A method and system for staged classification and identification of main controlling factors of landslides in the Three Gorges Reservoir area

Abstract

This invention provides a method and system for the staged classification and identification of main controlling factors of landslides in the Three Gorges Reservoir area. The method includes: establishing an initial factor set based on the geological characteristics and monitoring data of the landslides; optimizing and screening the initial factor set twice to obtain a core factor set; studying the influence of each core factor on landslide deformation, establishing a multiple stepwise regression model considering time, rainfall, and reservoir water level, decomposing the landslide deformation process into components, listing the calculation formulas for each component, and deriving the final multiple stepwise regression model; analyzing landslide deformation curves at different time periods, identifying stage inflection points by combining evaluation indicators, and accurately classifying the deformation stages of the landslides; analyzing the primary and secondary influences of internal and external influencing factors on the landslides at different stages by comparing the fitted amplitude values ​​with the measured amplitude values, and determining the main controlling factors for each stage and the current stage based on the changes in the proportion of each component in each stage. This invention can identify the main controlling factors of landslides at different stages.

Description

Technical Field

[0001] This invention relates to the field of landslide monitoring data analysis technology, and in particular to a method and system for classifying landslide stages and identifying main controlling factors in the Three Gorges Reservoir area. Background Technology

[0002] Landslides in the Three Gorges Reservoir area are a significant geological hazard posing a serious threat to the safety of the reservoir region. Analysis of long-term monitoring data helps to understand the inherent patterns of landslide deformation. However, traditional landslide monitoring and analysis methods often rely on single monitoring indicators, failing to comprehensively and systematically consider the landslide evolution process and the combined influence of multiple factors. They neglect the dynamic changes during landslide evolution, fail to effectively capture the key time points of each stage of landslide deformation, and struggle to accurately identify the main controlling factors at each stage based on changes in the monitoring curves. Summary of the Invention

[0003] To address the technical problems existing in the prior art, this invention provides a method and system for phased classification and identification of main controlling factors of landslides in the Three Gorges Reservoir area. The technical solution is as follows:

[0004] On the one hand, a method for the staged division of landslides and identification of main controlling factors in the Three Gorges Reservoir area is provided, which includes:

[0005] S1. Collect and preprocess landslide monitoring data in the Three Gorges Reservoir area;

[0006] S2. Based on the geological characteristics and monitoring data of the landslide, preliminary screening of factors that may affect the stability of the landslide was conducted, including rainfall and reservoir water level. An initial factor set of external influencing factors of the landslide was established around these two factors.

[0007] S3. Through the significance level index Sig and Pearson correlation coefficient analysis, the initial factor set was optimized twice to select the core factor set that best reflects the changes in landslide stability.

[0008] S4. Based on the optimized core factor set and landslide deformation data, separate and study the influence of each core factor on landslide deformation, establish a multivariate stepwise regression model that considers time, rainfall and reservoir water level, decompose the landslide deformation process into each component, list the calculation formulas of each component in turn, and obtain the final multivariate stepwise regression model.

[0009] S5. Based on the regression model fitting results, by analyzing the landslide deformation curves at different time periods and combining evaluation indicators, the turning points of the stages are identified, and the deformation stages of the landslide are accurately divided.

[0010] S6. By comparing the fitted amplitude value with the measured amplitude value, analyze the primary and secondary influences of internal and external influencing factors on the landslide at different stages, and determine the main controlling factors for each stage and the current stage based on the changes in the proportion of each component in each stage.

[0011] On the other hand, a system for phased classification and identification of main controlling factors of landslides in the Three Gorges Reservoir area is provided, the system comprising:

[0012] The data collection and preprocessing module is used to collect and preprocess landslide monitoring data in the Three Gorges Reservoir area;

[0013] The first module is used to preliminarily screen factors that may affect landslide stability based on the geological characteristics and monitoring data of the landslide, including rainfall and reservoir water level, and to establish an initial factor set of external influencing factors of the landslide around the two factors of rainfall and reservoir water level.

[0014] The screening module is used to optimize the initial factor set twice through significance level index Sig and Pearson correlation coefficient analysis to screen out the core factor set that best reflects the changes in landslide stability.

