A method for intelligent monitoring and support optimization of slope stability based on digital twins

Abstract

The present invention discloses a method for intelligent monitoring and support optimization of slope stability based on digital twins, comprising: extracting slope, curvature and geological stratification characteristics from a holographic terrain model, calculating the preliminary change trend of the stress distribution of the rock and soil body in combination with the finite element analysis method, and determining the spatial position distribution of the potential sliding surface of the slope; adjusting the data acquisition frequency of the sensor through a time series analysis method according to the spatial coordinate set of the real-time monitoring points, and obtaining a high-frequency monitoring data set reflecting the dynamic changes of stress and groundwater; extracting stress peaks and groundwater level fluctuation characteristics from the high-frequency monitoring data set, combining the spatial characteristics of the holographic terrain model, and using a data fusion algorithm to generate a multidimensional information matrix containing terrain, stress and water dynamics; judging according to the state parameters of the digital twin, if the stress distribution exceeds a preset threshold, recalculating the force distribution of the support structure through finite element analysis to obtain an adjusted slope support design scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and in particular to a method for intelligent monitoring and support optimization of slope stability based on digital twins. Background Art

[0002] Slope stability research is a core topic in the field of geotechnical engineering and geological disaster prevention and control. Its importance is directly related to the safety of infrastructure construction and the sustainability of the natural environment. Currently, traditional slope analysis methods mostly rely on manual surveys and two-dimensional imaging data, which have the disadvantages of limited coverage, delayed data updates, and insufficient adaptability to complex terrain. When faced with slopes with drastic terrain fluctuations or variable geological conditions, these methods often cannot accurately reflect the true three-dimensional characteristics, resulting in a lack of sufficient scientific basis for subsequent support design. In addition, existing monitoring methods are generally blind in the optimization of sensor layout and data acquisition frequency, making it difficult to fully capture the real-time characteristics of rock and soil stress distribution and groundwater dynamic changes.

[0003] In this field, the core challenges lie in achieving image data acquisition with strong adaptability to terrain undulations, improving the texture accuracy of 3D models, and coordinating and optimizing sensor layout and data acquisition. Because image acquisition fails to fully integrate terrain characteristics and temporal parameters, the resulting data lacks spatial resolution and dynamics, which in turn affects the accuracy of 3D modeling. Furthermore, the arbitrary design of sensor burial depth and spacing makes it difficult for monitoring data to accurately reflect the mechanical state and fluid motion patterns within the geotechnical structure. These unresolved technical factors make it difficult to achieve the organic integration of multidimensional information in slope support design when faced with complex geological conditions, creating a bottleneck in the application of the technology.

[0004] Therefore, how to generate a high-precision holographic terrain model through drone image acquisition and three-dimensional remote sensing technology, and on this basis optimize the sensor layout and data acquisition frequency to construct a multi-dimensional digital twin of the slope support construction process, has become a key issue in improving slope stability analysis and design reliability. Summary of the Invention

[0005] In response to the above-mentioned technical problems in related technologies, the present invention provides a digital twin-based intelligent monitoring and support optimization method for slope stability, which can solve the above-mentioned problems.

[0006] To achieve the above technical objectives, the technical solution of the present invention is implemented as follows:

[0007] A method for intelligent monitoring and support optimization of slope stability based on digital twins includes the following steps:

[0008] S100, for slope areas with dramatic terrain fluctuations, uses drones equipped with high-resolution cameras and lidar equipment to collect multi-angle image data, generating a comprehensive terrain image dataset with centimeter-level resolution.

[0009] S200, generating an initial three-dimensional model based on the collected terrain image data set, performing local interpolation processing on the texture fuzzy area, and obtaining a holographic terrain model with enhanced texture details;

[0010] S300: Extract slope, curvature, and geological stratification characteristics from the holographic terrain model, and use finite element analysis to calculate the initial trend of stress distribution in the rock and soil mass to determine the spatial distribution of the potential sliding surface of the slope.

[0011] S400, based on the position distribution of the potential sliding surface of the slope and combined with the buried depth and spacing of the sensors, obtain a set of spatial coordinates of the real-time monitoring points;

[0012] S500, adjusting the data acquisition frequency of the sensor by a time series analysis method based on the spatial coordinate set of the real-time monitoring points to obtain a high-frequency monitoring data set reflecting the dynamic changes of stress and groundwater;

[0013] S600 extracts stress peaks and groundwater level fluctuation characteristics from high-frequency monitoring data sets, combines them with the spatial characteristics of the holographic terrain model, and uses a data fusion algorithm to generate a multidimensional information matrix containing terrain, stress, and water dynamics;

[0014] S700: Based on the multi-dimensional information matrix, digital twin modeling technology is used to generate a dynamic simulation of the slope support construction process. The state parameters of the simulation are updated in combination with real-time monitoring data to obtain a digital twin that reflects the changes in slope stability.

