figure 4 , the pressure head rises from -1Mpa, the soil moisture content rises from the initial 10% until saturation, and the shear strength of the soil decreases from 1000 to 200. After 3 hours of rainfall, the maximum particle displacement of 2.3m can be observed.

[0033] Example 2 The main steps of the numerical simulation method of rainfall-induced landslide based on DEM in Example 2 are as follows: Step 1: Construct a fluid grid and set the initial boundary conditions of the grid; Step 2: Use the discrete element method in PFC3D to establish particles and walls and build a DEM module; Step 3: Use Python language to establish a CFD module, use the VG model to simulate the water-soil coupling behavior, and calculate the pressure head and pressure gradient force through the CFD module; Step 4: Establish a coupling module to exchange data between the CFD module and the DEM module. Specifically, use the program call command in the DEM module to call the CFD module, and update the soil contact information, soil moisture content information, and particle position information according to the pressure head. Step 5: Set the specified number of iteration steps to obtain the simulation results of particle displacement under rainfall.

[0034] The specific settings and simulation results are as follows: Use Gmsh to draw the tetrahedral mesh of the landslide and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish the fluid mesh. The pressures of the upper and lower surfaces of the given fluid mesh are 2 and -2MPa respectively, and the side is an impermeable boundary. In PFC3D, the discrete element method is used to establish a particle radius of 0.3, a density of 1100, a friction coefficient of 0.3, a wall stiffness of 1e9, and a solid phase particle motion control equation given by Newton's law of motion. The contact model between particles is a contact bonding model, thereby establishing a DEM module; use python for simulation, the variable is the pressure head, and the time step dt is set to 0.02 to solve the pressure head and moisture content of each grid. The pressure head rises from -2Mpa, the soil moisture content rises from the initial 10% until saturation, and the shear strength of the soil decreases from 1000 to 200. After 3 hours of rainfall, the maximum displacement of the particles can be observed to be 2.3m.

[0035] Example 3 This embodiment 3 refers to the steps and related basic values ​​of embodiment 1, and the specific settings and simulation results are as follows: Use Gmsh to draw the tetrahedral mesh of the landslide and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish the fluid mesh, and give the pressure of the upper and lower surfaces of the fluid mesh to be 1 and -1MPa respectively, and the side is an impermeable boundary. In PFC3D, the discrete element method is used to establish a particle radius of 0.4, a density of 1100, a friction coefficient of 0.3, a wall stiffness of 1e9, and a solid phase particle motion control equation given by Newton's law of motion. The contact model between particles is a contact bonding model, thereby establishing a DEM module; use python for simulation, the variable is the pressure head, and the time step dt is set to 0.02, and solve the pressure head and moisture content of each grid. The pressure head rises from -1Mpa, the soil moisture content rises from the initial 10% until saturation, and the shear strength of the soil decreases from 1000 to 200. After 3 hours of rainfall, the maximum displacement of the particles can be observed to be 2.1m.

[0036] Example 4 This embodiment 4 refers to the steps and related basic values ​​of embodiment 1, and the specific settings and simulation results are as follows: Use Gmsh to draw the tetrahedral mesh of the landslide and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish the fluid mesh, and give the pressure of the upper and lower surfaces of the fluid mesh to be 1 and -1MPa respectively, and the side is an impermeable boundary. In PFC3D, the discrete element method is used to establish a particle radius of 0.3, a density of 1100, a friction coefficient of 0.3, a wall stiffness of 1e10, and a solid phase particle motion control equation given by Newton's law of motion. The contact model between particles is a contact bonding model, thereby establishing a DEM module; use python for simulation, the variable is the pressure head, and the time step dt is set to 0.02, and solve the pressure head and moisture content of each grid. The pressure head rises from -1Mpa, the soil moisture content rises from the initial 10% until saturation, and the shear strength of the soil decreases from 1000 to 200. After 3 hours of rainfall, the maximum displacement of the particles can be observed to be 2.3m.

[0037] Example 5 This embodiment 5 refers to the steps and related basic values ​​of embodiment 1, and the specific settings and simulation results are as follows: Use Gmsh to draw the tetrahedral mesh of the landslide and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish the fluid mesh, and give the pressure of the upper and lower surfaces of the fluid mesh to be 1 and -1MPa respectively, and the side is an impermeable boundary. In PFC3D, the discrete element method is used to establish a particle radius of 0.3, a density of 1100, a friction coefficient of 0.4, a wall stiffness of 1e9, and a solid phase particle motion control equation given by Newton's law of motion. The contact model between particles is a contact bonding model, thereby establishing a DEM module; use python for simulation, the variable is the pressure head, and the time step dt is set to 0.02, and solve the pressure head and moisture content of each grid. The pressure head rises from -1Mpa, the soil moisture content rises from the initial 10% until saturation, and the shear strength of the soil decreases from 1000 to 200. After 3 hours of rainfall, the maximum displacement of the particles can be observed to be 1.6m.

