Creeping type landslide stability evaluation method and device

Abstract

The invention discloses a creep type landslide stability evaluation method and device, and relates to the field of landslide evaluation, and the method comprises the following steps: obtaining landslide basic information, and setting a three-dimensional landslide space model according to the landslide basic information; key point extraction is carried out on the set three-dimensional landslide space model, and multivariate real-time monitoring data at all key points are obtained; setting a basic evaluation model at the key point, and dynamically adjusting the basic evaluation model according to the multivariate real-time monitoring data to obtain a dynamic evaluation model; obtaining an association trigger release conditional relationship between corresponding data information in each key point, and setting a three-dimensional landslide key point association map; performing comprehensive stability evaluation on the multivariate real-time monitoring data according to the set three-dimensional landslide key point association map to obtain comprehensive stability evaluation data; the accuracy of the stability evaluation process is improved to a certain extent.

Description

A method and apparatus for evaluating the stability of creep-type landslides Technical Field

[0001] This invention relates to the field of landslide stability management technology, and in particular to a method and apparatus for evaluating the stability of creep-type landslides. Background Technology

[0002] In existing technologies, stability assessments of creep landslides largely rely on field monitoring data and empirical judgments, lacking systematic quantitative analysis methods. This results in highly subjective and inaccurate assessments. While some methods incorporate mechanical models for calculation, they fail to fully consider the time effects and multi-factor coupling during creep, making it difficult to accurately reflect the long-term evolution characteristics of landslides. Furthermore, existing devices are mostly limited to single-parameter monitoring, unable to achieve multi-source information fusion analysis, thus limiting the improvement of early warning capabilities. Therefore, there is an urgent need for an assessment method and supporting devices that can comprehensively consider geological conditions, environmental factors, and time-sensitivity characteristics to improve the scientific rigor and reliability of creep landslide stability assessments.

[0003] A search revealed Chinese invention patent CN111581694A, which discloses a method and apparatus for evaluating the stability of creep landslides, improving the reliability of stability evaluation for creep landslides. The method includes: establishing a creep landslide mechanical model under conditions where environmental variations alter the initial sliding conditions; calculating stability coefficients based on the established creep landslide mechanical model; and evaluating the stability of the creep landslide based on the calculated stability coefficients. This invention is applicable to the stability evaluation of creep landslides.

[0004] Compared with existing technologies, the invention patent with Chinese patent number CN111581694A calculates the stability coefficient by establishing a stability model for creep-type landslides under the condition that the initial sliding conditions are changed due to environmental variations, which can improve the reliability and accuracy of creep-type landslide stability evaluation.

[0005] However, in actual use, the above methods and devices cannot verify and update the established mechanical model of creep landslide based on the changes at different locations of the creep landslide, which to some extent affects the accuracy of the stability evaluation process. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of insufficient accuracy in existing technologies by proposing a method and apparatus for evaluating the stability of creep-type landslides.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for evaluating the stability of creeping landslides, comprising the following steps: Step S1: setting up a landslide stability evaluation platform, obtaining basic landslide information through the platform, and setting up a three-dimensional landslide spatial model based on the basic landslide information; Step S2: extracting key points based on the set three-dimensional landslide spatial model, setting up a data monitoring terminal based on the key point extraction results, and obtaining multivariate real-time monitoring data through the data monitoring terminal; Step S3: setting the model parameters of the basic evaluation model based on the location information of each key point in the three-dimensional landslide spatial model, dynamically adjusting the model parameters of the basic evaluation model based on the multivariate real-time monitoring data, and obtaining a dynamic evaluation model; Step S4: setting up association triggering and release conditions for the historical evaluation results of the dynamic evaluation model of each key point in the three-dimensional landslide spatial model, connecting each key point according to the key triggering conditions, and setting up a three-dimensional landslide key point association map; Step S5: performing a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map, and obtaining the comprehensive stability evaluation data of the corresponding landslide.

[0008] The above technical solution further includes: the process of setting up a three-dimensional landslide spatial model based on landslide basic information includes: setting up a landslide stability evaluation platform, which is equipped with an information input terminal, and obtaining landslide basic information through the information input terminal, including landslide attribute information, geological environment information, and monitored deformation information; generating three types of three-dimensional images of the landslide area in sequence based on the obtained landslide basic information, including topographic surface image, stratum slip surface image, and boundary load image; performing adaptive meshing on the generated three-dimensional images of the landslide area to obtain unit area meshes within the three-dimensional images of the landslide area; and performing parameter assignment processing within each unit area mesh to generate a three-dimensional landslide spatial model.

[0009] Furthermore, the process of acquiring multivariate real-time monitoring data through the data monitoring terminal includes: acquiring the parameter assignment processing results of each unit area grid within the three-dimensional landslide spatial model; comparing and analyzing the parameter assignment processing results of each adjacent unit area grid to obtain difference data; assessing the importance of the difference data; setting corresponding key points based on the importance assessment results; and marking the set key points within the three-dimensional landslide spatial model; setting up the data monitoring terminal based on the marking results within the three-dimensional landslide spatial model to which each key point belongs; the data monitoring terminal including monitoring terminals of multiple data types; acquiring multivariate real-time monitoring data at the corresponding locations of each key point through the data acquisition terminal; the multivariate real-time monitoring data including deformation monitoring data, stress monitoring data, hydrological monitoring data, and support monitoring data.