[0015] S800: If the stress distribution exceeds a preset threshold value based on the state parameters of the digital twin, recalculate the force distribution of the support structure through finite element analysis to obtain an adjusted slope support design scheme;

[0016] S900 extracts key parameters from the adjusted slope support design plan, combines them with the latest image data collected by drone remote sensing technology, and optimizes the geometric and mechanical characteristics of the digital twin through an iterative update algorithm to obtain the final digital twin that is synchronized with the actual slope state.

[0017] Furthermore, S100 specifically includes the following steps:

[0018] S110, using a drone equipped with a camera and a laser radar to obtain multi-angle image data of the slope area, adjusting the flight path to cover the terrain, and obtaining a preliminary image data set;

[0019] S120, using a planning algorithm to analyze the preliminary image data set, determine the optimized parameters of the flight path, adjust the acquisition angle and altitude, and generate optimized flight path data;

[0020] S130, controlling the UAV device using the optimized flight path data to obtain high-resolution image data, determine the coverage completeness of the terrain, and obtain a second image data set;

[0021] S140, using a laser radar to process the terrain undulation characteristics of the second image dataset, determine the depth information of the slope area, and generate a third image dataset containing the depth information;

[0022] S150, analyzing the third image data set using a random forest algorithm to determine the correlation between the image data and terrain undulations, thereby obtaining classified terrain feature data;

[0023] S160, based on the classified terrain feature data, optimizing the acquisition viewing angle using height adjustment parameters to obtain final image data and generate a data set with a centimeter-level resolution;

[0024] S170 , using the final image dataset and a rasterization processing method to determine the full coverage characteristics of the slope area, thereby obtaining a terrain image dataset.

[0025] Furthermore, S200 specifically includes the following steps:

[0026] S210, acquiring point cloud data through terrain images, and processing the data using a structured light algorithm to generate three-dimensional structure data;

[0027] S220, extracting motion recovery features from the three-dimensional structure data, fusing the point cloud data to generate a preliminary three-dimensional model;

[0028] S230, detecting texture fuzzy areas on the preliminary three-dimensional model, and repairing the fuzzy areas using a local interpolation technique to obtain an optimized three-dimensional model;

[0029] S240, analyzing texture detail features based on the optimized three-dimensional model, determining detail enhancement integrity, and generating holographic three-dimensional data;

[0030] S250, extracting image processing parameters through holographic three-dimensional data, adjusting the data fusion method, and obtaining a fusion optimized data set;

[0031] S260, using the fused optimized data set to generate a final holographic model, determining the consistency of the three-dimensional generation, and obtaining a full-coverage terrain model;

[0032] S270. Analyze the distribution of terrain features for the full-coverage terrain model, classify the features using a random forest algorithm, and obtain classified terrain data.

[0033] Furthermore, S300 specifically includes the following steps:

[0034] S310, extracting slope data and curvature data through the holographic terrain model, obtaining terrain feature distribution information, and obtaining an initial terrain parameter set;

[0035] S320, using the initial terrain parameter set in combination with geological stratification data to generate a rock and soil mass feature description and determine a rock and soil mass feature set;

[0036] S330, obtaining parameters required for stress distribution calculation from the rock and soil mass feature set, and using a finite element analysis method to obtain a stress distribution change trend;

[0037] S340, analyzing the spatial characteristics of the sliding surface based on the stress distribution change trend, and determining the preliminary distribution range of the potential sliding surface;

[0038] S350, adjusting the boundary conditions of the sliding surface position by combining the preliminary distribution range with the curvature data to obtain an optimized sliding surface distribution;

[0039] S360: If there is a deviation between the optimized sliding surface distribution and the geological stratification data, supplement the rock and soil characteristic information through data acquisition to determine the final spatial position of the sliding surface;

[0040] S370. Extract the classification basis of the terrain features according to the final spatial position, use the random forest algorithm to process the classification, and obtain the slope stability zoning data.

[0041] Furthermore, S400 specifically includes the following steps:

[0042] S410, generating an initial point grid based on the position distribution of the sliding surface, and using a genetic algorithm to optimize the burial depth and spacing to obtain adjusted sensor layout parameters;

[0043] S420: Determine the coverage of the key area based on the adjusted sensor layout parameters and obtain an optimized sensor layout plan;

[0044] S430, extracting the coordinates of the real-time monitoring points from the spatial set for the optimized sensor deployment plan to obtain monitoring point distribution data;

[0045] S440: If the monitoring point distribution data is inconsistent with the boundary conditions of the key area, the point coordinates are supplemented by the position distribution to determine the adjusted point distribution range;

[0046] S450, using the adjusted point distribution range, obtaining dynamic monitoring data of the sliding surface position, and determining the coverage integrity of the real-time monitoring;

[0047] S460: By analyzing the coverage integrity of real-time monitoring and the stability of point coordinates using a genetic algorithm, the final sensor deployment plan is obtained.

[0048] S470. According to the final sensor deployment plan, a monitoring point set of a key area is extracted from the spatial set to obtain a spatial coordinate result of real-time monitoring.

