Rainfall slope stability assessment method based on physical determinacy coupling model

Abstract

The invention relates to the technical field of rainfall slope evaluation, in particular to a rainfall slope stability evaluation method based on a physical certainty coupling model. The method comprises the following steps: acquiring digital elevation model data, engineering geological survey data and historical rainfall data of a target area; extracting topographic features from the digital elevation model data to obtain basic topographic parameters; the method comprises the following steps: carrying out longitudinal layering division on a rock-soil body based on engineering geological survey data, and carrying out physical and mechanical parameter setting on each layered rock-soil body to obtain a rock-soil body layering parameter set; and performing frequency analysis on the historical rainfall data to obtain an extreme rainfall intensity sequence. According to the method, by fusing the multi-source data and the three-dimensional physical coupling model, dynamic and accurate rainfall induced slope stability evaluation and risk early warning are realized, and the prediction precision is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rainfall slope evaluation, and in particular to a rainfall slope stability evaluation method based on a physical deterministic coupling model. BACKGROUND

[0002] In order to effectively evaluate and predict the stability of the slope under the action of rainfall, domestic and foreign researchers gradually develop from the empirical model to the physical mechanism model, and the physical deterministic model represented by TRIGRS gradually becomes a research hotspot for landslide prediction and risk assessment because it considers the coupling relationship of rainfall infiltration, pore water pressure change and slope stability response. The traditional slope stability analysis method mainly adopts two-dimensional limit equilibrium analysis, which only evaluates the safety factor on a specific section and ignores the three-dimensional nature of the terrain and the spatial complexity of the sliding surface.

[0003] There are still some deficiencies: first, the existing method has limited accuracy in processing the three-dimensional sliding surface space identification of the slope, and it is difficult to reflect the real sliding mechanism; second, some models only focus on a single rainfall duration scenario and cannot systematically simulate the dynamic response process of pore pressure and stability under different rainfall durations; third, the model results are mainly expressed in two-dimensional safety factor charts, and there is a lack of complete three-dimensional stability zoning results; in addition, most studies do not include landslide historical disaster data in the model verification process, which limits the practicality of the model in risk warning. SUMMARY

[0004] Therefore, it is necessary to provide a rainfall slope stability evaluation method based on a physical deterministic coupling model to solve at least one of the above technical problems.

[0005] To achieve the above purpose, a rainfall slope stability evaluation method based on a physical deterministic coupling model comprises the following steps: Step S1: Obtain the digital elevation model data, engineering geological survey data and historical rainfall data of the target area; extract the terrain features from the digital elevation model data to obtain the basic terrain parameters; perform longitudinal stratification division on the rock-soil mass based on the engineering geological survey data, and set the physical and mechanical parameters of each stratified rock-soil mass to obtain the stratified rock-soil mass parameter set; perform frequency analysis on the historical rainfall data to obtain an extreme rainfall intensity sequence; Step S2: Based on the basic terrain parameters, the stratified rock-soil mass parameter set and the extreme rainfall intensity sequence, a TRIGRS model is constructed, and the pore water pressure values under different rainfall durations in the target area are simulated to obtain two-dimensional pore water pressure distribution data; based on the two-dimensional pore water pressure distribution data, the stratified rock-soil mass parameter set and the pre-set slope mechanical relationship, an limit equilibrium model is constructed, and the safety factor of the slope unit is calculated to obtain two-dimensional slope safety factor; Step S3: Perform spatial interpolation processing on the two-dimensional pore water pressure distribution data, and construct a three-dimensional pore water pressure field by combining the digital elevation model data and the two-dimensional slope safety factor; construct a three-dimensional slope stability analysis model based on the three-dimensional pore water pressure field, the digital elevation model data, and the rock and soil layer parameter set, and perform spatial identification of the sliding surface and limit equilibrium calculation to obtain a three-dimensional minimum safety factor distribution result; Step S4: performing stability level classification on the three-dimensional minimum safety factor distribution result to obtain a three-dimensional stability partition diagram; Step S5: Obtain historical landslide hazard point distribution data; verify the three-dimensional stability partition map based on the historical landslide hazard point distribution data to obtain a model prediction accuracy evaluation index; Step S6: Based on the three-dimensional stability zoning map and the model prediction accuracy evaluation index, a distribution map of high-risk areas for rainfall-induced landslides is drawn, and a dynamic landslide risk early warning mechanism is established.

[0006] By integrating multi-source terrain, geological and meteorological data, the present invention establishes a three-dimensional slope stability assessment process that can dynamically respond to changes in rainfall conditions, significantly improving the spatial accuracy and physical credibility of landslide prediction. This method not only overcomes the problem of traditional two-dimensional analysis ignoring the three-dimensional spatial morphology of the sliding surface, but also systematically introduces the evolution of pore water pressure under different rainfall duration conditions to comprehensively reflect the stability change trend of the slope in continuous or sudden rainfall scenarios. By constructing a continuous three-dimensional pore pressure field and stability distribution field, a spatially explicit expression of slope stability is achieved, which helps to more intuitively identify potential unstable areas. At the same time, the method integrates historical landslide disaster data for verification and analysis, which can effectively quantify the prediction accuracy of the model and improve the reliability of risk identification. Finally, the prediction results are combined to generate a high-risk area distribution map and construct an early warning mechanism, which enhances the timeliness and guidance of landslide risk prevention and control, and has good engineering applicability and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 This is a schematic flow chart of the steps of a rainfall slope stability assessment method based on a physical deterministic coupling model according to the present invention; Figure 2 for Figure 1 Detailed step flow diagram of step S1; Figure 3 for Figure 1 Detailed step flow chart of step S3 in FIG. DETAILED DESCRIPTION