[0010] Furthermore, the process of setting the model parameters of the basic evaluation model based on the location information of each key point in the three-dimensional landslide spatial model includes: acquiring multivariate historical monitoring data at the location information of each key point; based on the parameter assignment processing results at the corresponding locations in the three-dimensional landslide spatial model where each key point is located, setting the corresponding key point datasets by setting the multivariate historical monitoring data and parameter assignment processing results at each key point; dividing the obtained key point datasets into multiple sub-datasets; analyzing and processing each obtained sub-dataset; and setting the corresponding basic evaluation model and model parameters of the basic evaluation model, wherein the basic evaluation model includes three types: limit equilibrium correction model, data simulation dynamic model, and machine learning early warning model.

[0011] Furthermore, the process of obtaining the dynamic evaluation model includes: performing a comparative correlation analysis between the parameter assignment processing results and multivariate historical monitoring data within the corresponding type of basic evaluation model subset and the corresponding model parameters to obtain a correlation table between the corresponding parameter assignment processing results and multivariate historical monitoring data and the corresponding type of model parameters; obtaining multivariate real-time monitoring data at the corresponding key point locations within the three-dimensional landslide spatial model; dynamically adjusting the reference assignment processing results at the key point locations based on the obtained multivariate real-time monitoring data to obtain real-time reference assignment processing results; comparing and analyzing the multivariate real-time monitoring data and real-time reference assignment processing results with the correlation table; and dynamically adjusting the model parameters of the basic evaluation model based on the comparison analysis results to obtain the dynamic evaluation model.

[0012] Furthermore, the process of setting the three-dimensional landslide key point association map includes: obtaining the historical evaluation results and historical evaluation time of the corresponding dynamic evaluation model at each key point in the three-dimensional landslide spatial model; setting vertical time-series dynamic change multivariate curves for the dynamic evaluation model and historical evaluation results in the three-dimensional landslide spatial model according to the historical evaluation time; the vertical time-series dynamic change multivariate curves include the model parameters of each dynamic evaluation model and the dynamic change connection of each type of historical evaluation result at each historical evaluation time; combining the elements of the vertical time-series dynamic change multivariate curves set at each key point in the three-dimensional landslide spatial model according to the key points between adjacent or spaced position information; and obtaining the corresponding vertical time-series dynamic change multivariate curves based on the element combination results. The data information within the dynamic multivariate curve is set up as an element combination dataset. Temporal features are extracted from the obtained element combination dataset based on its corresponding vertical dynamic multivariate curve. Data alignment is then performed on the data information within the element combination dataset based on the temporal feature extraction results. Based on the data alignment results, correlation determination analysis, triggering relationship determination analysis, and dissolution relationship determination analysis are performed sequentially to obtain the correlation, triggering, and dissolution conditions between different types of data information at various key points within the 3D landslide spatial model. Connection processing is then performed at each key point within the 3D landslide spatial model based on the correlation, triggering, and dissolution conditions corresponding to each element combination, constructing a 3D landslide key point correlation map corresponding to the 3D landslide spatial model.

[0013] Furthermore, the process of obtaining comprehensive stability assessment data for the corresponding landslide includes: acquiring multivariate real-time monitoring data at each key point within the three-dimensional landslide spatial model and model parameters of the dynamic evaluation model; performing an initial evaluation on the acquired multivariate real-time monitoring data based on the dynamic evaluation model with corresponding model parameters to obtain initial evaluation results; verifying and adjusting the initial evaluation results obtained at each key point within the three-dimensional landslide spatial model according to the association triggering and release conditions between different data types within each key point in the three-dimensional landslide key point association map; and performing a comprehensive stability evaluation on the initial evaluation results at each key point based on the verification and adjustment analysis results to obtain comprehensive stability assessment data.

[0014] A creep-type landslide stability evaluation device includes a data management module, a data acquisition module, a data processing module, a data analysis module, and a comprehensive evaluation module. The data management module acquires basic landslide information and sets up a three-dimensional landslide spatial model based on this information. The data acquisition module extracts key points from the set three-dimensional landslide spatial model, sets up a data monitoring terminal based on the key point extraction results, and acquires multivariate real-time monitoring data through the data monitoring terminal. The data processing module sets the model parameters of the basic evaluation model based on the location information of each key point within the three-dimensional landslide spatial model, and dynamically adjusts the model parameters of the set basic evaluation model based on the multivariate real-time monitoring data to obtain a dynamic evaluation model. The data analysis module sets association trigger and release conditions for the historical evaluation results of the dynamic evaluation model of each key point within the three-dimensional landslide spatial model, connects each key point according to the key trigger conditions, and sets up a three-dimensional landslide key point association map. The comprehensive evaluation module performs a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map to obtain the corresponding comprehensive stability evaluation data for the landslide.

[0015] The present invention has the following beneficial effects: 1. In the present invention, by analyzing and processing the obtained landslide basic information, setting a three-dimensional image of the landslide area, adaptively dividing the set three-dimensional image of the landslide area into grids, obtaining the unit area grid of the three-dimensional image of the landslide area, performing difference analysis based on the parameter assignment processing results in each unit area grid, setting corresponding key points, and setting multiple key points to conduct stability assessment of the creep-type landslide area respectively, the accuracy of the stability evaluation process can be improved to a certain extent.