[0049] Furthermore, S500 specifically includes the following steps:

[0050] S510, obtaining the initial data distribution of real-time monitoring through the spatial coordinate set, processing the time interval of data collection by time series analysis, and obtaining the adjusted collection frequency parameter;

[0051] S520, extracting the fluctuation characteristics of stress changes from the real-time monitoring based on the adjusted acquisition frequency parameters to obtain a time series set of high-frequency data;

[0052] S530: If the stress change in the time series set of high-frequency data exceeds a preset threshold, the sensor adjusts and updates the acquisition frequency to obtain a monitoring data set reflecting the dynamic change;

[0053] S540, obtaining the fluctuation characteristics of the groundwater level based on the monitoring data set reflecting the dynamic changes, and determining the time series correlation corresponding to the stress changes;

[0054] S550, analyzing the dynamic trend of changes through time series correlation, using clustering algorithm to divide the distribution area of ​​high-frequency data, and obtaining characteristic grouping of point distribution;

[0055] S560: Based on the feature grouping of point distribution, key monitoring points related to dynamic changes are extracted from spatial coordinates to obtain the spatial coverage of high-frequency data;

[0056] S570: Adjust sensor parameters for real-time monitoring according to the spatial coverage of the high-frequency data to obtain an optimized monitoring data set.

[0057] Furthermore, S600 specifically includes the following steps:

[0058] S610, extracting stress peaks and groundwater level fluctuation characteristics through high-frequency monitoring data to generate an initial feature set;

[0059] S620, using a data fusion algorithm to process the initial feature set, combined with the spatial distribution of the holographic terrain model, to obtain a multi-dimensional information matrix;

[0060] S630, dividing the dynamically changing areas of spatial distribution according to the multidimensional information matrix to determine the distribution range of the terrain features;

[0061] S640: If the dynamically changing area exceeds a preset threshold, adjust the high-frequency monitoring time interval based on the monitoring data to obtain an updated feature set;

[0062] S650: using a clustering algorithm to divide the distribution pattern of the fluctuation characteristics based on the updated feature set, and generating a classified information matrix;

[0063] S660, extracting key terrain features related to spatial distribution from the classified information matrix to obtain an optimized multi-dimensional information matrix;

[0064] S670. Analyze the correlation between the fluctuation characteristics and the dynamic changes through the optimized multi-dimensional information matrix to generate the final monitoring distribution set.

[0065] Furthermore, S700 specifically includes the following steps:

[0066] S710, obtaining spatial distribution characteristics of slope support through a multi-dimensional information matrix and generating an initial simulation shape;

[0067] S720, using digital twin modeling technology to process the initial simulation body shape to obtain a dynamic simulation body;

[0068] S730, extracting state parameters from the real-time monitoring data, updating the operating conditions of the dynamic simulation body, and obtaining an adjusted simulation body state;

[0069] S740, analyzing the spatial variation trend of the construction process for the adjusted simulated shape, and determining the distribution range of the stability variation;

[0070] S750: If the distribution range of the stability change exceeds a preset threshold, adjust the frequency of data updates through real-time monitoring data to obtain optimized state parameters;

[0071] S760, using a clustering algorithm to divide the fluctuation patterns of the dynamic simulation body according to the optimized state parameters, and obtaining a classified simulation body shape;

[0072] S770. Analyze the correlation between slope support and stability change through the classified simulation shape to generate a final simulation shape.

[0073] Furthermore, S800 specifically includes the following steps:

[0074] S810. If the stress distribution exceeds a preset threshold, the stress distribution of the support structure is processed by finite element analysis to obtain a preliminary adjustment plan;

[0075] S820, extract the change characteristics of the force distribution through the preliminary adjustment plan, and use the interpolation method to generate the distribution grid of the support structure;

[0076] S830, analyzing the spatial characteristics of the slope support based on the distribution grid to obtain an optimized force distribution;

[0077] S840: If the optimized force distribution still exceeds the preset threshold, adjust the boundary conditions of the finite element analysis using the state parameters to obtain an updated adjustment plan;

[0078] S850, dividing the stress-bearing areas of the branch retaining structure according to the updated adjustment plan, and using a clustering algorithm to determine the distribution pattern between the areas;

[0079] S860, generating dynamically updated parameters of slope support according to the distribution law, and obtaining a final support scheme;

[0080] S870. Analyze the fluctuation trend of stress distribution for the final support scheme and determine the frequency of dynamic updates.

[0081] Furthermore, S900 specifically includes the following steps:

[0082] S910, obtain the latest image data through remote sensing technology, use filtering methods to process the image data, and obtain a clear representation of the slope status;

[0083] S920: Extract key parameters from the slope state representation, combine them with the initial data in the design plan, and use an iterative update algorithm to adjust the geometric characteristics to obtain a preliminary updated digital twin;

[0084] S930: Analyze the mechanical properties of the initially updated digital twin through the extraction process, and if it is determined that the mechanical properties are inconsistent with the slope state, adjust the iterative step size of the update algorithm to obtain an optimized digital twin;

[0085] S940. Analyze the matching degree between the image data and the geometric characteristics based on the optimized digital twin, use interpolation methods to generate the spatial distribution of synchronous adjustment, and determine the boundary conditions of the mechanical characteristics;

[0086] S950: Extract change characteristics from the spatial distribution of synchronous adjustments, combine key parameters to update the design plan, and obtain a digital twin that is consistent with the actual slope state;

[0087] S960. For the consistent digital twin, obtain dynamic image data using remote sensing technology. If the image data exceeds a preset threshold, adjust the weights of the geometric and mechanical properties to obtain the final digital twin.