[0008] The technical method of the present application will be described clearly and completely below in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0009] In addition, the drawings are only schematic illustrations of the present application and are not necessarily drawn to scale. Identical reference signs in the drawings represent identical or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities, which do not necessarily have to correspond to physically or logically independent entities. The functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0010] It should be understood that although the terms "first", "second" and the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the exemplary embodiments, a first element can be referred to as a second element, and similarly a second element can be referred to as a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0011] To achieve the above-mentioned purpose, please refer to Figures 1 to 3 The present application provides a rainfall slope stability evaluation method based on a physically deterministic coupling model, which comprises the following steps: Step S1: Obtain digital elevation model data, engineering geological survey data and historical rainfall data of a target area; extract terrain features from the digital elevation model data to obtain basic terrain parameters; perform longitudinal layering division of rock-soil bodies based on the engineering geological survey data, and set physical and mechanical parameters of each layering rock-soil body to obtain a rock-soil body layering parameter set; perform frequency analysis on the historical rainfall data to obtain an extreme rainfall intensity sequence; Step S2: Construct a TRIGRS model based on the basic terrain parameters, the rock-soil body layering parameter set and the extreme rainfall intensity sequence, and simulate pore water pressure values under different rainfall durations in the target area to obtain two-dimensional pore water pressure distribution data; construct a limit equilibrium model based on the two-dimensional pore water pressure distribution data, the rock-soil body layering parameter set and a pre-set slope mechanical relationship, calculate the safety factor of the slope unit, and obtain two-dimensional slope safety factors; Step S3: spatial interpolation processing is performed on the two-dimensional pore water pressure distribution data, and a three-dimensional pore water pressure field is constructed in combination with the digital elevation model data and the two-dimensional slope safety factor; a three-dimensional slope stability analysis model is constructed based on the three-dimensional pore water pressure field, the digital elevation model data and the rock-soil layering parameter set, and sliding surface spatial recognition and limit equilibrium calculation are performed to obtain a three-dimensional minimum safety factor distribution result; Step S4: the three-dimensional minimum safety factor distribution result is subjected to stability grade division to obtain a three-dimensional stability partition map; Step S5: historical landslide disaster point distribution data are obtained; the three-dimensional stability partition map is verified based on the historical landslide disaster point distribution data to obtain a model prediction accuracy evaluation index; Step S6: a high-risk area distribution map of rainfall-induced landslides is drawn based on the three-dimensional stability partition map and the model prediction accuracy evaluation index, and a dynamic landslide risk early warning mechanism is constructed.

[0012] Preferably, step S1 comprises the following steps: Step S11: digital elevation model data, engineering geological survey data and historical rainfall data of the target area are obtained; Step S12: the digital elevation model data are extracted and analyzed to obtain basic terrain parameters; Step S13: rock-soil profile analysis and vertical structure modeling are performed based on the engineering geological survey data to obtain rock-soil longitudinal layering structure data; Step S14: classification and assignment of engineering geological parameters are performed on each layer of rock-soil based on the rock-soil longitudinal layering structure data to obtain a rock-soil layering parameter set; Step S15: Pearson III type frequency analysis is performed based on the historical rainfall data to obtain an extreme rainfall intensity sequence.

[0013] In the embodiment of the application, first, digital elevation model (DEM) data of the target area are obtained, and the resolution thereof should be no less than 8 meters to ensure that the terrain details can be accurately reflected. At the same time, engineering geological survey data are collected, which cover the type and distribution of rock-soil and other information, and historical rainfall data are collected, which record rainfall events in the past 30 years to ensure the statistical reliability of the data. The DEM data are processed by using geographic information system (GIS) software to extract terrain features such as slope and aspect. The slope calculation formula is The aspect is determined by calculating the elevation difference of 8 directions around each pixel point to obtain basic terrain parameters. Based on the engineering geological survey data, a rock-soil longitudinal layering model is adopted to divide the rock-soil into a soil layer and a bedrock weathering layer. The physical and mechanical parameters of the soil layer are set as cohesive force 10 kPa, internal friction angle 25° and unit weight 20 The parameters of the bedrock weathering layer are cohesive force 45 kPa, internal friction angle 40° and unit weight 22 , and form the longitudinal layered structure data of the rock-soil body. The engineering geological parameters are classified and valued for the longitudinal layered structure data of the rock-soil body. Specifically, for the soil layer, the permeability coefficient is set to , and the hydraulic dispersion coefficient is ; for the weathered bedrock layer, the permeability coefficient is set to , and the hydraulic dispersion coefficient is . Through the setting of these parameters, the layered parameter set of the rock-soil body is obtained. The historical rainfall data is subjected to Pearson III type frequency analysis. Specifically, first, the mean, standard deviation and skewness coefficient of the historical rainfall data are calculated, and then the extreme rainfall intensity sequence of different return periods (such as once every 100 years, once every 50 years, etc.) is calculated according to the formula of Pearson III type distribution. Through the above analysis, the extreme rainfall intensity sequence is obtained, such as the rainfall intensity of once every 100 years .

[0014] The present application realizes efficient acquisition and accurate modeling of the pre-analysis basic data of the slope stability by system integration of topographic, geological and meteorological multi-source data, significantly improves the completeness of the model input and the rationality of the parameter setting. The real topographic features are obtained by digital elevation model analysis, providing reliable spatial basis for subsequent hydraulic and mechanical simulation; the vertical structure modeling and parameter assignment process of the rock-soil body ensures the hierarchical expression of the rock-soil medium properties, makes the slope profile information closer to the actual geological conditions, and enhances the physical consistency and resolution capability of the analysis model; and the extreme rainfall intensity sequence is constructed by Pearson III type frequency analysis, which not only improves the scientificity and representativeness of the rainfall scenario simulation, but also provides multi-time and multi-intensity input basis for dynamic response analysis, thereby laying a solid data and theoretical support for subsequent pore water pressure simulation and slope stability evaluation. The combination of the steps effectively enhances the adaptability, prediction accuracy and engineering applicability of the model.

[0015] Preferably, the TRIGRS model is constructed based on the basic topographic parameters, the layered parameter set of the rock-soil body and the extreme rainfall intensity sequence in step S2, and the pore water pressure values under different rainfall durations in the target area are simulated, including: constructing the TRIGRS model based on the basic topographic parameters, the layered parameter set of the rock-soil body and the extreme rainfall intensity sequence, and initializing the parameters of the TRIGRS model to obtain the TRIGRS model; simulating the transient distribution of the pore water pressure in the target area under different rainfall durations based on the TRIGRS model to obtain two-dimensional pore water pressure distribution data , and the specific simulation process is as follows: ; ; ; ; ; ; wherein is the groundwater pressure head; is the thickness of the soil layer in the vertical direction, is the rainfall infiltration time, is the initial groundwater level depth of the rock-soil mass, is the time series, is the infiltration rate at the nth time period, is the vertical saturated permeability coefficient, is the convergence series, is the Heavyside step function, is the time of the nth rainfall intensity occurrence, is the depth of the basement, is the first integral value of the Gaussian complementary error function, is the slope, is the initial surface flux, is the saturated hydraulic diffusion coefficient, the degree of inclination of the slope surface, is the modified value of the saturated hydraulic diffusion coefficient.