[0016] 2. In this invention, the model parameters corresponding to the obtained basic evaluation model are dynamically adjusted based on the obtained multivariate real-time monitoring data to obtain a dynamic evaluation model. By analyzing the relationship between different model parameters and data types between the historical evaluation results of each key point at the historical evaluation time, a corresponding three-dimensional landslide key point association map is constructed. The stability evaluation process of creep landslide is verified and adjusted through the constructed three-dimensional landslide key point association map, which can improve the accuracy of the stability evaluation process to a certain extent. Attached Figure Description

[0017] Figure 1 is a flowchart of a creep-type landslide stability evaluation method proposed in this invention; Figure 2 is a structural diagram of a creep-type landslide stability evaluation device proposed in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, as shown in Figure 1, proposes a method for evaluating the stability of creep-type landslides, comprising the following steps: Step S1: Setting up a landslide stability evaluation platform, acquiring basic landslide information through the platform, and setting up a three-dimensional landslide spatial model based on the basic information; Step S2: Extracting key points from the set three-dimensional landslide spatial model, setting up a data monitoring terminal based on the key point extraction results, and acquiring multivariate real-time monitoring data through the data monitoring terminal; Step S3: Setting the model parameters of the basic evaluation model based on the location information of each key point in the three-dimensional landslide spatial model, dynamically adjusting the model parameters of the basic evaluation model based on the multivariate real-time monitoring data, and obtaining a dynamic evaluation model; Step S4: Setting up association triggering and release conditions for the historical evaluation results of the dynamic evaluation model of each key point in the three-dimensional landslide spatial model, connecting each key point according to the key triggering conditions, and setting up a three-dimensional landslide key point association map; Step S5: Performing a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map, and obtaining the comprehensive stability evaluation data of the corresponding landslide.

[0020] In the specific implementation process, the step S1 of setting up a landslide stability evaluation platform, obtaining basic landslide information through the landslide stability evaluation platform, and setting up a three-dimensional landslide spatial model based on the basic landslide information includes: S11: setting up a landslide stability evaluation platform, which is equipped with an information input terminal; S12: obtaining basic landslide information through the information input terminal, which includes landslide attribute information, geological environment information, and monitoring deformation information. The landslide attribute information includes geographical location information, scale characteristics, and morphological characteristics; the geological environment information includes topographic and geomorphological information, stratigraphic lithology information, hydrogeological information, and geological structure information; and the monitoring deformation information includes crack displacement information, meteorological monitoring information, and support information. Monitoring information and stress monitoring information; S13: Based on the obtained landslide basic information, three types of three-dimensional images of the landslide area are generated sequentially. The three types of three-dimensional images of the landslide area are divided into topographic surface images, stratum slip surface images and boundary load images; S131: Spatial geometric data information includes topographic elevation (DEM), landslide perimeter, slip surface morphology, stratum layer thickness, borehole coordinates / depth, etc.; Physical and mechanical information includes unit weight (γ), shear strength (c / φ), elastic modulus (E), Poisson's ratio (μ), permeability coefficient (k), etc. of each layer of rock and soil; Boundary condition information includes groundwater level, rainfall, seismic motion parameters, location / magnitude of engineering load, peak strength (c / φ), residual strength (cᵣ / φᵣ), etc.

[0021] S132: Standardize the obtained spatial geometric data, and use the standardized spatial geometric data to generate a 3D topographic surface image of the landslide area based on a software interpolation algorithm to describe the topographic undulations of the landslide area; S133: Standardize the obtained physical and mechanical information, obtain the stratigraphic profile and slip surface within the landslide area, and visualize the stratigraphic structure and slip surface morphology of the stratigraphic profile and slip surface within the landslide area respectively to generate a 3D stratigraphic and slip surface image of the landslide area; S134: Standardize and extract the obtained boundary condition information, obtain the mechanical boundary, seepage boundary, and external load within the landslide area, and map them to the corresponding locations in the landslide area to generate a boundary load image of the landslide area; S14: Adaptively mesh the generated 3D image of the landslide area, obtain the unit area grid within the 3D image of the landslide area, and perform parameter assignment within each unit area grid to generate a 3D landslide spatial model; S141: Overlay the generated 3D images of the landslide area, and then... Preset unit area grids are used to divide the 3D images of landslide areas of various types, and the types of reference data within the 3D images of landslide areas corresponding to each preset unit area grid are obtained; S142: If the types of reference data are singular, no adaptive adjustment is made to the unit area grid; if the types of reference data include multiple types, the unit area grid is adaptively reduced until the types of reference data corresponding to the adaptively reduced unit area grid are singular; it should be noted that: the reference data types being singular means that the spatial geometric type data information, physical and mechanical type information, and boundary condition type information corresponding to the unit area grid are all of the same type; in addition, during the adaptive adjustment of the unit area grid, an adaptive adjustment minimum unit is set, and the corresponding unit area grid is enlarged or reduced according to the adaptive adjustment minimum unit; S143: The superposition result of the 3D images of the landslide area is processed based on the basic landslide information according to the corresponding grid unit size, and a 3D landslide spatial model is generated based on the parameter assignment result of the superposition result of the 3D images of the landslide area.

