Landslide digital twin framework construction method

Abstract

The invention relates to the technical field of landslide simulation, and discloses a landslide digital twinborn framework construction method, which comprises the following steps: carrying out analog simulation on slope instability conditions of different soil properties and different gradients, forming a digital twinborn body construction foundation, constructing a three-dimensional scene substrate, and fusing an instability threshold value and a rainfall field to form a landslide digital twinborn body. And constructing a three-dimensional visual scene of the landslide, and carrying out visual monitoring and early warning on the landslide process. The framework technology constructed by the method supports dynamic updating and response of measured data, and can provide theoretical technical support for accurate prediction and intelligent early warning of the landslide process. The technology has a function of updating physical attributes of the earth surface in real time, can realize visualization of the whole process of the landslide, provides a theoretical basis for landslide prevention and control decision, realizes real-time monitoring of areas where the landslide easily occurs, saves a lot of manpower and material resources compared with traditional monitoring, and has important significance for intelligent monitoring and management of the landslide.

Description

Technical Field

[0001] This invention relates to the field of landslide simulation technology, specifically a method for constructing a digital twin framework for landslides. Background Technology

[0002] Digital twin technology originated in the industrial sector, initially used to digitally map physical products. With technological advancements, its applications have expanded to broader fields such as digital twin cities, digital twin river basins, and digital twin Earths. This technology constructs virtual mirror images of physical entities, enabling real-time data analysis throughout the entity's entire lifecycle and providing support for decision-making. Implementation requires precise digital replication of the physical entity's characteristics, operational status, and dynamic behavior, constructing a highly realistic "mirror image" of the entity in virtual space. This allows researchers to conduct performance analysis, prediction, and optimization based on the virtual model.

[0003] In high-altitude mountainous areas, landslides are a prominent type of geological hazard, seriously threatening human life and property. The complex terrain and fragile soil and rock structures of these areas, coupled with the influence of heavy rainfall and other climatic conditions, make them high-risk landslide zones. While landslides can be categorized into various types based on their causative factors, over 80% are triggered by rainfall, or rainfall is the dominant factor. Traditional prevention methods, such as soil anchoring, slope reinforcement, and drainage systems, are limited in their effectiveness due to insufficient understanding of landslide formation mechanisms. In recent years, with the rapid innovation of digital technology and the continuous development of earth sciences, digital geology has become a cutting-edge direction in landslide research. How to establish a digital scenario of the entire landslide process, integrate multi-source data, and update it in real time has become an urgent problem to be solved.

[0004] Therefore, integrating knowledge from multiple disciplines such as geology, rock and soil mechanics, and meteorology, and leveraging modern information technology to construct a digital twin framework for landslide disasters based on multi-source data fusion, is of great significance for in-depth research on landslide formation mechanisms, development of accurate and efficient prediction models, improvement of prevention and control levels, and ensuring safety. Constructing digitally driven landslide models is of great value in revealing the mechanisms of regional landslide gestation and evolution, and can also improve the accuracy and effectiveness of landslide disaster prevention and control. Against this backdrop, this invention proposes a landslide digital twin framework applicable to complex geographical conditions and explores its practical application paths and value in landslide prevention and control, in order to promote the innovative development of landslide research and prevention technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for constructing a digital twin framework for landslides, which has advantages such as improving the accuracy and effectiveness of landslide disaster prevention and control, and solves the aforementioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for constructing a digital twin framework for landslides, comprising the following steps: S1: Analyze the periods when landslides are most likely to occur and construct the causes of landslides; S2: Based on the landslide causes obtained in S1, the instability of slopes with different soil properties and different slope gradients is simulated. S3: Based on simulation, key parameters of slope stability before and after evolution are obtained, including stress field, void ratio, and displacement field. The changes in stress and land cover properties before and after landslide are analyzed. S4: Obtain the land feature attributes and stress distribution simulated in S3 to form the basis for building a digital twin, and store the stress values ​​corresponding to the slope height; S5: Based on the digital twin obtained in S4, construct a three-dimensional scene base, integrate instability threshold and rainfall field to form a landslide digital twin, construct a three-dimensional landslide scene visualization scene, and perform visual monitoring and early warning of the landslide process.