[0015] The second module is used to separate and study the influence of each core factor on landslide deformation based on the optimized core factor set and landslide deformation data, establish a multivariate stepwise regression model that considers time, rainfall and reservoir water level, decompose the landslide deformation process into each component, list the calculation formulas of each component in turn, and obtain the final multivariate stepwise regression model.

[0016] The segmentation module is used to accurately segment the deformation stages of landslides by analyzing the landslide deformation curves at different time periods based on the regression model fitting results and combining evaluation indicators to identify the stage inflection points.

[0017] The determination module is used to analyze the primary and secondary influences of internal and external influencing factors on landslides at different stages by comparing the fitted amplitude values ​​with the measured amplitude values. Based on the changes in the proportion of each component in each stage, the main controlling factors for each stage and the current stage are determined.

[0018] On the other hand, an electronic device is provided, which includes a processor and a memory, wherein the memory stores at least one instruction, which is loaded and executed by the processor to realize the above-mentioned method for staged division and identification of main control factors of landslides in the Three Gorges Reservoir area.

[0019] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, and the at least one instruction is loaded and executed by a processor to implement the above-mentioned method for the staged division and identification of the main controlling factors of landslides in the Three Gorges Reservoir area.

[0020] The beneficial effects of the technical solution provided by this invention include at least the following:

[0021] 1. Dynamic Identification: This invention utilizes a multivariate stepwise regression model to achieve phased classification of landslide deformation, accurately capturing key time points of external factor transformation. At different stages of landslide deformation, external triggering factors (such as rainfall and reservoir water levels) have varying degrees of impact on the landslide. By dynamically identifying these influencing factors, this invention can adjust monitoring priorities in real time, improving the response speed of the landslide early warning system.

[0022] 2. Improved Accuracy: This invention employs a multivariate stepwise regression analysis method, combining various monitoring data such as displacement, rainfall, and reservoir water level, to establish a comprehensive and highly accurate landslide deformation assessment model. Through regression analysis, the model's multiple correlation coefficient (R) can reach above 0.98, and the residual standard deviation (S) is less than 1 mm, indicating that this invention has high accuracy in predicting and warning of landslide deformation.

[0023] 3. Optimized Prevention and Control: This invention can accurately divide the deformation stages of landslides and adjust monitoring priorities according to changes in external factors. For example, in the early stages of a landslide, monitoring influencing factors such as reservoir water levels and rainfall is conducted; while in the rainfall-dominated stage, monitoring of rainfall warnings is strengthened. This optimization scheme based on multi-factor analysis and stage division can provide a more scientific basis for landslide disaster prevention and control, reducing the risk of potential disasters. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of a method for phased division and identification of main controlling factors of landslides in the Three Gorges Reservoir area provided by an embodiment of the present invention;

[0026] Figure 2 This is a flowchart of the impact factor establishment process provided in an embodiment of the present invention;

[0027] Figure 3 This is a flowchart of the multivariate stepwise regression model establishment provided in the embodiments of the present invention;

[0028] Figure 4 This is a landslide stage division and external influencing factor primary and secondary analysis diagram provided in the embodiments of the present invention;

[0029] Figure 5 This is a block diagram of a system for phased division and identification of main controlling factors of landslides in the Three Gorges Reservoir area provided by an embodiment of the present invention;

[0030] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0031] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0032] This invention provides a method for analyzing the stage division and main controlling factors of landslides in the Three Gorges Reservoir area based on multi-spectral dual-branch deep fusion of flotation foam color. This method can be implemented by an electronic device, which can be a terminal or a server. Figure 1 The diagram shown is a flowchart of the method. The processing flow may include the following steps:

[0033] S1. Collect and preprocess monitoring data of landslides in the Three Gorges Reservoir area (taking Yemaomian landslide as an example);

[0034] In this embodiment of the invention, monitoring equipment (such as displacement gauges, rain gauges, vibration sensors, etc.) installed in the landslide area is used to collect data such as the deformation of the landslide body, rainfall, reservoir water level, and seismic acceleration in real time.

[0035] Data preprocessing in this embodiment of the invention: The collected raw data is cleaned, imputed, and standardized to remove outliers and determine data, thereby ensuring data quality for subsequent analysis.

[0036] S2. Based on the geological characteristics and monitoring data of the landslide, preliminary screening of factors that may affect the stability of the landslide was conducted, including rainfall and reservoir water level. An initial factor set of external influencing factors of the landslide was established around these two factors.