[0088] S970. Based on the final digital twin, a clustering algorithm is used to divide the regional characteristics of the slope state and determine the update frequency of the synchronous adjustment.

[0089] Beneficial effects of the invention: This application realizes all-round monitoring of slope topography, stress and water dynamics, as well as real-time optimization of support schemes, which significantly improves the safety and efficiency of the slope construction process and provides an intelligent solution for slope projects under complex terrain conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0091] The present invention will be described in further detail below with reference to the accompanying drawings.

[0092] Figure 1 This is a flow chart of a method for intelligent monitoring and support optimization of slope stability based on digital twins described in an embodiment of the present invention. DETAILED DESCRIPTION

[0093] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention are within the scope of protection of the present invention.

[0094] like Figure 1 As shown, according to the present invention, a method for intelligent monitoring and support optimization of slope stability based on digital twin is disclosed, comprising the following steps:

[0095] S100, for slope areas with dramatic terrain undulations, uses drones equipped with high-resolution cameras and lidar equipment to collect multi-angle image data, resulting in a comprehensive terrain image dataset with centimeter-level resolution.

[0096] Using drones equipped with cameras and lidar, multi-angle image data of the slope area is acquired. The flight path is adjusted to cover the terrain, generating a preliminary image dataset. A planning algorithm is used to analyze the preliminary image dataset, determine optimal flight path parameters, adjust the acquisition angle and altitude, and generate optimized flight path data. The optimized flight path data is used to control the drone to acquire high-resolution image data and determine the completeness of terrain coverage, generating a second image dataset. Lidar is used to process the terrain characteristics of the second image dataset, determine the depth information of the slope area, and generate a third image dataset containing this depth information. A random forest algorithm is used to analyze the third image dataset, determine the correlation between the image data and the terrain, and generate classified terrain feature data. Based on the classified terrain feature data, the acquisition angle is optimized using altitude adjustment parameters to obtain final image data, generating a dataset with centimeter-level resolution. The final image dataset is then rasterized to determine the complete coverage of the slope area, generating a terrain image dataset.

[0097] S200 , generating an initial three-dimensional model based on the collected terrain image data set, performing local interpolation processing on the texture fuzzy area, and obtaining a holographic terrain model with enhanced texture details.

[0098] Point cloud data is acquired from terrain imagery and processed using a structured light algorithm to generate 3D structural data. Motion recovery features are extracted from the 3D structural data and fused with the point cloud data to generate a preliminary 3D model. Texture blurring areas are detected within the preliminary 3D model and repaired using local interpolation techniques to obtain an optimized 3D model. Texture detail features are analyzed based on the optimized 3D model to determine the completeness of detail enhancement and generate holographic 3D data. Image processing parameters are extracted from the holographic 3D data and the data fusion method is adjusted to obtain a fused and optimized dataset. The fused and optimized dataset is used to generate the final holographic model, and the 3D generation consistency is determined to obtain a fully covered terrain model. The distribution of terrain features within the fully covered terrain model is analyzed, and the features are classified using a random forest algorithm to obtain classified terrain data.

[0099] S300. Extract the slope, curvature and geological stratification characteristics from the holographic terrain model, combine the finite element analysis method to calculate the preliminary change trend of the stress distribution of the rock and soil mass, and determine the spatial position distribution of the potential sliding surface of the slope.

[0100] The slope data and curvature data are extracted through the holographic model to obtain the terrain feature distribution information and obtain the initial terrain parameter set. The calculation formula is: , S represents the terrain characteristic parameter, z represents the elevation value, and x represents the spatial coordinate. Indicates the slope, represents the curvature, and n represents the number of spatial coordinates.

[0101] The initial terrain parameter set is combined with geological layering data to generate a rock and soil feature description and determine the rock and soil feature set. The parameters required for stress distribution calculation are obtained from the rock and soil feature set. The finite element analysis method is used to obtain the stress distribution change trend. The stress distribution function is: , σ represents the stress distribution function, represents vertical load, A represents the action area, represents the bending moment, represents the moment of inertia, and h represents the calculation depth.

[0102] The spatial characteristics of the sliding surface are analyzed based on the stress distribution trend to determine the preliminary distribution range of the potential sliding surface. By combining the preliminary distribution range with the curvature data, the boundary conditions of the sliding surface are adjusted to obtain the optimized sliding surface distribution. The calculation formula of the sliding surface distribution index is: , D represents the sliding surface distribution index, r represents the curvature radius, d represents the depth parameter, α represents the inclination coefficient, and P represents the number of sliding surfaces.