[0022] In the specific implementation process, the step S2, which involves extracting key points based on the established three-dimensional landslide spatial model, setting up a data monitoring terminal based on the key point extraction results, and acquiring multi-dimensional real-time monitoring data through the data monitoring terminal, includes: S21: acquiring the parameter assignment processing results of each unit area grid within the three-dimensional landslide spatial model, comparing and analyzing the parameter assignment processing results of each adjacent unit area grid, and acquiring the difference data; it should be further explained that in the process of acquiring the difference data in step S21, the parameter assignment processing results within the unit area grid are compared with the parameter assignment processing results of each unit area grid within the three-dimensional landslide spatial model. Sort the parameters according to their corresponding assignment types; compare and analyze the parameter assignment results of each assignment type with the corresponding sorting results, and obtain the type sorting difference data of the parameter assignment results according to the comparison analysis results of each sorting result. Mark the type sorting difference data separately, and obtain the difference data corresponding to the parameter assignment results in each adjacent unit area grid; S22: Perform importance assessment based on the difference data, set corresponding key points according to the importance assessment results, and mark the set key points in the three-dimensional landslide space model; it needs to be further explained that In the process of assessing the importance of differential data, physical assessment, mutation assessment, location sensitivity assessment, and stability contribution assessment corresponding to different parameter assignment results are obtained. The intervals to which each assessment type belongs are obtained based on the corresponding differential data. An importance grading system database is set up according to the intervals to which they belong. The importance grading system database contains the importance level of each interval to which the differential data belongs. Differential data are matched according to the importance grading system database to obtain the corresponding importance assessment results; S23: Data monitoring terminals are set up according to the marking results in the three-dimensional landslide spatial model to which each key point belongs. The data monitoring terminals include multiple... The monitoring terminal for each data type acquires multi-dimensional real-time monitoring data at the corresponding locations of each key point through the data acquisition terminal. The multi-dimensional real-time monitoring data includes deformation monitoring data, stress monitoring data, hydrological monitoring data, and support monitoring data. It should be further explained that the process of assessing the importance of the differential data in step S22 is as follows: the comparison type of the parameter assignment processing results corresponding to the differential data is obtained. The comparison type corresponding to the parameter assignment processing results is the superposition result of the three-dimensional image of the landslide area to which the corresponding parameter assignment processing results belong and the corresponding basic information of the landslide, for example, the parameter assignment processing result of the corresponding sliding surface morphology in the topographic surface image.The corresponding importance weights are assigned to the parameter assignment results of the corresponding types. The differential data of each type within adjacent unit area grids are weighted according to their respective importance weights to obtain the corresponding importance assessment results. An importance assessment threshold is set. If the importance assessment result is greater than or equal to the importance assessment threshold, adjacent unit area grids are set as key points, and the set key points are mapped to the corresponding locations within the 3D landslide spatial model.