[0016] In the embodiment of the present application, first, the TRIGRS model is constructed by using the basic terrain parameters, the rock-soil mass layering parameter set and the extreme rainfall intensity sequence, and the model is parameterized. The basic terrain parameters include the slope and the elevation, which are extracted from the digital elevation model (DEM) data. The rock-soil mass layering parameter set contains the physical and mechanical parameters of the soil layer and the bedrock weathering layer. The cohesion of the soil layer is 10 kPa, the internal friction angle is 25°, and the unit weight is 20 . The cohesion of the bedrock weathering layer is 45 kPa, the internal friction angle is 40°, and the unit weight is 22 . The extreme rainfall intensity sequence is obtained by Pearson type III frequency analysis, and the 100-year rainfall intensity is . In the TRIGRS model, the parameterization includes setting the initial groundwater level depth to 1 meter, the vertical saturated permeability coefficient to , and the saturated hydraulic diffusion coefficient to . The model is based on the linear solution of the Richards equation of Iverson (2000) to simulate the transient distribution of pore water pressure under different rainfall durations. The specific calculation formula is: ; ; ; ; ; ; wherein is the groundwater pressure head, which is related to the soil thickness and time; is the thickness of the soil layer in the vertical direction, which is positive in the vertical downward direction, is the rainfall infiltration time, is the initial groundwater level depth of the rock-soil body, is the time sequence, is the infiltration rate in the nth time period, is the vertical saturated permeability coefficient, is the convergence series, is the Heavyside step function, which is related to the rainfall intensity at the nth time and is a time function of the rainfall intensity, is the time at which the nth rainfall intensity occurs, is the depth of the base, is the first integral value of the Gaussian complementary function, which converges rapidly when the expression is in an infinite series, is the slope, is the initial surface flux, is the saturated hydraulic diffusion coefficient, is the degree of inclination of the slope surface, is the modified value of the saturated hydraulic diffusion coefficient. Through the above formula, the pore water pressure values under different rainfall durations are calculated to form two-dimensional pore water pressure distribution data. When the nonlinear formula of the Richards equation in the prior art is used, the transient distribution of the pore water pressure can be obtained, and by applying the linear solution formula of the Richards equation of Iverson (2000) provided in the present application, the transient distribution of the pore water pressure can be more accurately calculated.

[0017] The application realizes dynamic response simulation of pore water pressure under different rainfall durations by introducing a TRIGRS physical mechanism model, combining multi-source spatial terrain parameters, rock-soil body properties and extreme rainfall scenarios, and effectively improves the accuracy and timeliness of the modeling of the hydro-mechanical process of the slope. In the parameter initialization stage, the topography, rock-soil layer characteristics and soil permeability difference are fully considered to ensure the physical rationality and spatial consistency of the water infiltration and pore pressure change process. By simulating the transient pore water pressure distribution, the soil saturation evolution and pore pressure accumulation process caused by rainfall infiltration can be accurately captured, and the potential instability mechanism and its spatio-temporal evolution law are revealed, which provides key hydraulic input for subsequent slope safety factor calculation and three-dimensional stability analysis, and significantly enhances the response sensitivity of the overall evaluation system to rainfall intensity and duration changes and the engineering adaptability.

[0018] Preferably, the step S2 of constructing a limit equilibrium model based on the two-dimensional pore water pressure distribution data, the rock-soil body layered parameter set and the preset slope mechanical relationship, and calculating the safety factor of the slope unit comprises: calculating the safety factor of each slope grid unit based on the two-dimensional pore water pressure distribution data, the rock-soil body layered parameter set and the preset slope mechanical relationship to obtain a two-dimensional slope safety factor, and the specific calculation process is: ; wherein, the safety factor is, the internal friction angle of the soil layer is, the cohesion of the soil layer is, the groundwater unit weight is, the slope is, the pressure water head is a function of depth and time.

[0019] In the embodiment of the application, first, the pore water pressure values of each slope grid unit under different rainfall durations are extracted from the two-dimensional pore water pressure distribution data obtained from the TRIGRS model. These data reflect the changes of the pore water pressure in the rainfall infiltration process and are one of the key inputs for calculating the safety factor. Secondly, combined with the rock-soil body layered parameter set, the physical and mechanical parameters of the rock-soil body layer to which each slope grid unit belongs are determined. As described above, the cohesion of the soil layer is 10 kPa, the internal friction angle is 25°, and the unit weight is 20 The cohesion of the weathered bedrock layer is 45 kPa, the internal friction angle is 40°, and the unit weight is 22 These parameters directly affect the anti-sliding capacity of the slope. Then, according to the preset slope mechanical relationship, i.e., the limit equilibrium theory, the safety factor of each slope grid unit is calculated. The specific calculation formula is: ; wherein, the safety factor is, is the internal friction angle of the soil layer, For soil cohesion, is the groundwater bulk density, is the slope, is the pressure head, which is a function of depth and time, and the cohesion and The smaller the value of and the higher the groundwater level, the smaller the safety factor and the more unstable the slope. Using the above formula, the safety factor for each slope grid cell within the target area is calculated under different rainfall durations. While conventional limit equilibrium theory formulas (such as the Mohr-Coulomb criterion) can be used to determine the safety factor, the safety factor can be calculated more accurately by applying the safety factor calculation formula provided by the present invention, which combines limit equilibrium theory with the infinite slope model, particularly when considering the influence of pore water pressure.

[0020] Based on the introduction of dynamic simulation results of pore water pressure, the present invention combines the relationship between rock and soil stratification properties and slope mechanics to achieve refined and dynamic calculation of safety factors in slope stability assessment, effectively enhancing the response ability of the analysis model to changes in soil shear strength under the influence of rainfall. By quantitatively calculating the safety factor for each grid unit, it is possible to fully capture the spatial distribution characteristics of slope stability in the region, identify local potential unstable areas, and improve the spatial resolution of risk identification. At the same time, this process fully integrates the hydraulic-mechanical coupling effect, avoiding the simplified error of traditional methods that ignore the impact of pore pressure changes on slope stability, enhancing the physical basis and engineering reliability of the assessment results, and providing key two-dimensional stability basic data for subsequent three-dimensional stability modeling and high-risk area identification.