[0103] If there is a deviation between the optimized sliding surface distribution and the geological stratification data, the rock and soil characteristics information is supplemented through data acquisition to determine the final spatial position of the sliding surface. The classification basis of the terrain characteristics is extracted based on the final spatial position, and the random forest algorithm is used to process the classification to obtain the slope stability partition data. The corresponding slope stability index is: , F represents the slope stability index, c represents cohesion, σ represents normal stress, φ represents the internal friction angle, T represents shear force, and m represents the number of sliding surfaces.

[0104] S400: Based on the position distribution of the potential sliding surface of the slope and combined with the buried depth and spacing of the sensors, a spatial coordinate set of the real-time monitoring points is obtained.

[0105] An initial point grid is generated based on the position distribution of the sliding surface. A genetic algorithm is used to optimize the burial depth and spacing, resulting in adjusted sensor layout parameters. Based on the adjusted sensor layout parameters, the coverage range of the critical area is determined, resulting in an optimized sensor placement plan. Based on the optimized sensor placement plan, the coordinates of the real-time monitoring points are extracted from the spatial set to obtain the monitoring point distribution data. If the monitoring point distribution data is inconsistent with the boundary conditions of the critical area, the point coordinates are supplemented using the position distribution to determine the adjusted point distribution range. Using the adjusted point distribution range, dynamic monitoring data of the sliding surface is obtained to determine the coverage integrity of the real-time monitoring. The coverage integrity of the real-time monitoring is combined with the genetic algorithm to analyze the stability of the point coordinates and obtain the final sensor placement plan. Based on the final sensor placement plan, the monitoring point set for the critical area is extracted from the spatial set to obtain the spatial coordinate results of the real-time monitoring.

[0106] S500 , according to the spatial coordinate set of the real-time monitoring points, the data acquisition frequency of the sensor is adjusted by a time series analysis method to obtain a high-frequency monitoring data set reflecting the dynamic changes of stress and groundwater.

[0107] The initial data distribution for real-time monitoring is obtained using a spatial coordinate set. Time series analysis is used to process the data collection intervals to obtain an adjusted acquisition frequency parameter. Based on this adjusted acquisition frequency parameter, the fluctuation characteristics of stress changes are extracted from the real-time monitoring data to obtain a time series set of high-frequency data. If the stress change in the high-frequency data time series exceeds a preset threshold, the sensor adjusts and updates the acquisition frequency to obtain a monitoring data set that reflects dynamic changes. Based on this monitoring data set, the fluctuation characteristics of the groundwater level are obtained, and the time series correlation corresponding to the stress change is determined.

[0108] The fluctuation characteristics of groundwater level are expressed by the groundwater level fluctuation characteristic function F(s): , W represents the water level observation value, m represents the number of observations, α represents the fluctuation coefficient, β represents the spatial impact factor, and s represents the spatial distance. It represents the difference in observation time between two adjacent water level observations.

[0109] By analyzing dynamic trends through time series correlation, a clustering algorithm is used to segment the distribution of high-frequency data, generating characteristic groupings of point distribution. Based on these characteristic groupings, key monitoring points associated with dynamic changes are extracted from spatial coordinates to determine the spatial coverage of the high-frequency data. Based on this spatial coverage, the parameters of the real-time monitoring sensors are adjusted to produce an optimized monitoring dataset.

[0110] The sensor parameter optimization function S(v) is expressed as: , v p represents the current parameter value, k represents the number of parameters, γ represents the optimization coefficient, and v max and v min Represent the maximum and minimum values ​​of the parameters respectively.

[0111] S600 extracts stress peaks and groundwater level fluctuation characteristics from high-frequency monitoring data sets, combines them with the spatial characteristics of the holographic terrain model, and uses a data fusion algorithm to generate a multidimensional information matrix containing terrain, stress, and water dynamics.

[0112] High-frequency monitoring data is used to extract the fluctuation characteristics of stress peaks and groundwater levels to generate an initial feature set. A data fusion algorithm is used to process the initial feature set, and combined with the spatial distribution of the holographic terrain model, a multidimensional information matrix is ​​obtained. The dynamic change areas of spatial distribution are divided according to the multidimensional information matrix to determine the distribution range of terrain features. If the dynamic change area exceeds the preset threshold, the high-frequency monitoring time interval is adjusted based on the monitoring data to obtain an updated feature set. For the updated feature set, a clustering algorithm is used to divide the distribution pattern of the fluctuation characteristics to generate a classified information matrix. Key terrain features related to spatial distribution are extracted from the classified information matrix to obtain an optimized multidimensional information matrix. The correlation between the fluctuation characteristics and dynamic changes is analyzed using the optimized multidimensional information matrix to generate the final monitoring distribution set.

[0113] S700, for the multi-dimensional information matrix, uses digital twin modeling technology to generate a dynamic simulation of the slope support construction process. Combined with real-time monitoring data, the state parameters of the simulation are updated to obtain a digital twin that reflects the changes in slope stability.