[0007] As a preferred technical solution of the present invention, the simulation analysis in S2 is based on Abaqus software to establish simulation models with different slopes. The specific process is as follows: starting from a slope of 40°, simulation is performed once for every 5° increase, up to 80°. The overall length of the simulation model is 50 meters, the height is 30 meters, and the foundation soil layer is 10 meters. The simulation model parameters include dry density, deformation modulus, Poisson's ratio, saturated permeability coefficient, initial void ratio, effective cohesion, effective internal friction angle, and rainfall level.

[0008] As a preferred embodiment of the present invention, the key parameters for simulation in S3 also include Mises stress.

[0009] As a preferred technical solution of the present invention, the specific steps for constructing a three-dimensional visualization scene of a landslide in S5 are as follows: construct a three-dimensional terrain based on ue5 software, extract several slope points on the basis of the three-dimensional terrain, and overlay a solid model based on the slope points. Each solid model is used as a basic unit, and the stress value corresponding to the slope height stored in S4 is input into each basic unit, and real-time simulation is performed.

[0010] Compared with the prior art, the present invention provides a method for constructing a digital twin framework for landslides, which has the following beneficial effects: This invention enables the construction of landslide scenarios based on physical characteristics, which is of great value in revealing the mechanisms of regional landslide formation and evolution. It can also improve the accuracy and effectiveness of landslide disaster prevention and control, and has the function of real-time updating of surface attributes. It can realize the visualization of the entire landslide process, provide a theoretical basis for landslide prevention and control decisions, and realize real-time monitoring of landslide-prone areas. Compared with traditional monitoring, it saves a lot of manpower and material resources, and is of great significance for the intelligent monitoring and management of landslides. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating an embodiment of a landslide digital twin framework construction method according to the present invention. Figure 2 To obtain simulation models of the physical properties of land features at different slopes; Figure 3 Line graphs showing stress at different slope heights before and after the landslide; Figure 4 The process and simulation results for creating a digital twin scenario; Figure 5 Construct a schematic diagram for the 3D scene; Figure 6 This is a schematic diagram of a real-world application scenario. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] Please see Figures 1-6 A method for constructing a digital twin framework for landslides, comprising: S1. Analyze the time period during which landslides are most likely to occur, identify the locations prone to landslides, and analyze the causes of landslides.

[0014] Specifically, data on landslides in Zhaotong, Yunnan Province, from 2010 to 2024 were compiled, totaling over 500 incidents. Analysis of the data extracted the timing, location, and dominant factors of each landslide. The study analyzed the most frequent time of year for landslides, finding that July saw the highest number, with over 400 incidents occurring between May and July, accounting for more than 85% of the total. Since May to July coincides with the rainy season, rainfall was the dominant factor in over 80% of the landslide data. Furthermore, the majority of landslides occurred on slopes exceeding 40 degrees. This work corresponds to... Figure 1 The first part of the flowchart.

[0015] S2. Based on the landslide causes obtained in S1, design the model architecture, complete the decision level of the model, and simulate and analyze the soil properties and slope instability at different slopes. Specifically, since most landslides are caused by rainfall, the initial rainfall gradually increases the water content of the soil and rock, setting the stage for the impact of subsequent short-duration torrential rains. These short-duration torrential rains, as a key triggering factor, bring in a large amount of rainwater in a short period, drastically altering the stability conditions of the slope. Rainfall affects slope stability through multiple physical processes. Rainwater infiltration is a crucial link; after rainwater seeps into the slope's soil and rock mass, it causes a significant increase in the weight of the soil and rock. This is because water is denser than air; rainwater fills the pores in the soil and rock mass, increasing the weight per unit volume of soil and rock, thus increasing the slope's sliding force. Therefore, rainfall-induced landslides typically occur on slopes that already have cracks. Based on the analysis of landslide formation, before constructing a twin scenario, it is necessary to examine whether cracks exist at the slope crest to provide decision-making assistance for the twin scenario.