[0021] Preferably, step S3 includes the following steps: Step S31: performing spatial interpolation processing on the two-dimensional pore water pressure distribution data, and performing elevation registration in combination with the digital elevation model data and the two-dimensional slope safety factor to obtain a three-dimensional pore water pressure field; Step S32: constructing a three-dimensional slope grid structure based on the three-dimensional pore water pressure field, digital elevation model data and rock and soil layer parameter set, abstracting each DEM grid unit into a spatial cylindrical unit, and obtaining a three-dimensional slope analysis cylindrical model; Step S33: performing spatial identification of sliding surfaces based on the three-dimensional slope analysis column model to obtain a set of sliding surface candidate sets; Step S34: Calculate the safety factor corresponding to each sliding surface based on the three-dimensional pore water pressure field, the rock and soil layer parameter set, and the sliding surface candidate set to obtain a three-dimensional minimum safety factor distribution result.

[0022] In the embodiment of the present application, firstly, the two-dimensional pore water pressure distribution data is subjected to spatial interpolation processing. The bilinear interpolation method is adopted to convert the two-dimensional pore water pressure data from a regular grid to an irregular grid to adapt to complex terrain. Combined with digital elevation model (DEM) data and two-dimensional slope safety factor, elevation registration is performed. The specific operation is to match the elevation information in the DEM data with the two-dimensional pore water pressure data, to ensure that each pore water pressure value corresponds to the correct elevation position, thereby obtaining a three-dimensional pore water pressure field. Based on the three-dimensional pore water pressure field, digital elevation model data and rock-soil body layering parameter set, a three-dimensional slope grid structure is constructed. Each DEM grid unit is abstracted as a spatial column unit, and the height of each column unit is determined by the DEM data, and the width and length are consistent with the resolution of the DEM grid. The physical and mechanical parameters of each column unit are determined according to the rock-soil body layering to which it belongs, such as the parameters of the soil layer and the weathered layer of bedrock. In this way, a three-dimensional slope analysis column model is obtained. Based on the three-dimensional slope analysis column model, the spatial recognition of the sliding surface is performed. The spherical surface search method is adopted to search for the possible sliding surface with each column unit as the center. The specific operation is to set an initial radius, expand outward from the center point, and gradually increase the radius until a surface that meets the sliding condition is found. The sliding condition is based on the limit equilibrium theory, considering the pore water pressure and rock-soil body parameters. In this way, a sliding surface candidate set is obtained. Based on the three-dimensional pore water pressure field, rock-soil body layering parameter set and sliding surface candidate set, the safety factor corresponding to each sliding surface is calculated. For each sliding surface, the pore water pressure distribution thereon is calculated, and combined with the rock-soil body parameters, the Bishop limit equilibrium theory is used to calculate the safety factor. The specific formula is: ; wherein, is the cohesion of the i-th layer, is the sliding surface length of the i-th layer, is the unit weight of the i-th layer, is the height of the i-th layer, is the slope of the i-th layer, is the groundwater unit weight, is the pore water pressure height. Through the above formula, the safety factor of each sliding surface is calculated, and the three-dimensional minimum safety factor distribution result is obtained.

[0023] The present application realizes fine reconstruction of the pore water pressure field from two-dimensional plane to three-dimensional space by introducing spatial interpolation and height registration means, and improves the spatial coupling accuracy between the hydraulic conditions and the topographic features. Combined with three-dimensional column modeling, the slope profile is converted into a spatially continuous calculation unit, which facilitates the comprehensive expression of the vertical variation characteristics of the rock-soil hierarchical structure and hydraulic state, and enhances the physical reality and geometric expression of the three-dimensional modeling. On this basis, the spatial recognition of the sliding surface is carried out, which breaks through the dependence of the traditional two-dimensional analysis on the preset sliding surface form, can automatically capture the potential sliding surface form inside the slope, and reflects a more complex instability mechanism. Finally, through the stability evaluation of each candidate sliding surface, the three-dimensional minimum safety factor distribution result is obtained, the quantitative identification and positioning of the slope instability risk in the spatial scale are realized, and the precision, depth and engineering application value of the overall evaluation system are significantly improved. When using the conventional Bishop limit equilibrium theory formula in the field, the three-dimensional minimum safety factor can be obtained, and by applying the Scoops3D three-dimensional stability analysis method (based on the Bishop limit equilibrium theory) provided by the present application, the three-dimensional minimum safety factor can be more accurately calculated.

[0024] Preferably, step S33 comprises the following steps: Step S331: performing spatial position traversal and profile division on the three-dimensional slope analysis column model, constructing a sliding surface search domain based on a spherical surface search method, and obtaining a sliding surface search initial domain data set; Step S332: based on the sliding surface search initial domain data set, rotating to construct spherical surface profiles of different radii, inclination angles and azimuth angle combinations with the DEM grid unit as the three-dimensional column center, and obtaining multiple groups of sliding surface geometric parameter combinations; Step S333: performing three-dimensional spatial fitting analysis on the multiple groups of sliding surface geometric parameter combinations, determining the contact area and cutting relationship of the multiple groups of sliding surface geometric parameter combinations and the three-dimensional slope analysis column model, extracting a sliding path, and obtaining a sliding surface candidate set.

[0025] In the embodiment of the present application, firstly, the spatial position of the three-dimensional slope analysis column model is traversed and the profile is divided. Starting from the starting position of the model, the spatial position of each column unit is checked one by one, and the entire model is divided into multiple profiles according to the preset grid resolution. Each profile contains several column units, and the division direction of the profile is consistent with the potential sliding direction of the slope. Based on the spherical surface search method, a sliding surface search domain is constructed with each column unit as the center. The initial radius of the sliding surface search domain is set to 10 meters, the inclination angle range is from 0° to 90°, and the azimuth angle range is from 0° to 360°, and the step size is 5°. In this way, the initial domain data set of the sliding surface search is obtained. Based on the initial domain data set of the sliding surface search, the DEM grid unit is taken as the center of the three-dimensional column, and the spherical surface profile of different radius, inclination angle and azimuth angle combinations is constructed by rotation. The specific operation is as follows: for each column unit, starting from the initial radius, the radius is gradually increased by 5 meters each time until the preset maximum radius of 50 meters is reached. For each radius, the spherical surface profile is constructed by rotation according to the step size of the inclination angle and the azimuth angle. In this way, multiple groups of sliding surface geometric parameter combinations are obtained, each group containing a specific radius, inclination angle and azimuth angle. The three-dimensional space fitting analysis is performed on the multiple groups of sliding surface geometric parameter combinations. The specific operation is as follows: compare each group of sliding surface geometric parameter combinations with the three-dimensional slope analysis column model to determine the contact area and cutting relationship with the model. The contact area refers to the intersection of the sliding surface and the slope surface, and the cutting relationship refers to the path of the sliding surface through the interior of the slope. By extracting these contact points and cutting paths, the sliding path is obtained. Finally, all the sliding paths that meet the sliding conditions are collected to form a sliding surface candidate set. The sliding condition is based on the limit equilibrium theory, considering the pore water pressure and rock-soil parameters.