[0023] In the specific implementation process, the step S3, which involves setting the model parameters of the basic evaluation model based on the location information of each key point in the three-dimensional landslide spatial model, and dynamically adjusting the model parameters of the basic evaluation model based on multivariate real-time monitoring data to obtain the dynamic evaluation model, includes: S31: Obtaining multivariate historical monitoring data at the location information of each key point; based on the parameter assignment processing results at the corresponding locations within the three-dimensional landslide spatial model where each key point is located, setting the corresponding key point datasets using the multivariate historical monitoring data and parameter assignment processing results obtained at each key point; S32: Dividing the obtained key point datasets into multiple sub-datasets, and analyzing and processing each obtained sub-dataset separately. Set the corresponding basic evaluation model and model parameters of the basic evaluation model. The basic evaluation model includes three types: limit equilibrium correction model, data simulation dynamic model, and machine learning early warning model; S321: Obtain the sub-datasets divided by the key point dataset, and compare the parameter assignment processing results of the set sub-datasets with the corresponding multivariate historical monitoring data at the same time. Obtain the monitoring collection time corresponding to the multivariate historical monitoring data, and compare the parameter assignment processing results of each multivariate historical monitoring data at the same time according to the monitoring collection time. That is, compare the multivariate historical monitoring data and parameter assignment processing results at the same time at the corresponding position; S322: According to the set sub-datasets Based on the simultaneous comparison of multivariate historical monitoring data and parameter assignment processing results within the data set, the analysis was performed using the Mohr-Coulomb criterion to construct a corresponding limit equilibrium correction model and corresponding model parameters. It should be further explained that the process of constructing the limit equilibrium correction model and corresponding model parameters based on the Mohr-Coulomb criterion includes: obtaining multivariate historical monitoring data and parameter assignment processing results; dividing the multivariate historical monitoring data and parameter assignment processing results into regions to obtain sliding body regions and sliding zone regions; and assigning model parameters according to the set sliding body regions, sliding zone regions, and sliding bed regions respectively. It is assumed that the failure of the soil and rock mass within the landslide area is mainly caused by shear stress, and that the shear stress on the shear failure surface is linearly correlated with the normal stress. The following steps are performed: A yield function and flow rule are constructed to obtain failure criteria under three-dimensional stress states. Differential values ​​are assigned to the sliding body region, slip zone region, and sliding bed region based on these criteria. For the sliding body region and the sliding bed region, Poisson's ratio (μ), unit weight (γ), elastic modulus (E), and peak strength (c / φ) of each soil layer are input into the input layer. For the slip zone region, residual strength (cᵣ / φᵣ) is input into the input layer. The input layers in the sliding body region and the slip zone region are coupled to obtain the corresponding effective stress. Stress calculations are performed based on the effective stress, and the stress calculation results are compared with the corresponding failure criteria to obtain the failure determination results. The corresponding model parameters are then obtained based on the failure determination results.S323: Based on the multivariate historical monitoring data and the time-of-flight comparison results of the parameter assignment processing results within the set subset, data inversion processing is performed using FLAC3D to construct the corresponding data simulation dynamic model and corresponding model parameters corresponding to the multivariate historical monitoring data and parameter assignment processing results. It should be further explained that the FLAC3D data inversion process includes: performing parameter inversion on the multivariate historical monitoring data and parameter assignment processing results corresponding to the time-of-flight comparison processing results, and dividing them into inversion monitoring data and verification monitoring data for displacement inversion and seepage inversion respectively, wherein: displacement inversion: comparing the inversion monitoring data with the corresponding... The limit equilibrium correction model and the corresponding model parameters were compared and analyzed with the verification monitoring data. The elastic modulus (E), Poisson's ratio (μ), and residual strength (cᵣ / φᵣ) of the sliding body and sliding zone regions were corrected by the least squares method, with the goal of making the relative error e ≤ 10%. Seepage inversion: The groundwater level and pore water pressure in the inversion monitoring data were compared and analyzed with the verification monitoring data calculated by the model to obtain the consistency between the phreatic line and the actual monitoring. The comparison results of each set of inversion monitoring data and verification monitoring data were analyzed and processed, and the model parameters and boundary conditions were automatically updated to obtain the corresponding data simulation dynamic model and corresponding model parameters. S324: Based on the multivariate historical monitoring data and parameter assignment processing results within the set subset, the results of the same-time comparison processing of the parameters, and the output results of the limit equilibrium correction model and the data simulation dynamic model, analysis and processing are performed using machine learning algorithms. A mapping relationship between the corresponding multivariate historical monitoring data and stability state under the corresponding parameter assignment processing results is established, the risk level is output, and a machine learning early warning model and corresponding model parameters are constructed. It should be further explained that during the construction of the machine learning early warning model and corresponding model parameters: the multivariate historical monitoring data and parameter assignment processing results within the subset corresponding to the same-time comparison processing results are compared with the corresponding limit equilibrium correction model... The outputs of the positive model and the data simulation dynamic model are used to construct a dataset, which includes a training set, a validation set, and a test set. The obtained training set is then processed using a Long Short-Term Memory (LSTM) network. The corresponding multivariate historical monitoring data and parameter assignment results within the training set are used for sample classification, and corresponding sample labels are set based on the classification results. Feature statistics are performed based on the sample labels corresponding to the classification results within the training set, and corresponding feature data is obtained. This feature data includes statistical features, trend change features, and physical coupling features. The obtained feature data is then used to train the training set using the LTM network, and a loss function is set. : ,in, As a balance factor, The modulation coefficient and pt accuracy data are used. The loss function is iteratively validated based on the validation set, and the training is stopped and validated based on the set test set. The test results are obtained, and it is determined whether the test results meet the standard. If they do, the corresponding machine learning early warning model and the corresponding model parameters are output. S33: The parameter assignment processing results and multivariate historical monitoring data in the corresponding type of basic evaluation model subset are compared and correlated with the corresponding model parameters to obtain the comparison and correlation table between the corresponding parameter assignment processing results and multivariate historical monitoring data and the corresponding type of model parameters. It should be further explained that in the process of comparing and correlated with the parameter assignment processing results and multivariate historical monitoring data in the basic evaluation model subset and the corresponding model parameters, the parameter assignment processing results and multivariate historical monitoring data in the basic evaluation model subset are first set as independent variables X, and the model parameters are set as dependent variables Y. Then, single variable extraction and multivariate extraction are performed in sequence. The linear relationship strength between the independent variable and the dependent variable is calculated by the Pearson linear relationship strength. The calculation formula is r=cov(X,Y) / (σ_X). σ_Y), where cov(X,Y) represents the covariance of X and Y, and σ_X and σ_Y represent the standard deviations of X and Y, respectively. The absolute value of the linear relationship strength is greater than 0.8, indicating a strong correlation; 0.3-0.8 indicates a moderate correlation; and less than 0.3 indicates a weak correlation. Positive values ​​indicate a positive correlation, and negative values ​​indicate a negative correlation. The statistical significance of the correlation is determined by the significance level (p-value). A p-value less than 0.05 indicates a significant correlation at the 95% confidence level, and a p-value less than 0.01 indicates a significant correlation at the 99% confidence level. Based on the results of univariate and multivariate extraction, the system automatically outputs the linear relationship strength matrix and significance level results. For example, "the linear relationship strength between the parameter assignment processing result and the model parameters is 0.85 (p<0.01)" indicates a strong positive correlation and statistical significance, while "the linear relationship strength between the multivariate historical monitoring data and the model parameters is -0.62 (p<0.05)" indicates a moderate negative correlation and statistical significance.Based on these correlation analysis results, the system automatically generates correlation comparison data containing descriptions of linear relationship strength, significance level, and correlation strength, and then constructs a correlation table comparing the corresponding parameter assignment processing results and multivariate historical monitoring data with model parameters; S34: Obtain multivariate real-time monitoring data at corresponding key point locations within the three-dimensional landslide spatial model, dynamically adjust the reference assignment processing results at key point locations based on the obtained multivariate real-time monitoring data, and obtain real-time reference assignment processing results; S35: Compare and analyze the multivariate real-time monitoring data and real-time reference assignment processing results with the correlation table, and adjust the model parameters of the basic evaluation model based on the comparison analysis results. The data is dynamically adjusted to obtain a dynamic evaluation model; S351: Obtain the multivariate real-time monitoring data and real-time reference assignment processing results at the corresponding location, as well as the comparison and correlation table. The obtained multivariate real-time monitoring data and real-time parameter assignment processing results are retrieved and compared in the comparison and correlation table. Based on the retrieval and comparison analysis processing results, the model parameters corresponding to the multivariate real-time monitoring data and real-time parameter assignment processing results are obtained; S352: Based on the model parameters obtained from the retrieval and comparison analysis processing results, the model parameters corresponding to the current basic evaluation model are dynamically adjusted to obtain the dynamically adjusted model parameters. Based on the obtained model parameters, the corresponding dynamic evaluation model is obtained.