[0016] In the simulation, rainfall field is used as the dominant factor. Slope simulation models with different slopes are established, and real soil physical parameters are input. A gravity field is also added to simulate the situation before and after a landslide on a real surface. To obtain the stress distribution of the soil before and after slope instability, simulations are performed in Abaqus. First, simulation models with different slopes are established. According to S1 statistics, landslides with slopes above 40 degrees are mainly affected by gravity at the top of the slope, so the correlation with the slope is minimal. However, apart from the top, the stress situation of all parts of the slope is greatly affected by the slope. Therefore, this paper simulates the stress situation at different slopes. The shape of the model is as follows. Figure 2 As shown, the model is 30 meters high, with a 10-meter-thick base soil layer and a length of 40 meters. Simulations were performed every 5 degrees increasing from a 40° slope, up to 80°. The overall model length is 50 meters, the height is 30 meters, and the base soil layer is 10 meters. Soil physical properties were incorporated, and the simulated soil parameters are shown in Table 1. An external field model of rainfall infiltration was added, with the rainfall magnitude set to an extreme rainstorm, i.e., 24-hour rainfall exceeding 250 mm. This work corresponds to... Figure 1 The second part of the flowchart.

[0017] Table 1 Simulation parameter settings S3. Based on simulation, obtain key parameters such as stress field, void ratio, and displacement field before and after slope stability evolution, and compare various parameters of different slopes simulated. Analyze the changes in stress before and after the landslide.

[0018] Specifically, the principal stresses in the simulation results are crucial mechanical parameters in fields such as materials mechanics and geotechnical mechanics. Mises stress can be expressed by the following two formulas: In the formula: For Mises stress, The first principal stress, This is the second principal stress. The third principal stress, For stress acting perpendicular to the x-axis, For the stress acting perpendicular to the y-axis, This represents the stress acting perpendicular to the z-axis. It is the shear stress acting in the y-direction on a plane perpendicular to the x-axis. It is the shear stress acting in the z-direction on a plane perpendicular to the y-axis. Mises stress is the shear stress acting on a plane perpendicular to the z-axis along the x-direction. It can be calculated using the three principal stresses. In the Cartesian coordinate system, Mises stress can be described by stress components in various directions, as shown in the second formula.

[0019] Comparison of simulation results with different slope gradients revealed that as the slope gradient increases, both the shear stress and normal stress components at specific locations on the slope surface exhibit an increasing trend. Figure 3 As shown, this phenomenon is particularly pronounced in the lower part of the slope, i.e., at lower elevations, mainly due to the greater self-weight stress and lateral pressure experienced by the lower slope. As the elevation approaches the slope crest, regardless of the slope ratio, the stress gradient gradually decreases, exhibiting a stress stabilization trend. Furthermore, quantitative analysis of the simulation data reveals that the stress amplitude on the slope surface decreases with increasing elevation. This simulation result highly agrees with the stress distribution law derived from soil mechanics principles, further validating the effectiveness and reliability of the constructed numerical model in simulating the evolution of slope stress fields.

[0020] Simulations reveal displacement at various points, void ratios in different sections, and various stress results. Detailed analysis of these simulations clearly shows that the Mises stress exhibits a spatially regular variation, increasing gradually in a trapezoidal pattern from the slope surface to the interior of the slope. Focusing on the stress characteristics of the slope surface, the simulation results show that the stress at the slope crest is largely unaffected by the increase in slope gradient. In stark contrast, the increase in slope gradient has a crucial impact on the stress at the slope base. The stress at different slope heights (2m-20m) before and after instability shows a significant increasing trend in stress at the slope base as the slope gradient increases. This is because the change in slope gradient alters the stress boundary conditions and stress transmission path of the slope. This analysis aligns with the concept of slope in geotechnical mechanics, indirectly verifying the accuracy of the simulation results. The simulated stress values ​​at different slope gradients before and after the landslide are extracted and embedded into a digital twin scenario of the landslide.