[0037] Optionally, S2 specifically includes:

[0038] (1) Rainfall factor

[0039] Landslide deformation exhibits a certain lag in its response to rainfall. The rainfall factor is the cumulative rainfall factor and the maximum effective rainfall factor within the previous 30 days, specifically defined as follows:

[0040]

[0041] In the formula: P N p represents the cumulative rainfall over the previous N days. i P represents the cumulative daily rainfall up to day i; N represents the maximum number of days of cumulative rainfall considered; α＝aThe maximum effective continuous rainfall within the previous 30 days; α is the effective rainfall attenuation coefficient, where the value of a ranges from (0,1] (in this embodiment, a = 0.6, 0.7, 0.8, 0.9, 1.0); n is the duration of the maximum continuous rainfall process within the previous 30 days;

[0042] (2) Reservoir water level factors

[0043] The impact of reservoir water level fluctuations on landslide deformation, like rainfall, exhibits a lag. In addition to considering the effect of the reservoir water level fluctuation factor, the influence of the current reservoir water level elevation on the landslide deformation response is also taken into account. The specific definitions of the reservoir water level fluctuation factor and the reservoir water level elevation factor are as follows:

[0044]

[0045] In the formula: The average reservoir water level elevation over the previous N days; N is the number of days the reservoir water level is considered to be affected, N = 10, 15, 30, 60; r i The reservoir water level elevation is the level of the day before day i. This represents the increase in reservoir water level over the previous N days. ΔR represents the rate of decrease in reservoir water level over the previous N days. N Δ|R represents the relative change in reservoir water level over the previous N days; N | represents the absolute change in reservoir water level over the previous N days;

[0046] The initial factor set for external influencing factors of landslides is as follows:

[0047] Rainfall:

[0048] X1: Cumulative rainfall P3 within the previous 3 days; X2: Cumulative rainfall P5 within the previous 5 days; X3: Cumulative rainfall P7 within the previous 7 days; X4: Cumulative rainfall P within the previous 15 days 15 X5: Cumulative rainfall P in the previous 30 days 30 X6: Cumulative rainfall P within the previous 45 days 45 X7: Cumulative rainfall P within the previous 60 days 60 X8: Maximum consecutive rainfall P within the previous 30 days α＝0.6 X9: Maximum consecutive rainfall P within the previous 30 days α＝0.7 ;X 10 Maximum consecutive rainfall P within the previous 30 days α＝0.8 ;X 11 Maximum consecutive rainfall P within the previous 30 days α＝0.9 ;X 12 Maximum consecutive rainfall P within the previous 30 days α＝1.0 ;

[0049] Reservoir water level:

[0050] X13 The increase in reservoir water level over the past 10 days X 14 The increase in reservoir water level over the past 15 days X 15 The increase in reservoir water level over the first 30 days X 16 The increase in reservoir water level over the past 60 days X 17 The percentage decrease in reservoir water level over the first 10 days X 18 The percentage decrease in reservoir water level over the first 15 days X 19 The percentage decrease in reservoir water level over the first 30 days X 20 The percentage decrease in reservoir water level over the first 60 days X 21 The relative fluctuation range of the reservoir water level in the first 10 days ΔR 10 ;X 22 The relative fluctuation range of the reservoir water level in the first 15 days ΔR 15 ;X 23 The relative fluctuation range of reservoir water level ΔR over the first 30 days 30 ;X 24 The relative fluctuation range of reservoir water level ΔR over the previous 60 days 60 ;X 25 The absolute change in reservoir water level over the previous 10 days Δ|R 10 |;X 26 The absolute change in reservoir water level over the previous 15 days Δ|R 15 |;X 27 The absolute change in reservoir water level Δ|R over the previous 30 days 30 |;X 28 The absolute variation amplitude of the reservoir water level in the previous 60 days Δ|R 60 |;X 29 Average water level elevation of the reservoir over the previous 10 days X 30 Average water level of the reservoir over the previous 15 days X 31 Average reservoir water level elevation over the previous 30 days X 32 Average reservoir water level elevation over the previous 60 days As shown in Table 1:

[0051] Table 1 Initial Factor Set of External Influencing Factors of Landslides

[0052]

[0053] S3. Through the significance level index Sig and Pearson correlation coefficient analysis, the initial factor set was optimized twice to select the core factor set that best reflects the changes in landslide stability.