[0026] The present application realizes the active detection and multi-parameter combination matching of the geometric shape of the sliding surface in the three-dimensional space by constructing a sliding surface spatial recognition mechanism based on the spherical surface search method, effectively breaking through the prior assumption limit of the shape and position of the sliding surface in the traditional two-dimensional analysis. By traversing the three-dimensional slope analysis column model in space and dividing the profile, the search process ensures full coverage of the entire slope area, improving the comprehensiveness and accuracy of the potential sliding area recognition; the multi-parameter combination construction method fully considers the different azimuths, inclination angles and curvature changes that the sliding surface may have, improving the adaptability of the search results to the actual complex sliding mechanism; the fitting analysis and contact relationship determination make the sliding path extraction have high spatial consistency and mechanical rationality, providing a diverse and reliable sliding surface candidate basis for subsequent safety factor calculation, significantly enhancing the intelligence, automation level and engineering applicability of the sliding surface recognition in three-dimensional slope stability evaluation.

[0027] Preferably, step S4 comprises the following steps: Step S41: attribute extraction and raster reclassification processing are performed on the three-dimensional minimum safety factor distribution result, and stability grade value is assigned to each grid unit to obtain stability grade grid data; Step S42: classification processing is performed based on the stability grade grid data to obtain a stability grade division layer, wherein the classification processing is specifically that a safety factor less than 0.75 is an extremely unstable area, a safety factor greater than or equal to 0.75 and less than 1 is an unstable area, a safety factor greater than or equal to 1 and less than 1.25 is a basically stable area, a safety factor greater than or equal to 1.25 and less than 1.5 is a stable area, and a safety factor greater than or equal to 1.5 is an extremely stable area; Step S43: three-dimensional space fusion rendering is performed based on the digital elevation model data and the stability grade division layer to construct a three-dimensional stability division map.

[0028] In the embodiment of the present application, first, attribute extraction and raster reclassification processing are performed on the three-dimensional minimum safety factor distribution result. The safety factor value of each grid unit is extracted from the three-dimensional minimum safety factor distribution result. According to the preset stability grade standard, each grid unit is reclassified, the grid unit with a safety factor less than 0.75 is assigned as an extremely unstable area, the grid unit with a safety factor greater than or equal to 0.75 and less than 1 is assigned as an unstable area, the grid unit with a safety factor greater than or equal to 1 and less than 1.25 is assigned as a basically stable area, the grid unit with a safety factor greater than or equal to 1.25 and less than 1.5 is assigned as a stable area, and the grid unit with a safety factor greater than or equal to 1.5 is assigned as an extremely stable area. In this way, stability grade grid data is obtained. Classification processing is performed based on the stability grade grid data to obtain a stability grade division layer. The specific operation is that the stability grade grid data is imported into a geographic information system (GIS) software, the classification function of the GIS is used, the grid data is divided into five categories according to the above-mentioned stability grade standard, and the five categories correspond to the extremely unstable area, the unstable area, the basically stable area, the stable area and the extremely stable area respectively. Each category is identified by different colors or symbols to generate the stability grade division layer. Three-dimensional space fusion rendering is performed based on the digital elevation model (DEM) data and the stability grade division layer to construct a three-dimensional stability division map. The specific operation is that the DEM data and the stability grade division layer are imported into a three-dimensional visualization software such as ArcGIS Pro or QGIS 3D. In the software, the DEM data is used as the terrain basis, and the stability grade division layer is superimposed on the DEM. By setting the transparency, color and symbol properties, the areas with different stability can be clearly visible in the three-dimensional space. The three-dimensional rendering function of the software is used to generate the three-dimensional stability division map to intuitively display the stability conditions of different areas.

[0029] The application realizes the standardized expression and spatial partition of the slope stability state by attribute reclassification and grade assignment of the three-dimensional minimum safety factor distribution results, effectively enhances the visualization and operability of the slope stability evaluation results. The stability grade division standard is clear, covering the whole process state from extremely unstable to extremely stable, which is convenient for engineering management personnel to quickly identify the risk level and develop differentiated protection strategies. Further fusion of digital elevation model for three-dimensional space rendering makes the stability partition results have intuitive topographic dependence and spatial realism, improving the expression effect and use experience. The overall processing procedure not only improves the interpretation efficiency and application value of the analysis results, but also provides clear spatial basis and decision support for subsequent landslide risk assessment, disaster prevention layout optimization and dynamic early warning system.

[0030] Preferably, step S5 comprises the following steps: Step S51: Obtain historical landslide disaster point distribution data and perform spatial formatting processing to obtain a standardized landslide point vector data set; Step S52: Perform spatial overlay analysis based on the three-dimensional stability partition map and the standardized landslide point vector data set, count the number and proportion of landslide points falling into each stability level area, and obtain landslide point spatial matching statistical results; Step S53: Determine the area proportion of the high-risk area in the three-dimensional stability partition map, wherein the high-risk area is the area with a safety factor less than 1; Step S54: Perform quantitative evaluation of the model prediction accuracy based on the landslide point spatial matching statistical results and the area proportion of the high-risk area, and obtain a prediction accuracy evaluation index , and the specific quantitative evaluation process is as follows: ; Wherein S is the landslide point proportion, and A is the prediction area proportion.