[0024] In the specific implementation process, the step S4, which involves setting the historical evaluation results of the dynamic evaluation models of each key point in the three-dimensional landslide spatial model to establish association trigger release conditions, connecting each key point according to the key trigger conditions, and setting the association map of the three-dimensional landslide key points, includes: S41: obtaining the historical evaluation results and historical evaluation time of the dynamic evaluation models corresponding to each key point in the three-dimensional landslide spatial model, and setting vertical time-series dynamic change multivariate curves for the dynamic evaluation models and historical evaluation results in the three-dimensional landslide spatial model according to the historical evaluation time. The vertical time-series dynamic change multivariate curve includes each dynamic evaluation model... S42: Connect the model parameters of the type and the dynamic changes of historical assessment results of each type at each historical assessment time; S42: Combine the vertical time-series dynamic change multivariate curves set at each key point in the three-dimensional landslide spatial model with elements based on the key points between adjacent or spaced positions, and obtain the data information in the corresponding vertical time-series dynamic change multivariate curves based on the element combination results to set the element combination dataset; S421: Obtain the vertical time-series dynamic change multivariate curves set at each key point in the three-dimensional landslide spatial model, and combine the obtained vertical time-series dynamic change multivariate curves with the key points and various multivariate real-time monitoring data and actual... The historical evaluation results of the time parameter assignment processing are set as the corresponding analysis elements, and the set analysis elements are marked on the vertical time-series dynamic change multivariate curve; S422: Based on the position information of each analysis element in the vertical time-series dynamic change multivariate curve and the marking results on the vertical time-series dynamic change multivariate curve, element cross-combination is performed. The element cross-combination process is to respectively combine different analysis elements on the vertical time-series dynamic change multivariate curve within the same key point, the same analysis elements on the corresponding vertical time-series dynamic change multivariate curve within adjacent key points, and different analysis elements on the corresponding vertical time-series dynamic change multivariate curve within adjacent key points. S43: Analyze the elements and combine them. Based on the element combination results, obtain the data information within the corresponding vertical time-series dynamic change multivariate curve and set up the element combination dataset; S44: Extract time-series features from the obtained element combination dataset according to the vertical time-series dynamic change multivariate curve to which it belongs. Based on the time-series feature extraction results, perform data alignment processing on the data information within the element combination dataset; S45: Based on the data alignment processing results, sequentially perform correlation relationship determination analysis, trigger relationship determination analysis, and dissolution relationship determination analysis to obtain the correlation, trigger, and dissolution condition relationships between different types of data information at various key points in the three-dimensional landslide spatial model.S441: During the correlation analysis of the data alignment results, statistical analysis is performed based on Pearson correlation analysis, Spearman rank correlation analysis, and mutual information entropy analysis to obtain quantitative correlation data between corresponding element combinations. Furthermore, based on the principles of geotechnical mechanics, creep law statistics are performed on the data alignment results within the element combination dataset to obtain mechanical correlation data. Logical verification analysis is then performed on the quantitative correlation data and the mechanical correlation data. The correlation relationship between the corresponding element combination datasets is obtained based on the logical verification analysis results. S442: During the trigger relationship analysis of the data alignment results, triggering basic constraint information is set. This triggering basic constraint information includes temporal constraints, threshold constraints, and response constraints. The element combination datasets with correlation relationships are analyzed and processed based on the triggering constraint information. Statistical analysis is performed on the data alignment results within the element combination datasets, and the corresponding triggering basic constraints are determined. The component information is adaptively adjusted, and the triggering relationship of the corresponding element combination dataset is obtained based on the adaptive adjustment result; S443: In the process of analyzing the removal relationship of the data alignment processing result, based on the corresponding association and triggering relationship within each element combination set, the basic constraint condition information for removal is set, which includes time-series constraint conditions, threshold constraint conditions, and response constraint conditions; the element combination dataset with association and triggering relationship is analyzed and processed according to the basic constraint condition information for removal, the data alignment processing result within the element combination dataset is statistically analyzed, the corresponding basic constraint condition information for removal is adaptively adjusted, and the removal relationship of the corresponding element combination dataset is obtained based on the adaptive adjustment result; S45: According to the association trigger removal condition relationship corresponding to each element combination, the connection processing is performed at each key point in the three-dimensional landslide spatial model to construct the three-dimensional landslide key point association map corresponding to the three-dimensional landslide spatial model.