[0054] The initial factor set typically contains many factors that are highly correlated, leading to redundancy within the factor set. Therefore, this embodiment of the invention optimizes the established initial factor set to reasonably reduce its complexity and eliminate the impact of multicollinearity among factors on the model analysis results.

[0055] Factor set optimization utilizes significance level index (Sig) and Pearson correlation coefficient (P) analysis to screen feature factors. The significance level index indicates whether a relationship exists between factors, while the correlation coefficient indicates the strength of that relationship. A Sig value less than 0.05 indicates a 95% probability of a relationship between factors; that is, Sig < 0.05 indicates a significant relationship. A Pearson correlation coefficient above 0.7 indicates a very strong relationship; between 0.4 and 0.7 indicates a strong relationship; and between 0.2 and 0.4 indicates a moderate relationship.

[0056] Optionally, S3 specifically includes:

[0057] Preliminary optimization: According to different factor categories, calculate the significance level index Sig and Pearson correlation coefficient between each factor in the initial factor set and the deformation event. Based on the calculation results, sort the factors of each type in the factor set according to the magnitude of the correlation coefficient between each factor and the deformation event, and remove factors with low correlation with the deformation event, thereby achieving the initial optimization of the candidate factor set.

[0058] Secondary optimization: Calculate the Pearson correlation coefficient and significance level index Sig among the candidate factors after the initial optimization, and further screen them according to the correlation between each factor to obtain the final core factor set.

[0059] This invention optimizes the initial factor set based on displacement time series data from the TP06 monitoring point. This monitoring point is located in the front part of the landslide body, has the longest monitoring time, and provides complete and representative monitoring data. Following the optimization method described above, rainfall-related factors and reservoir water level factors are sequentially optimized and screened.

[0060] (1) Optimization process of rainfall-type factors

[0061] Table 2 shows the Pearson correlation coefficients and significance levels between each rainfall-type factor and deformation event. The significance levels of all 12 rainfall-type factors and deformation events are less than 0.01, indicating significant correlations. The correlation coefficients of X1, X2, and X3 with deformation events are all less than 0.2, indicating weak correlations, and these three factors can be excluded. The correlation coefficients of the other 9 rainfall-type factors with deformation events are all greater than 0.2, indicating strong correlations. The factors were ranked according to their correlation coefficients, and the ranking result was: X7 > X... 12 =X6=X 11 >X 10 =X5=X9>X8>X4, meaning that among the nine rainfall-related factors after the initial optimization, the cumulative rainfall X7 within the first 60 days has the highest correlation with the deformation event.

[0062] Table 2. Pearson correlation coefficients between rainfall-type factors and deformation events.

[0063]

[0064] After the initial optimization of the rainfall factors, nine rainfall-related factors remain, which increases the difficulty of establishing the subsequent regression model and affects the effectiveness and accuracy of the final regression model. Therefore, this embodiment of the invention analyzes the correlation between the remaining nine factors, selecting those with a high correlation to landslide deformation and no redundant information with other factors, thus achieving a secondary optimization of the factor set. After the initial optimization, the Sig values ​​of all rainfall factors are less than 0.01, indicating a significant correlation. The Pearson correlation coefficients are shown in Table 3. X7 and X 12 X7 had the highest correlation coefficient and was also highly correlated with the other seven preliminarily optimized factors. Therefore, in the optimization process of the rainfall-related factors, X7 and X... 12 These two factors represent the optimization results of rainfall-related factors.

[0065] Table 3. Pearson correlation coefficients among rainfall-related factors after initial optimization.

[0066]