[0114] The spatial distribution characteristics of slope support are obtained through a multidimensional information matrix to generate an initial simulation shape. Digital twin modeling technology is used to process the initial simulation shape to obtain a dynamic simulation body. State parameters are extracted from real-time monitoring data, and the operating conditions of the dynamic simulation body are updated to obtain an adjusted simulation shape. Based on the adjusted simulation shape, the spatial variation trend of the construction process is analyzed to determine the distribution range of stability changes. If the distribution range of stability changes exceeds the preset threshold, the frequency of data updates is adjusted through real-time monitoring data to obtain optimized state parameters. Based on the optimized state parameters, a clustering algorithm is used to divide the fluctuation pattern of the dynamic simulation body to obtain a classified simulation shape. The correlation between slope support and stability changes is analyzed through the classified simulation shape to generate the final simulation shape.

[0115] S800: Based on the state parameters of the digital twin, if the stress distribution exceeds a preset threshold, the force distribution of the support structure is recalculated through finite element analysis to obtain an adjusted slope support design scheme.

[0116] If the stress distribution exceeds the preset threshold, the force distribution of the support structure is processed through finite element analysis to obtain a preliminary adjustment plan. The changing characteristics of the force distribution are extracted through the preliminary adjustment plan, and the distribution grid of the support structure is generated by the interpolation method. The spatial characteristics of the slope support are analyzed based on the distribution grid to obtain the optimized force distribution. If the optimized force distribution still exceeds the preset threshold, the boundary conditions of the finite element analysis are adjusted through the state parameters to obtain an updated adjustment plan. According to the updated adjustment plan, the force areas of the branch retaining structure are divided, and the distribution pattern between areas is determined by the clustering algorithm. The dynamic update parameters of the slope support are generated based on the distribution pattern to obtain the final support plan. The fluctuation trend of the stress distribution is analyzed for the final support plan to determine the frequency of dynamic updates.

[0117] S900 extracts key parameters from the adjusted slope support design plan, combines them with the latest image data collected by drone remote sensing technology, and optimizes the geometric and mechanical characteristics of the digital twin through an iterative update algorithm to obtain the final digital twin that is synchronized with the actual slope state.

[0118] Remote sensing technology is used to acquire the latest image data, which is then processed using filtering methods to obtain a clear representation of the slope state. Key parameters are extracted from the slope state representation and, combined with the initial data from the design solution, an iterative update algorithm is used to adjust the geometric characteristics, resulting in a preliminary updated digital twin. For this preliminary updated digital twin, mechanical properties are analyzed during the extraction process. If the mechanical properties are inconsistent with the slope state, the iterative step size of the update algorithm is adjusted to obtain an optimized digital twin. Based on the optimized digital twin, the degree of match between the image data and the geometric properties is analyzed. An interpolation method is used to generate a spatial distribution of synchronous adjustments and determine the boundary conditions for the mechanical properties. Change characteristics are extracted from the synchronously adjusted spatial distribution and, combined with key parameters, the design solution is updated to obtain a digital twin consistent with the actual slope state. For this consistent digital twin, dynamic image data is acquired using remote sensing technology. If the image data exceeds a preset threshold, the weights of the geometric and mechanical properties are adjusted to obtain the final digital twin. Based on the final digital twin, a clustering algorithm is used to partition the regional characteristics of the slope state and determine the update frequency for synchronous adjustments.

[0119] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for intelligent monitoring and support optimization of slope stability based on digital twins, characterized in that: The steps include: S100, for slope areas with dramatic terrain fluctuations, uses drones equipped with high-resolution cameras and lidar equipment to collect multi-angle image data, generating a comprehensive terrain image dataset with centimeter-level resolution. S200, generating an initial three-dimensional model based on the collected terrain image data set, performing local interpolation processing on the texture fuzzy area, and obtaining a holographic terrain model with enhanced texture details; S300: Extract slope, curvature, and geological stratification characteristics from the holographic terrain model, and use finite element analysis to calculate the initial trend of stress distribution in the rock and soil mass to determine the spatial distribution of the potential sliding surface of the slope. S400, based on the position distribution of the potential sliding surface of the slope and combined with the buried depth and spacing of the sensors, obtain a set of spatial coordinates of the real-time monitoring points; S500, adjusting the data acquisition frequency of the sensor by a time series analysis method based on the spatial coordinate set of the real-time monitoring points to obtain a high-frequency monitoring data set reflecting the dynamic changes of stress and groundwater; S600 extracts stress peaks and groundwater level fluctuation characteristics from high-frequency monitoring data sets, combines them with the spatial characteristics of the holographic terrain model, and uses a data fusion algorithm to generate a multidimensional information matrix containing terrain, stress, and water dynamics; S700: Based on the multi-dimensional information matrix, digital twin modeling technology is used to generate a dynamic simulation of the slope support construction process. The state parameters of the simulation are updated in combination with real-time monitoring data to obtain a digital twin that reflects the changes in slope stability. S800: If the stress distribution exceeds a preset threshold value based on the state parameters of the digital twin, recalculate the force distribution of the support structure through finite element analysis to obtain an adjusted slope support design scheme; S900 extracts key parameters from the adjusted slope support design plan, combines them with the latest image data collected by drone remote sensing technology, and optimizes the geometric and mechanical characteristics of the digital twin through an iterative update algorithm to obtain the final digital twin that is synchronized with the actual slope state.

2. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S100 specifically includes the following steps: S110, using a drone equipped with a camera and a laser radar to obtain multi-angle image data of the slope area, adjusting the flight path to cover the terrain, and obtaining a preliminary image data set; S120, using a planning algorithm to analyze the preliminary image data set, determine the optimized parameters of the flight path, adjust the acquisition angle and altitude, and generate optimized flight path data; S130, controlling the UAV device using the optimized flight path data to obtain high-resolution image data, determine the coverage completeness of the terrain, and obtain a second image data set; S140, using a laser radar to process the terrain undulation characteristics of the second image dataset, determine the depth information of the slope area, and generate a third image dataset containing the depth information; S150, analyzing the third image data set using a random forest algorithm to determine the correlation between the image data and terrain undulations, thereby obtaining classified terrain feature data; S160, based on the classified terrain feature data, optimizing the acquisition viewing angle using height adjustment parameters to obtain final image data and generate a data set with a centimeter-level resolution; S170 , using the final image dataset and a rasterization processing method to determine the full coverage characteristics of the slope area, thereby obtaining a terrain image dataset.

3. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S200 specifically includes the following steps: S210, acquiring point cloud data through terrain images, and processing the data using a structured light algorithm to generate three-dimensional structure data; S220, extracting motion recovery features from the three-dimensional structure data, fusing the point cloud data to generate a preliminary three-dimensional model; S230, detecting texture fuzzy areas on the preliminary three-dimensional model, and repairing the fuzzy areas using a local interpolation technique to obtain an optimized three-dimensional model; S240, analyzing texture detail features based on the optimized three-dimensional model, determining detail enhancement integrity, and generating holographic three-dimensional data; S250, extracting image processing parameters through holographic three-dimensional data, adjusting the data fusion method, and obtaining a fusion optimized data set; S260, using the fused optimized data set to generate a final holographic model, determining the consistency of the three-dimensional generation, and obtaining a full-coverage terrain model; S270. Analyze the distribution of terrain features for the full-coverage terrain model, classify the features using a random forest algorithm, and obtain classified terrain data.

4. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S300 specifically includes the following steps: S310, extracting slope data and curvature data through the holographic terrain model, obtaining terrain feature distribution information, and obtaining an initial terrain parameter set; S320, using the initial terrain parameter set in combination with geological stratification data to generate a rock and soil mass feature description and determine a rock and soil mass feature set; S330, obtaining parameters required for stress distribution calculation from the rock and soil mass feature set, and using a finite element analysis method to obtain a stress distribution change trend; S340, analyzing the spatial characteristics of the sliding surface based on the stress distribution change trend, and determining the preliminary distribution range of the potential sliding surface; S350, adjusting the boundary conditions of the sliding surface position by combining the preliminary distribution range with the curvature data to obtain an optimized sliding surface distribution; S360: If there is a deviation between the optimized sliding surface distribution and the geological stratification data, supplement the rock and soil characteristic information through data acquisition to determine the final spatial position of the sliding surface; S370. Extract the classification basis of the terrain features according to the final spatial position, use the random forest algorithm to process the classification, and obtain the slope stability zoning data.

5. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S400 specifically includes the following steps: S410, generating an initial point grid based on the position distribution of the sliding surface, and using a genetic algorithm to optimize the burial depth and spacing to obtain adjusted sensor layout parameters; S420: Determine the coverage of the key area based on the adjusted sensor layout parameters and obtain an optimized sensor layout plan; S430, extracting the coordinates of the real-time monitoring points from the spatial set for the optimized sensor deployment plan to obtain monitoring point distribution data; S440: If the monitoring point distribution data is inconsistent with the boundary conditions of the key area, the point coordinates are supplemented by the position distribution to determine the adjusted point distribution range; S450, using the adjusted point distribution range, obtaining dynamic monitoring data of the sliding surface position, and determining the coverage integrity of the real-time monitoring; S460: By analyzing the coverage integrity of real-time monitoring and the stability of point coordinates using a genetic algorithm, the final sensor deployment plan is obtained. S470. According to the final sensor deployment plan, a monitoring point set of a key area is extracted from the spatial set to obtain a spatial coordinate result of real-time monitoring.

6. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S500 specifically includes the following steps: S510, obtaining the initial data distribution of real-time monitoring through the spatial coordinate set, processing the time interval of data collection by time series analysis, and obtaining the adjusted collection frequency parameter; S520, extracting the fluctuation characteristics of stress changes from the real-time monitoring based on the adjusted acquisition frequency parameters to obtain a time series set of high-frequency data; S530: If the stress change in the time series set of high-frequency data exceeds a preset threshold, the sensor adjusts and updates the acquisition frequency to obtain a monitoring data set reflecting the dynamic change; S540, obtaining the fluctuation characteristics of the groundwater level based on the monitoring data set reflecting the dynamic changes, and determining the time series correlation corresponding to the stress changes; S550, analyzing the dynamic trend of changes through time series correlation, using clustering algorithm to divide the distribution area of ​​high-frequency data, and obtaining characteristic grouping of point distribution; S560: Based on the feature grouping of point distribution, key monitoring points related to dynamic changes are extracted from spatial coordinates to obtain the spatial coverage of high-frequency data; S570: Adjust sensor parameters for real-time monitoring according to the spatial coverage of the high-frequency data to obtain an optimized monitoring data set.

7. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S600 specifically includes the following steps: S610, extracting stress peaks and groundwater level fluctuation characteristics through high-frequency monitoring data to generate an initial feature set; S620, using a data fusion algorithm to process the initial feature set, combined with the spatial distribution of the holographic terrain model, to obtain a multi-dimensional information matrix; S630, dividing the dynamically changing areas of spatial distribution according to the multidimensional information matrix to determine the distribution range of the terrain features; S640: If the dynamically changing area exceeds a preset threshold, adjust the high-frequency monitoring time interval based on the monitoring data to obtain an updated feature set; S650: using a clustering algorithm to divide the distribution pattern of the fluctuation characteristics based on the updated feature set, and generating a classified information matrix; S660, extracting key terrain features related to spatial distribution from the classified information matrix to obtain an optimized multi-dimensional information matrix; S670. Analyze the correlation between the fluctuation characteristics and the dynamic changes through the optimized multi-dimensional information matrix to generate the final monitoring distribution set.

8. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S700 specifically includes the following steps: S710, obtaining spatial distribution characteristics of slope support through a multi-dimensional information matrix and generating an initial simulation shape; S720, using digital twin modeling technology to process the initial simulation body shape to obtain a dynamic simulation body; S730, extracting state parameters from the real-time monitoring data, updating the operating conditions of the dynamic simulation body, and obtaining an adjusted simulation body state; S740, analyzing the spatial variation trend of the construction process for the adjusted simulated shape, and determining the distribution range of the stability variation; S750: If the distribution range of the stability change exceeds a preset threshold, adjust the frequency of data updates through real-time monitoring data to obtain optimized state parameters; S760, using a clustering algorithm to divide the fluctuation patterns of the dynamic simulation body according to the optimized state parameters, and obtaining a classified simulation body shape; S770. Analyze the correlation between slope support and stability change through the classified simulation shape to generate a final simulation shape.

9. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1 is characterized in that: S800 specifically includes the following steps: S810. If the stress distribution exceeds a preset threshold, the stress distribution of the support structure is processed by finite element analysis to obtain a preliminary adjustment plan; S820, extract the change characteristics of the force distribution through the preliminary adjustment plan, and use the interpolation method to generate the distribution grid of the support structure; S830, analyzing the spatial characteristics of the slope support based on the distribution grid to obtain an optimized force distribution; S840: If the optimized force distribution still exceeds the preset threshold, adjust the boundary conditions of the finite element analysis using the state parameters to obtain an updated adjustment plan; S850, dividing the stress-bearing areas of the branch retaining structure according to the updated adjustment plan, and using a clustering algorithm to determine the distribution pattern between the areas; S860, generating dynamically updated parameters of slope support according to the distribution law, and obtaining a final support scheme; S870. Analyze the fluctuation trend of stress distribution for the final support scheme and determine the frequency of dynamic updates.

10. The method for intelligent monitoring and support optimization of slope stability based on digital twin according to claim 1, characterized in that: S900 specifically includes the following steps: S910, obtain the latest image data through remote sensing technology, use filtering methods to process the image data, and obtain a clear representation of the slope status; S920: Extract key parameters from the slope state representation, combine them with the initial data in the design plan, and use an iterative update algorithm to adjust the geometric characteristics to obtain a preliminary updated digital twin; S930: Analyze the mechanical properties of the initially updated digital twin through the extraction process, and if it is determined that the mechanical properties are inconsistent with the slope state, adjust the iterative step size of the update algorithm to obtain an optimized digital twin; S940. Analyze the matching degree between the image data and the geometric characteristics based on the optimized digital twin, use interpolation methods to generate the spatial distribution of synchronous adjustment, and determine the boundary conditions of the mechanical characteristics; S950: Extract change characteristics from the spatial distribution of synchronous adjustments, combine key parameters to update the design plan, and obtain a digital twin that is consistent with the actual slope state; S960. For the consistent digital twin, obtain dynamic image data using remote sensing technology. If the image data exceeds a preset threshold, adjust the weights of the geometric and mechanical properties to obtain the final digital twin. S970. Based on the final digital twin, a clustering algorithm is used to divide the regional characteristics of the slope state and determine the update frequency of the synchronous adjustment.

Images