[0031] In the embodiment of the present application, first, historical landslide disaster point distribution data is acquired, which should include the accurate geographic coordinates of the landslide points. The data is processed in a spatial format to convert it into a standardized vector data set, ensuring that the position information of each landslide point is accurate. Specifically, the landslide point data is imported into a geographic information system (GIS) software, and the vectorization function of the GIS is used to convert each landslide point into a vector point with a unique identifier, forming a standardized landslide point vector data set. Spatial overlay analysis is performed based on the three-dimensional stability zoning map and the standardized landslide point vector data set. Specifically, the standardized landslide point vector data set is overlaid with the three-dimensional stability zoning map, and the spatial analysis function of the GIS is used to count the number of landslide points falling into each stability level area. The number of landslide points in each stability level area is calculated, and the proportion of the total number of landslide points is calculated to obtain the spatial matching statistical result of the landslide points. The area proportion of the high-risk area in the three-dimensional stability zoning map is determined. The high-risk area is defined as the area with a safety factor less than 1. Specifically, the area with a safety factor less than 1 is extracted from the three-dimensional stability zoning map, the total area of these areas is calculated, and the proportion of the total area of the study area is calculated to obtain the area proportion of the high-risk area. Based on the spatial matching statistical result of the landslide points and the area proportion of the high-risk area, the prediction accuracy of the model is quantitatively evaluated. Specifically, the proportion of landslide points falling into the high-risk area is calculated, i.e. the ratio of the number of landslide points in the high-risk area to the total number of landslide points. At the same time, the area proportion of the high-risk area is calculated. By comparing the two proportions, the prediction accuracy of the model is evaluated. The specific formula is: ; where S is the landslide point proportion, and A is the predicted area proportion.

[0032] The present application realizes objective verification and quantitative precision evaluation of slope stability prediction results by introducing historical landslide disaster point data and performing spatial coupling analysis with the three-dimensional stability zoning results, significantly improving the credibility and application persuasiveness of the model results. Spatial matching analysis can intuitively reflect the consistency between the actual occurrence position of landslide disasters and the predicted high-risk areas, providing data support for evaluating the discrimination ability of the model; the extraction of the high-risk area makes the model results have a more explicit spatial warning function, which helps to evaluate the risk coverage and regional control strength; finally, the prediction accuracy index established by the comprehensive statistical results provides a quantitative basis for the adaptability of different regional models and the effectiveness of early warning, enhancing the guidance and evaluation standardization level of the model in practical engineering applications.

[0033] Preferably, step S6 comprises the following steps: Step S61: Extract the extremely unstable area and unstable area grid range based on the three-dimensional stability zoning map, and perform spatial merging processing to obtain an initial landslide high-risk area judgment layer; Step S62: based on the prediction accuracy evaluation index and the landslide high-risk area preliminary judgment layer, weighted analysis processing is carried out to obtain an optimized landslide high-risk area distribution map; Step S63: real-time rainfall monitoring data is obtained; the real-time rainfall monitoring data and the optimized landslide high-risk area distribution map are superimposed and analyzed, and a landslide early warning response logic is constructed to obtain a landslide dynamic early warning rule set; Step S64: based on the optimized landslide high-risk area distribution map and the landslide dynamic early warning rule set, a landslide risk grading early warning model is constructed to obtain a grading response early warning mechanism of rainfall-induced landslides.

[0034] In the embodiment of the application, first, the grid ranges of the extremely unstable area (safety factor less than 0.75) and the unstable area (safety factor greater than or equal to 0.75 and less than 1) are extracted based on the three-dimensional stability zoning map. The grids of the two areas are spatially merged and processed using the spatial analysis function of the geographic information system (GIS) software to form a continuous layer, and an initial landslide high-risk area judgment layer is obtained. Based on the prediction accuracy evaluation index and the initial landslide high-risk area judgment layer, weighted analysis processing is carried out. Specifically, the prediction accuracy evaluation index is used as a weight factor to perform weighted processing on each grid in the initial landslide high-risk area judgment layer. The weight factor is standardized according to the numerical range of the prediction accuracy evaluation index to ensure that it is between 0 and 1. Through weighted processing, an optimized landslide high-risk area distribution map is obtained, which more accurately reflects the landslide high-risk area. Real-time rainfall monitoring data is obtained, which should include rainfall intensity and rainfall duration. The real-time rainfall monitoring data and the optimized landslide high-risk area distribution map are superimposed and analyzed. Specifically, the real-time rainfall data and the high-risk area distribution map are superimposed using the GIS software to analyze the influence of rainfall conditions on the high-risk area. Based on the superimposed analysis result, a landslide early warning response logic is constructed, and the specific rules are as follows: when the rainfall intensity exceeds a preset threshold (such as 10 mm / hour) and the rainfall duration exceeds a preset time (such as 12 hours), the landslide early warning is triggered. In this way, a landslide dynamic early warning rule set is obtained. Based on the optimized landslide high-risk area distribution map and the landslide dynamic early warning rule set, a landslide risk grading early warning model is constructed. Specifically, the high-risk area distribution map and the dynamic early warning rule set are combined, and according to the rainfall conditions and the regional stability, the landslide risk is divided into four grades: low risk, medium risk, high risk, and extremely high risk. Each grade corresponds to different early warning response measures.

[0035] The application effectively integrates stability zoning and historical prediction accuracy information through spatial merging and weighting analysis, improves the accuracy and spatial continuity of landslide high-risk area identification, and enhances the scientificity and practicality of risk area discrimination. Combined with real-time rainfall monitoring data and optimized high-risk area distribution, dynamic linkage of landslide warning response is realized, which can timely reflect the influence of rainfall change on slope stability and improve the timeliness and sensitivity of the warning system. The hierarchical warning model constructed according to the risk degree for hierarchical response is helpful to realize precise management and differentiated prevention and control of landslide risk, significantly enhances the pertinence and scientific decision support capability of landslide disaster prevention, promotes the intelligentization and dynamicization of the warning mechanism, and improves the effectiveness and engineering popularization value of the overall landslide risk management.

[0036] Preferably, step S63 comprises the following steps: Step S631: Obtain real-time rainfall monitoring data and perform time series analysis and spatial registration processing to obtain a real-time rainfall intensity data set; Step S632: Perform spatial overlap analysis based on the real-time rainfall intensity data set and the optimized landslide high-risk area distribution map, and determine whether each grid meets the trigger condition that the cumulative rainfall intensity is greater than or equal to the preset extreme rainfall intensity threshold to obtain a landslide response trigger unit layer; Step S633: Identify the stability zoning level of the landslide response trigger unit layer in the optimized landslide high-risk area distribution map and perform secondary discrimination processing to obtain a landslide dynamic response data set; Step S634: Obtain a landslide dynamic warning rule set based on the landslide dynamic response data set, the prediction accuracy evaluation index, and the constructed warning trigger condition logic.