[0067] (2) Optimization process of reservoir water type factors

[0068] Table 4 shows the Pearson correlation coefficients and significance levels between various reservoir water type factors and deformation events. From the significance level index, X... 17 X 18 X 19The Sig values ​​of the three factors related to the deformation event were all greater than 0.05, indicating a clear lack of correlation, and these three factors can be eliminated. The Sig values ​​of the other 17 reservoir-type factors related to the deformation event were all less than 0.01, showing a significant correlation. From the Pearson correlation coefficient, X... 13 X 20 X 14 X 15 X 16 The p-values ​​were all less than 0.15, indicating no significant correlation; therefore, these five factors can be removed. 21 X 22 X 25 X 26 X 27 The correlation coefficients of these five reservoir water factors ranged from 0.15 to 0.2, indicating a moderate correlation; X 23 X 24 X 28 The correlation coefficient of these three factors is greater than 0.2, indicating a strong correlation; X 29 X 30 X 31 X 32 The correlation coefficients of these four factors are all greater than 0.45, indicating a strong correlation. The factors are then sorted according to their correlation coefficients, resulting in the following ranking: X 32 >X 31 =X 30 =X 29 >X 24 >X 23 >X 28 >X 22 >X 27 >X 21 >X 26 =X 25 That is, among the 12 reservoir water type factors after initial optimization, the average reservoir water level elevation X over the first 60 days. 32 It has the highest correlation with deformation events.

[0069] Table 4. Pearson correlation coefficients between reservoir water type factors and deformation events.

[0070]

[0071] After the initial optimization, the Sig values ​​of all reservoir water factors were less than 0.01, and the Pearson correlation coefficients are shown in Table 5. Combining the Pearson correlation coefficients among the reservoir water factors and the ranking results after the initial optimization, X... 32 X 30 X 24 X 22 These four factors are relatively independent of each other, and there are significant correlations between them and other factors, so X is retained.32 X 30 X 24 X 22 The four factors are the final optimization results of the reservoir water type factors.

[0072] Table 5. Pearson values ​​of each factor after initial optimization of reservoir water type factors.

[0073]

[0074] Optionally, the final core factor set includes: X7, X 12 X 32 X 30 X 24 X 22 .

[0075] The flowchart for establishing the influence factor in this invention is as follows: Figure 2 As shown.

[0076] S4. Based on the optimized core factor set and landslide deformation data, the influence of each core factor on landslide deformation is separated and studied. A multiple stepwise regression model considering time, rainfall, and reservoir water level is established. The landslide deformation process is decomposed into components, and the calculation formulas for each component are listed in sequence. The final multiple stepwise regression model is then obtained, as shown below. Figure 3 As shown;

[0077] Optionally, S4 specifically includes:

[0078] S41. Establish a multiple stepwise regression model that considers time, reservoir water level, and rainfall, decomposing the landslide deformation process into its components. The multiple stepwise regression model is expressed as follows:

[0079] δ=δ θ +δ P +δ R +D

[0080] In the formula: δ is the deformation; δ θ The deformation component caused by aging; δ P The deformation component caused by changes in rainfall; δ R D represents the deformation component caused by reservoir water level fluctuations; D is a statistical constant.

[0081] S42. The rainfall and reservoir water level data are processed by using [0,1] linear normalization to map data from different ranges to a unified interval, so as to ensure consistent data scale and improve model stability.

[0082] S43. Decompose the landslide deformation process into components and list the calculation formulas for each component in turn.

[0083] (1) The time-dependent component δ θ

[0084] The deformation component caused by aging is expressed as:

[0085] δ θ =a1θ+a2lnθ

[0086] In the formula: θ is the cumulative number of days from the displacement observation date to the initial measurement date divided by 100; a1 and a2 are the regression coefficients of the deformation components caused by time-related changes, which can be obtained by regression analysis;

[0087] (2) Rainfall component δ P

[0088] Based on the optimization results of the rainfall pattern factor, the cumulative rainfall P within the previous 60 days was used. 60 and the maximum continuous rainfall P in the previous 30 days α＝1.0 The formula for calculating rainfall components is as follows:

[0089] δ P =b1P 60 +b2P α＝1.0

[0090] In the formula: P 60 This refers to the cumulative rainfall over the previous 60 days; P α＝1.0 The maximum continuous rainfall within the previous 30 days, with a rainfall attenuation coefficient α = 1.0; b i The regression coefficients for the deformation components caused by changes in rainfall can be obtained through regression analysis.