[0038] Example 6 This embodiment 6 refers to the steps and related basic values ​​of embodiment 1, and the specific settings and simulation results are as follows: Use Gmsh to draw the tetrahedral mesh of the landslide and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish the fluid mesh, and give the pressure of the upper and lower surfaces of the fluid mesh to be 1 and -1MPa respectively, and the side is an impermeable boundary. In PFC3D, the discrete element method is used to establish a particle radius of 0.3, a density of 1200, a friction coefficient of 0.3, a wall stiffness of 1e9, and a solid phase particle motion control equation given by Newton's law of motion. The contact model between particles is a contact bonding model, thereby establishing a DEM module; use python for simulation, the variable is the pressure head, and the time step dt is set to 0.02, and solve the pressure head and moisture content of each grid. The pressure head rises from -1Mpa, the soil moisture content rises from the initial 10% until saturation, and the shear strength of the soil decreases from 1000 to 200. After 3 hours of rainfall, the maximum displacement of the particles can be observed to be 2.3m.

[0039] Conclusion: The larger the friction coefficient, the smaller the particle displacement. As the rainfall duration increases, the moisture content of the particles increases, the shear strength of the particles gradually decreases, and the soil saturation increases until saturation. The stiffness of the wall does not affect the displacement of the particles.

Claims

1. The numerical simulation method of rainfall-induced landslide based on DEM is characterized by: Follow these steps to implement: Step 1: Construct a fluid grid and set the initial boundary conditions of the grid; Step 2: Use the discrete element method in PFC3D to establish particles and walls and build a DEM module; Step 3: Use Python language to establish a CFD module, use the VG model to simulate the water-soil coupling behavior, and calculate the pressure head and pressure gradient force through the CFD module; Step 4: Establish a coupling module to exchange data between the CFD module and the DEM module. Specifically, use the program call command in the DEM module to call the CFD module, and update the soil contact information, soil moisture content information, and particle position information according to the pressure head. Step 5: Set the specified number of iteration steps to obtain the simulation results of particle displacement under rainfall.

2. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1 is characterized in that: In step 1, the specific process is: Use Gmsh to draw the tetrahedral mesh of the landslide and save the mesh as a .msh file; then import the .msh file into PFC3D to read and establish the fluid mesh, give the pressure on the upper and lower surfaces of the fluid mesh, and use the side as an impermeable boundary.

3. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1 is characterized in that: In step 2, the specific process is: In PFC3D, the discrete element method is used to establish the relationship between particles and walls. The control equations of moving solid particles, the normal contact force model between particles, and the tangential contact force model between particles are given to establish the DEM module. Construct particles within the fluid grid, and give the particle radius, density, stiffness emod, and friction coefficient fric; The solid particles use the built-in contact bonding model of PFC3D, and the solid particles control equation is as follows: ， ， in, is the particle velocity vector; t It’s time; is the sum of the additional forces acting on the particle; is the pressure gradient force exerted by the fluid on the particle; is the particle mass; g is the acceleration due to gravity; is the particle angular velocity; is the particle moment of inertia; is the torque acting on the particle.

4. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1 is characterized in that: In step 3, the specific process is: use fipy to solve the three-dimensional Richards equation to obtain the pressure head, and the VG water-soil characteristic curve relationship is as follows: ， ， ， ， ， The three-dimensional Richards equation is given by the following equation: ， in, Indicates volumetric water content in cm 3 cm -3 ; Indicates saturated moisture content, in cm 3 cm -3 ; is the residual moisture content, in cm 3 cm -3 ; , a, n, and m are model-related parameters; Indicates saturation; It indicates the pressure head, the unit is cmH2O, and it takes positive value; , They represent total head and position head respectively; represents the fluid force, r represents the particle radius, Represents the change in pressure gradient force; represents the hydraulic conductivity along the z direction, , Same reason.

5. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1 is characterized in that: In step 4, the specific process is: soil contact strength includes normal contact strength and tangential contact strength , the expressions are as follows: ， 。 6. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1, characterized in that: In step 4, the relationship between soil contact strength and saturation is defined as follows: When the saturation is 0, the normal contact strength The value of 0; when the saturation is 1, the normal contact strength The value of 1.

7. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1, characterized in that: In step 4, the relationship between soil contact strength and saturation is defined as follows: When the saturation is 0, the tangential contact strength The value of 0; when the saturation is 1, the tangential contact strength The value of 1.

8. The DEM-based rainfall-induced landslide numerical simulation method according to claim 1, characterized in that: In step 5, the number of iteration steps is set to 400,000.

Images