[0025] In the specific implementation process, the step S5, which involves performing a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map, and obtaining the corresponding comprehensive stability assessment data of the landslide, includes: S51: Obtaining the multivariate real-time monitoring data at each key point within the three-dimensional landslide spatial model and the model parameters of the dynamic evaluation model; performing an initial evaluation on the obtained multivariate real-time monitoring data based on the dynamic evaluation model with the corresponding model parameters, and obtaining the initial evaluation results; S52: Verifying and adjusting the initial evaluation results obtained at each key point within the three-dimensional landslide spatial model according to the association triggering and release conditions between different data types within each key point in the three-dimensional landslide key point association map, and performing a comprehensive stability evaluation on the initial evaluation results at each key point based on the verification and adjustment analysis results, and obtaining the comprehensive stability assessment data of the landslide. Qualitative assessment data; S521: Verify and analyze the initial evaluation results at each key point location based on the association conditions between different data types of key points within the 3D landslide key point association map, and determine whether the corresponding initial evaluation results meet the corresponding association conditions. If they meet the corresponding association conditions, the verification is complete; if they do not meet the corresponding association conditions, the corresponding initial evaluation results do not meet the verification requirements; S522: Compare and match the assessment process corresponding to the initial evaluation results that do not meet the verification requirements with the triggering and deactivation conditions within the 3D landslide key point association map, and make comprehensive adjustments to the corresponding assessment process based on the comparison and matching results; S523: Adjust the evaluation process of the obtained multivariate real-time monitoring data based on the comprehensive adjustment results until the initial evaluation results are verified and output the corresponding comprehensive stability assessment data.

[0026] As shown in Figure 2, based on Embodiment 1, this invention proposes a creep-type landslide stability evaluation device, comprising a data management module, a data acquisition module, a data processing module, a data analysis module, and a comprehensive evaluation module. The specific implementation process includes: the data management module acquiring basic landslide information and setting up a three-dimensional landslide spatial model based on this information; the data acquisition module extracting key points from the set three-dimensional landslide spatial model, setting up a data monitoring terminal based on the key point extraction results, and acquiring multi-dimensional real-time monitoring data through the data monitoring terminal; and the data processing module processing data based on various parameters within the three-dimensional landslide spatial model. The location information of key points sets the model parameters of the basic evaluation model. Based on multivariate real-time monitoring data, the model parameters of the basic evaluation model are dynamically adjusted to obtain a dynamic evaluation model. The data analysis module is used to set association trigger and release conditions for the historical evaluation results of the dynamic evaluation model of each key point within the three-dimensional landslide spatial model. Based on the key trigger conditions, each key point is connected to create a three-dimensional landslide key point association map. The comprehensive evaluation module is used to perform a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map, obtaining the comprehensive stability assessment data of the corresponding landslide.

[0027] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for evaluating the stability of creep-type landslides, characterized in that, Includes the following steps: Step S1: Set up a landslide stability evaluation platform, obtain basic landslide information through the platform, and set up a three-dimensional landslide spatial model based on the basic landslide information; S2: Extract key points based on the set three-dimensional landslide spatial model, set up a data monitoring terminal based on the key point extraction results, and obtain multivariate real-time monitoring data through the data monitoring terminal; S3: Set up the model parameters of the basic evaluation model based on the location information of each key point in the three-dimensional landslide spatial model, and dynamically adjust the model parameters of the basic evaluation model based on the multivariate real-time monitoring data to obtain a dynamic evaluation model; S4: Set up association triggering and release conditions for the historical evaluation results of the dynamic evaluation model of each key point in the three-dimensional landslide spatial model, connect each key point according to the key triggering conditions, and set up a three-dimensional landslide key point association map; S5: Perform a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map to obtain the comprehensive stability evaluation data of the corresponding landslide.

2. The method for evaluating the stability of creep-type landslides according to claim 1, characterized in that, The process of setting up a three-dimensional landslide spatial model based on landslide basic information includes: setting up a landslide stability evaluation platform, which is equipped with an information input terminal to obtain landslide basic information, including landslide attribute information, geological environment information, and monitored deformation information; generating three types of three-dimensional images of the landslide area based on the obtained landslide basic information, including topographic surface images, stratum slip surface images, and boundary load images; performing adaptive meshing on the generated three-dimensional images of the landslide area to obtain unit area meshes within the three-dimensional images of the landslide area; and performing parameter assignment processing within each unit area mesh to generate a three-dimensional landslide spatial model.

3. The method for evaluating the stability of creep-type landslides according to claim 2, characterized in that, The process of acquiring multivariate real-time monitoring data through a data monitoring terminal includes: acquiring the parameter assignment processing results of each unit area grid within the three-dimensional landslide spatial model; comparing and analyzing the parameter assignment processing results of adjacent unit area grids to obtain difference data; assessing the importance of the difference data; setting corresponding key points based on the importance assessment results; marking the set key points within the three-dimensional landslide spatial model; setting up a data monitoring terminal based on the marking results within the three-dimensional landslide spatial model to which each key point belongs; the data monitoring terminal including monitoring terminals for multiple data types; acquiring multivariate real-time monitoring data at the corresponding locations of each key point through a data acquisition terminal; the multivariate real-time monitoring data including deformation monitoring data, stress monitoring data, hydrological monitoring data, and support monitoring data.