[0091] (3) δ component of reservoir water level R

[0092] Based on the optimization results of the reservoir water type factor, the formula for calculating the reservoir water level component is as follows:

[0093]

[0094] In the formula: These are the average reservoir water level elevations over the previous 60 days and 15 days, respectively; ΔR 60 ,ΔR 15 These represent the relative fluctuations in reservoir water levels over the previous 60 days and 15 days, respectively; c i The regression coefficients of the deformation components caused by reservoir water level fluctuations are i = 1, 2, 3, 4;

[0095] S44. The final multiple stepwise regression model is as follows:

[0096]

[0097] Optionally, the evaluation indicators include: multiple correlation coefficient R, residual standard deviation S, and F value;

[0098] The multiple correlation coefficient R is an indicator that measures the degree of linear correlation between a variable and several other variables. Its value ranges from 0 to 1. The closer the absolute value is to 1, the stronger the linear relationship between the two variables; the closer it is to 0, the weaker the linear relationship. In regression analysis, the focus is on the correlation coefficient between predicted displacement and observed displacement. The closer it is to 1, the better the model fit. It measures the correlation between a variable y and several other variables x1, x2, ..., xn. k The correlation coefficient between them requires constructing a coefficient for x1, x2, ..., x k The linear combination of x1, x2, ..., x2 is used to calculate the simple correlation coefficient between this linear combination and y as the variable y and x1, x2, ..., x2. k The specific steps for determining the multiple correlation coefficient between them are as follows:

[0099] Let y be the ratio of x1, x2, ..., x k By regression, we get:

[0100]

[0101] Calculate the simple correlation coefficient, as y versus x1, x2, ..., x k The multiple correlation coefficient between them

[0102] The formula for calculating the multiple correlation coefficient is:

[0103]

[0104] The residual standard deviation (S) is a metric for measuring model fitting error. It represents the standard deviation of the difference between observed and predicted values. The smaller the residual standard deviation, the smaller the model fitting error. Its calculation formula is:

[0105]

[0106] In the formula: y i These are the observed values; n is the predicted value; n is the number of samples; p is the number of model parameters;

[0107] The F-value is used to test the overall significance of a regression model, that is, whether the overall effect of all independent variables on the dependent variable is significant. The larger the F-value, the stronger the explanatory power of the independent variables on the dependent variable. The calculation formula is:

[0108]

[0109] In the formula: p is the mean of the dependent variable; n is the number of model parameters; and n is the number of samples.

[0110] S5. Based on the regression model fitting results, by analyzing the landslide deformation curves at different time periods and combining evaluation indicators, the turning points of the stages are identified, and the deformation stages of the landslide are accurately divided.

[0111] Optionally, the division of deformation stages includes:

[0112] Phase 1: Initial deformation phase of the landslide, from June 2003 to September 2006;

[0113] The second phase: the sensitive period affected by reservoir water levels, from October 2006 to September 2018;

[0114] The third stage: the landslide stabilization and deformation stage, which lasts from October 2018 to February 2024;

[0115] Current state (March 2025): Based on the deformation trend, it is inferred that the current stage is the third stage.

[0116] S6. By comparing the fitted amplitude value with the measured amplitude value, analyze the primary and secondary influences of internal and external influencing factors on the landslide at different stages, and determine the main controlling factors for each stage and the current stage based on the changes in the proportion of each component in each stage.

[0117] Optionally, S6 specifically includes:

[0118] In the first stage, the time-dependent component accounts for the main part of the deformation. The influence of reservoir water level on the landslide front edge is significantly greater than that of rainfall. According to the data from the initial stage, the landslide was generally in a stable state before water impoundment. The first phase of experimental water impoundment led to a significant rise in the water level in front of the slope, resulting in significant physical and mechanical effects, which caused the landslide to revive through front edge traction deformation.

[0119] In the second stage, the time-dependent component of landslide deformation is the main component. Compared with the initial deformation stage, the proportion of the time-dependent component in this stage has increased significantly. The main external influencing factor is the fluctuation of reservoir water level, while the proportion of rainfall component is relatively small. The deformation mode of the landslide changes from the leading edge traction in the early stage of water storage to overall creep.

[0120] In the third stage, the deformation pattern of the landslide did not change substantially, but its overall creep rate decreased significantly. This indicates that the deterioration effect caused by large-scale water-rock interaction tended to weaken. After years of reservoir water level fluctuations, the landslide gradually adapted to the changes in the groundwater environment caused by reservoir water level fluctuations. The time-dependent component was the main deformation component. Compared with the second stage, the proportion of the time-dependent component in this stage was roughly the same. This indicates that the deterioration caused by water-rock interaction was basically completed in the second stage. The creep rate of the landslide further decreased compared with the previous stage, and the landslide deformation trend slowed down. The external influencing factor changed from reservoir water level fluctuations in the second stage to seasonal rainfall. Comparing the proportions of rainfall and reservoir water components, it can be seen that at this time, the main external influencing factor changed from reservoir water level fluctuations to rainfall.