[0021] S4: Summarize the land feature attributes and stress distribution simulated in S3 to form the physical basis of the digital twin.

[0022] Specifically, the stress at a slope height of 20 meters is approximately 2 kPa, with little difference before and after the landslide. However, the stress changes more significantly with decreasing slope height, showing a trend of increasing stress with increasing slope gradient. Specific values ​​are as follows... Figure 4 As shown in the figure. The left figure shows the stress situation before the landslide, and the right figure shows the stress situation after the landslide.

[0023] S5: Based on the stress value obtained in S4, construct a three-dimensional scene base and integrate the instability threshold and rainfall field to form a landslide digital twin.

[0024] Specifically, the 3D scene is constructed in UE5. To build a visualized 3D landslide scene based on a structured database combining numerical simulation and physical properties, a 3D terrain is first constructed as a base, and then points with the same slope are extracted. For example... Figure 5 As shown: Figure a is the base of the 3D scene; Figure b shows the extraction of points with the same slope gradient; Figure c is a solid model superimposed on Figure b, using spheres as the basic unit. The stress values ​​for different slope gradients and heights obtained from S4 were added to each basic unit to obtain the 3D scene representation. Figure d shows that after the operation started, the location where the landslide first occurred was at one-tenth of the slope height in the area of ​​maximum slope gradient, which is consistent with the conclusion summarized in S4 that the higher the slope, the lower the slope height. This forms a digital twin that integrates physical properties and rainfall field data from multiple sources, capable of reflecting the entire landslide process and providing real-time simulation. This work corresponds to... Figure 1 The third part of the flowchart.

[0025] In practical applications, after the aforementioned numerical simulations and model embedding, adding rainfall as a triggering factor allows for application in real-world scenarios, such as... Figure 6 The actual application scenarios shown are as follows: Figure a represents 15 mm of rainfall per 24 hours, Figure b represents 25 mm per 24 hours, and Figure c represents 60 mm per 24 hours. However, no landslides occurred in the first three scenarios. The rainfall in Figure d just matches the extremely heavy rain level mentioned earlier, 260 mm per 24 hours. It is under this condition that a landslide occurred.

[0026] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a digital twin framework for landslides, characterized in that: Includes the following steps: S1: Analyze the periods when landslides are most likely to occur and construct the causes of landslides; S2: Based on the landslide causes obtained in S1, the instability of slopes with different soil properties and different slope gradients is simulated. S3: Based on simulation, key parameters of slope stability before and after evolution are obtained, including stress field, void ratio, and displacement field. The changes in stress and land cover properties before and after landslide are analyzed. S4: Obtain the land feature attributes and stress distribution simulated in S3 to form the basis for building a digital twin, and store the stress values ​​corresponding to the slope height; S5: Based on the digital twin obtained in S4, construct a three-dimensional scene base, integrate instability threshold and rainfall field to form a landslide digital twin, construct a three-dimensional landslide scene visualization scene, and perform visual monitoring and early warning of the landslide process.

2. The method for constructing a landslide digital twin framework according to claim 1, characterized in that: The simulation in S2 is based on Abaqus software, which establishes simulation models with different slopes. The specific process is as follows: starting from a slope of 40°, the simulation is performed once for every 5° increase, up to 80°. The overall length of the simulation model is 50 meters, the height is 30 meters, and the foundation soil layer is 10 meters. The simulation model parameters include dry density, deformation modulus, Poisson's ratio, saturated permeability coefficient, initial void ratio, effective cohesion, effective internal friction angle, and rainfall level.

3. The method for constructing a landslide digital twin framework according to claim 1, characterized in that: The key parameters for simulation in S3 also include Mises stress.

4. The method for constructing a digital twin framework for landslides according to claim 1, characterized in that: The specific steps for constructing a three-dimensional visualization scene of a landslide in S5 are as follows: construct a three-dimensional terrain based on the UE5 software, extract several slope points based on the three-dimensional terrain, and overlay a solid model based on the slope points. Use each solid model as a basic unit, input the stress value corresponding to the slope height stored in S4 into each basic unit, and perform real-time simulation.

Images