[0037] In the embodiment of the application, first, real-time rainfall monitoring data is obtained, which should include rainfall intensity and rainfall time. The real-time rainfall monitoring data is analyzed in time series to extract the rainfall intensity value at each time point, and spatial registration processing is performed to ensure that each rainfall intensity value matches the corresponding geographic coordinates. Specifically, the real-time rainfall data is imported into a geographic information system (GIS) software, the time series of the rainfall data is analyzed using the time series analysis function of the GIS, and each rainfall intensity value is matched with the geographic coordinates in the DEM data through the spatial registration function to obtain a real-time rainfall intensity data set. Spatial overlap analysis is performed based on the real-time rainfall intensity data set and the optimized landslide high-risk area distribution map. Specifically, the real-time rainfall intensity data set and the optimized landslide high-risk area distribution map are superimposed, and the spatial analysis function of the GIS is used to determine whether the cumulative rainfall intensity of each grid unit meets the trigger condition of the preset extreme rainfall intensity threshold. The preset extreme rainfall intensity threshold is corresponding to a 100-year rainfall intensity. In this way, a landslide response trigger cell layer is obtained, which identifies the grid cells that meet the trigger condition. The landslide response trigger cell layer is identified in the stability zoning level of the optimized landslide high-risk area distribution map. Specifically, the landslide response trigger cell layer is superimposed and analyzed with the optimized landslide high-risk area distribution map, and the stability zoning level of each trigger cell is extracted. According to the stability zoning level, secondary discrimination processing is performed, specifically: if the trigger cell is located in the extremely unstable area or unstable area, it is directly determined as a high-risk response unit; if it is located in the basically stable area or stable area, its rainfall intensity and duration are further evaluated, and if the rainfall intensity exceeds the preset threshold and the duration exceeds the preset time, it is determined as a medium-risk response unit. In this way, a landslide dynamic response dataset is obtained. Based on the landslide dynamic response dataset, the prediction accuracy evaluation index and the construction of the early warning trigger condition logic, a landslide dynamic early warning rule set is obtained. Specifically, the landslide dynamic response dataset is combined with the prediction accuracy evaluation index, and according to the numerical range of the prediction accuracy evaluation index, each response unit in the landslide dynamic response dataset is weighted. At the same time, according to the early warning trigger condition logic, specific early warning rules are set. For example, if the prediction accuracy evaluation index of the high-risk response unit is greater than 0.8, a first-level early warning is triggered; if the prediction accuracy evaluation index of the medium-risk response unit is between 0.6 and 0.8, a second-level early warning is triggered.

[0038] The present application realizes accurate spatio-temporal matching of rainfall information and landslide high-risk areas by high-precision time series analysis and spatial registration of real-time rainfall data, and improves the real-time and spatial targeting of rainfall-induced landslide early warning. Combined with cumulative rainfall intensity threshold determination, potential landslide response trigger units are effectively identified, and scientific and quantitative management of early warning triggering is realized. Through secondary discrimination of the stability level of the trigger unit, the accuracy of the dynamic response data and the rationality of the risk classification are enhanced. The finally constructed dynamic early warning rule set integrates historical prediction accuracy and real-time response conditions, significantly improving the flexibility, intelligence level and response efficiency of the landslide early warning system, providing solid data support and technical support for timely prevention and control of landslide disasters.

[0039] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, the scope of the present application is not limited by the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the application file are intended to be included in the present application.

[0040] The foregoing is considered as illustrative only of the principles of the application. Numerous modifications and changes will readily occur to those skilled in the art, and it is intended to embrace all such modifications and changes that fall within the scope of the application. Accordingly, the application is not to be restricted in scope to the specific embodiments disclosed herein but is to be accorded the full scope that the principles and novel features request appropriately granted.

Claims

1. A rainfall slope stability assessment method based on a physical deterministic coupling model, characterized in that: The following steps are involved: Step S1: Acquire digital elevation model data, engineering geological survey data, and historical rainfall data of the target area; extract terrain features from the digital elevation model data to obtain basic terrain parameters; perform vertical stratification of the rock and soil based on the engineering geological survey data, and set physical and mechanical parameters for each layer of rock and soil to obtain a rock and soil layer parameter set; perform frequency analysis on the historical rainfall data to obtain an extreme rainfall intensity sequence; Step S2: Based on the basic terrain parameters, the rock and soil layer parameter set, and the extreme rainfall intensity sequence, a TRIGRS model is constructed, and the pore water pressure values ​​under different rainfall durations in the target area are simulated to obtain two-dimensional pore water pressure distribution data; based on the two-dimensional pore water pressure distribution data, the rock and soil layer parameter set, and the preset slope mechanics relationship, a limit equilibrium model is constructed to calculate the safety factor of the slope unit to obtain the two-dimensional slope safety factor; Step S3: Perform spatial interpolation processing on the two-dimensional pore water pressure distribution data, and construct a three-dimensional pore water pressure field by combining the digital elevation model data and the two-dimensional slope safety factor; construct a three-dimensional slope stability analysis model based on the three-dimensional pore water pressure field, the digital elevation model data, and the rock and soil layer parameter set, and perform spatial identification of the sliding surface and limit equilibrium calculation to obtain a three-dimensional minimum safety factor distribution result; Step S4: performing stability level classification on the three-dimensional minimum safety factor distribution result to obtain a three-dimensional stability partition diagram; Step S5: Obtain historical landslide hazard point distribution data; verify the three-dimensional stability partition map based on the historical landslide hazard point distribution data to obtain a model prediction accuracy evaluation index; Step S6: Based on the three-dimensional stability zoning map and the model prediction accuracy evaluation index, a distribution map of high-risk areas for rainfall-induced landslides is drawn, and a dynamic landslide risk early warning mechanism is established.

2. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: obtaining digital elevation model data, engineering geological survey data and historical rainfall data of the target area; Step S12: extracting and analyzing the digital elevation model data to obtain basic terrain parameters; Step S13: performing rock and soil profile analysis and vertical structure modeling based on engineering geological survey data to obtain rock and soil longitudinal layered structure data; Step S14: classifying and assigning engineering geological parameters to each layered rock and soil body based on the rock and soil body longitudinal layered structure data to obtain a rock and soil body layered parameter set; Step S15: Perform Pearson III frequency analysis based on historical rainfall data to obtain an extreme rainfall intensity series.

3. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: In step S2, the TRIGRS model is constructed based on the basic terrain parameters, the rock and soil layer parameter set, and the extreme rainfall intensity sequence, and the pore water pressure values ​​under different rainfall durations in the target area are simulated, including: The TRIGRS model is constructed based on basic terrain parameters, rock and soil layer parameter sets and extreme rainfall intensity sequence, and the parameters of the TRIGRS model are initialized to obtain the TRIGRS model; Based on the TRIGRS model, the transient distribution of pore water pressure under different rainfall durations in the target area is simulated to obtain two-dimensional pore water pressure distribution data. , the specific simulation process is: ； ； ； ； ； ； in is the groundwater pressure head; is the thickness of the soil layer in the vertical direction, is the rainfall infiltration time, is the initial groundwater level depth of the rock and soil mass, is a time series, is the infiltration rate in the nth time period, is the vertical saturated permeability coefficient, is a convergent series, is the Heavyside step function, is the time when the nth rainfall intensity occurs, is the depth of the base, is the first integral value of the Gaussian complementary difference function, is the slope, is the initial surface flux, is the saturated hydraulic diffusion coefficient, is the slope degree, is the corrected value of the saturated hydraulic diffusion coefficient.

4. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: In step S2, a limit equilibrium model is constructed based on the two-dimensional pore water pressure distribution data, the rock and soil layer parameter set, and the preset slope mechanics relationship. The safety factor of the slope unit is calculated, including: The safety factor of each slope grid cell is calculated based on the two-dimensional pore water pressure distribution data, the rock and soil layer parameter set, and the preset slope mechanics relationship to obtain the two-dimensional slope safety factor. The specific calculation process is as follows: ； in, is the safety factor, is the internal friction angle of the soil layer, For soil cohesion, is the groundwater bulk density, is the slope, is the pressure head.

5. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: performing spatial interpolation processing on the two-dimensional pore water pressure distribution data, and performing elevation registration in combination with the digital elevation model data and the two-dimensional slope safety factor to obtain a three-dimensional pore water pressure field; Step S32: constructing a three-dimensional slope grid structure based on the three-dimensional pore water pressure field, digital elevation model data and rock and soil layer parameter set, abstracting each DEM grid unit into a spatial cylindrical unit, and obtaining a three-dimensional slope analysis cylindrical model; Step S33: performing spatial identification of sliding surfaces based on the three-dimensional slope analysis column model to obtain a set of sliding surface candidate sets; Step S34: Calculate the safety factor corresponding to each sliding surface based on the three-dimensional pore water pressure field, the rock and soil layer parameter set, and the sliding surface candidate set to obtain a three-dimensional minimum safety factor distribution result.

6. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 5 is characterized in that: Step S33 includes the following steps: Step S331: performing spatial position traversal and cross-section division on the three-dimensional slope analysis column model, constructing a sliding surface search domain based on a spherical surface search method, and obtaining a sliding surface search initial domain data set; Step S332: Based on the sliding surface search initial domain dataset, with the DEM grid unit as the 3D cylinder center, spherical surface profiles with different radius, inclination and azimuth combinations are constructed by rotation to obtain multiple sets of sliding surface geometric parameter combinations; Step S333: Perform three-dimensional spatial fitting analysis on multiple sets of sliding surface geometric parameter combinations, determine the contact area and cutting relationship between the multiple sets of sliding surface geometric parameter combinations and the three-dimensional slope analysis column model, extract the sliding path, and obtain a sliding surface candidate set.

7. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: performing attribute extraction and grid reclassification processing on the three-dimensional minimum safety factor distribution result, and assigning a stability level to each grid unit to obtain stability level grid data; Step S42: Classification processing is performed based on the stability level grid data to obtain a stability level division layer, wherein the classification processing is specifically as follows: a safety factor less than 0.75 is an extremely unstable area, a safety factor greater than or equal to 0.75 and less than 1 is an unstable area, a safety factor greater than or equal to 1 and less than 1.25 is a basically stable area, a safety factor greater than or equal to 1.25 and less than 1.5 is a stable area, and a safety factor greater than or equal to 1.5 is an extremely stable area; Step S43: Perform three-dimensional space fusion rendering based on the digital elevation model data and the stability level classification layer to construct a three-dimensional stability partition map.

8. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: Step S5 includes the following steps: Step S51: Obtain historical landslide disaster point distribution data and perform spatial formatting to obtain a standardized landslide point vector dataset; Step S52: performing spatial overlay analysis based on the three-dimensional stability zoning map and the standardized landslide point vector dataset, counting the number and proportion of landslide points falling into each stability level area, and obtaining the landslide point spatial matching statistical results; Step S53: determining the area ratio of high-risk areas in the three-dimensional stability zoning map, wherein the high-risk areas are areas with a safety factor less than 1; Step S54: Based on the landslide point spatial matching statistical results and the proportion of high-risk areas, the model prediction accuracy is quantitatively evaluated to obtain the prediction accuracy evaluation index. ,The specific quantitative evaluation process is; ； Where S is the proportion of landslide points and A is the proportion of predicted area.

9. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 1 is characterized in that: Step S6 includes the following steps: Step S61: extracting the grid ranges of the extremely unstable area and the unstable area based on the three-dimensional stability zoning map, and performing spatial merging processing to obtain an initial landslide high-risk area judgment layer; Step S62: performing weighted analysis based on the prediction accuracy evaluation index and the initial judgment layer of the landslide high-risk area to obtain an optimized landslide high-risk area distribution map; Step S63: acquiring real-time rainfall monitoring data; performing overlay analysis on the real-time rainfall monitoring data and the optimized landslide high-risk area distribution map, and constructing landslide early warning response logic to obtain a landslide dynamic early warning rule set; Step S64: Based on the optimized landslide high-risk area distribution map and the landslide dynamic warning rule set, a landslide risk classification warning model is constructed to obtain a graded response warning mechanism for rainfall-induced landslides.

10. The rainfall slope stability assessment method based on physical deterministic coupling model according to claim 9 is characterized in that: Step S63 includes the following steps: Step S631: Acquire real-time rainfall monitoring data, perform time series analysis and spatial registration processing, and obtain a real-time rainfall intensity dataset; Step S632: Perform spatial overlap analysis based on the real-time rainfall intensity dataset and the optimized landslide high-risk area distribution map, and determine whether each grid meets the trigger condition that the cumulative rainfall intensity is greater than or equal to the preset extreme rainfall intensity threshold, thereby obtaining a landslide response trigger unit layer; Step S633: identifying the stability zoning level of the landslide response triggering unit layer in the optimized landslide high-risk area distribution map, and performing secondary discrimination processing to obtain a landslide dynamic response dataset; Step S634: Based on the landslide dynamic response data set, the prediction accuracy evaluation index and the construction of the warning trigger condition logic, a landslide dynamic warning rule set is obtained.

Images