4. The method for evaluating the stability of creep-type landslides according to claim 3, characterized in that, The process of setting the model parameters of the basic evaluation model based on the location information of each key point in the three-dimensional landslide spatial model includes: acquiring multivariate historical monitoring data at the location information of each key point; based on the parameter assignment processing results at the corresponding locations in the three-dimensional landslide spatial model where each key point is located, setting the corresponding key point datasets by setting the multivariate historical monitoring data and parameter assignment processing results at each key point; dividing the obtained key point datasets into multiple sub-datasets; analyzing and processing each obtained sub-dataset; and setting the corresponding basic evaluation model and model parameters of the basic evaluation model. The basic evaluation model includes three types: limit equilibrium correction model, data simulation dynamic model, and machine learning early warning model.

5. The method for evaluating the stability of creep-type landslides according to claim 4, characterized in that, The process of obtaining the dynamic evaluation model includes: performing a comparative correlation analysis between the parameter assignment processing results and multivariate historical monitoring data within the corresponding type of basic evaluation model subset and the corresponding model parameters to obtain a correlation table between the corresponding parameter assignment processing results and multivariate historical monitoring data and the corresponding type of model parameters; obtaining multivariate real-time monitoring data at the corresponding key point locations within the three-dimensional landslide spatial model; dynamically adjusting the reference assignment processing results at the key point locations based on the obtained multivariate real-time monitoring data to obtain real-time reference assignment processing results; comparing and analyzing the multivariate real-time monitoring data and real-time reference assignment processing results with the correlation table; and dynamically adjusting the model parameters of the basic evaluation model based on the comparison analysis results to obtain the dynamic evaluation model.

6. The method for evaluating the stability of creep-type landslides according to claim 5, characterized in that, The process of setting up a three-dimensional landslide key point association map includes: acquiring the historical evaluation results and historical evaluation time of the corresponding dynamic evaluation model at each key point in the three-dimensional landslide spatial model; setting vertical time-series dynamic change multivariate curves for the dynamic evaluation model and historical evaluation results in the three-dimensional landslide spatial model according to the historical evaluation time; combining the vertical time-series dynamic change multivariate curves set at each key point in the three-dimensional landslide spatial model based on key points between adjacent or spaced locations; and obtaining the corresponding vertical time-series dynamic change multivariate curves based on the element combination results. The data information within the changing multivariate curve is set into an element combination dataset; the obtained element combination dataset is subjected to time series feature extraction according to the vertical time series dynamic changing multivariate curve to which it belongs; the data information within the element combination dataset is subjected to data alignment processing based on the time series feature extraction results; based on the data alignment processing results, correlation relationship determination analysis, trigger relationship determination analysis, and dissolution relationship determination analysis are performed in sequence to obtain the correlation, trigger, and dissolution condition relationships between different types of data information at each key point location in the 3D landslide spatial model; according to the correlation, trigger, and dissolution condition relationships corresponding to each element combination, connection processing is performed at each key point in the 3D landslide spatial model to construct the 3D landslide key point correlation map corresponding to the 3D landslide spatial model.

7. The method for evaluating the stability of creep-type landslides according to claim 6, characterized in that, The process of obtaining comprehensive stability assessment data for the corresponding landslide includes: acquiring multivariate real-time monitoring data and model parameters of the dynamic evaluation model at each key point within the three-dimensional landslide spatial model; performing an initial evaluation on the acquired multivariate real-time monitoring data based on the dynamic evaluation model with corresponding model parameters to obtain initial evaluation results; verifying and adjusting the initial evaluation results obtained at each key point within the three-dimensional landslide spatial model according to the association triggering and release conditions between different data types within each key point in the three-dimensional landslide key point association map; and performing a comprehensive stability assessment on the initial evaluation results at each key point based on the verification and adjustment analysis results to obtain comprehensive stability assessment data.

8. A creep-type landslide stability evaluation device corresponding to the creep-type landslide stability evaluation method according to any one of claims 1 to 7, characterized in that, The system includes a data management module, a data acquisition module, a data processing module, a data analysis module, and a comprehensive evaluation module. The data management module acquires basic landslide information and sets up a three-dimensional landslide spatial model based on this information. The data acquisition module extracts key points from the three-dimensional landslide spatial model, sets up a data monitoring terminal based on the extraction results, and acquires multivariate real-time monitoring data through the terminal. The data processing module sets the model parameters of the basic evaluation model based on the location information of each key point within the three-dimensional landslide spatial model, dynamically adjusts these parameters based on the multivariate real-time monitoring data, and obtains a dynamic evaluation model. The data analysis module sets associated trigger and release conditions for the historical evaluation results of the dynamic evaluation model for each key point within the three-dimensional landslide spatial model, connects the key points based on these trigger conditions, and sets up a three-dimensional landslide key point association map. The comprehensive evaluation module performs a comprehensive stability evaluation on the multivariate real-time monitoring data at each key point based on the obtained three-dimensional landslide key point association map, and obtains the corresponding comprehensive stability assessment data for the landslide.

Images