[0121] Currently, we are in the third stage. The main controlling factor, which is the most important external influencing factor of landslides, has changed from reservoir water level to rainfall. This means that the focus of landslide monitoring and prevention should shift to the forecasting of heavy rainfall and the early warning and prevention of disasters that may be caused by it.

[0122] The landslide stage division and the primary and secondary analysis of external influencing factors in this invention embodiment are shown in the following diagram. Figure 4 As shown, more specifically:

[0123] 1. Initial deformation stage of landslide

[0124] The first stage is the initial deformation stage, which lasted from June 2003 to September 2006. Table 6 shows the regression model coefficients for this stage. Except for TP07, the multiple correlation coefficients R for other monitoring points were all higher than 0.98, the residual standard deviations S were all less than 1, and the F values ​​were all much greater than 10. Therefore, the regression model in the first stage has high accuracy.

[0125] Table 6. Regression coefficients of the regression model in the first stage (June 2003 to September 2006)

[0126]

[0127]

[0128] Table 7 shows that in the first stage, the landslide was in the initial deformation phase. The deformation at five monitoring points—AL01 and AL02 at the rear edge, AL03 and AL04 at the middle, and TP06 at the leading edge—was too small, and the amplitude and proportion of the three components were not representative. The amplitude values ​​at the leading edge monitoring points TP07, TP08, and TP09 ranged from 4.5 to 11 mm, with the time-dependent component accounting for the majority of the deformation. The influence of reservoir water level on the leading edge of the landslide was significantly greater than that of rainfall. The data from the initial stage indicate that the Yemaomian landslide was generally stable before water impoundment. The first phase of experimental water impoundment led to a significant rise in the water level at the front of the slope, resulting in significant physical and mechanical effects and causing the landslide to revive through leading edge deformation.

[0129] Table 7. Amplitude values ​​and model separation results for the first stage (June 2003 to September 2006)

[0130]

[0131] 2. Sensitive periods affected by reservoir water levels

[0132] After multiple regression calculations, the sensitive period for the impact of reservoir water level was determined to be from October 2006 to September 2018. During this period, the landslide deformation was highly sensitive to the impact of reservoir water level, which can better reflect the complete cycle of reservoir water level fluctuations affecting the deformation of the Yemaomian landslide. The regression model calculations are similar to those in the first stage, and the tables are omitted here.

[0133] During this period, from October 2006 to September 2008, the reservoir water level fluctuated between 145 and 156 m, with a fluctuation range of 11 m. Experimental impoundment to 175 m began on September 28, 2008, after which the reservoir water level fluctuated between 145 and 175 m, with a fluctuation range of 30 m. Table 8 shows that the time-dependent component constitutes the majority of the landslide deformation, and its proportion has significantly increased compared to the initial deformation stage. AL02, located on the eastern side of the rear edge of the landslide, is furthest from the bank and is less affected by reservoir water level fluctuations. For the other seven monitoring points, the main external influencing factor is reservoir water level fluctuation, with rainfall accounting for a relatively small proportion. Therefore, this stage can be considered a sensitive stage affected by reservoir water level fluctuations. During this stage, except for AL02, the deformation at the other monitoring points was roughly the same, indicating that the deformation mode of the Yemaomian landslide changed from leading-edge traction in the initial impoundment stage to overall creep.

[0134] Table 8. Amplitude values ​​and model separation results for the second stage (October 2006 to September 2018)

[0135]

[0136] 3. Landslide stabilization and deformation stage

[0137] The third stage, from October 2018 to February 2024, saw the reservoir water level fluctuate between 145 and 175 meters. Table 9 shows that the landslide's deformation pattern did not change substantially during this stage, but its overall creep rate decreased significantly. This indicates that the deterioration effect caused by large-scale water-rock interaction tended to weaken, and after years of reservoir water level fluctuations, the Yemaomian landslide gradually adapted to the changes in the groundwater environment caused by these fluctuations.

[0138] Table 9. Amplitude values ​​and model separation results for the third stage (October 2018 to February 2024)

[0139]

[0140] Table 9 also shows that the time-dependent component is the main deformation component, and its proportion is roughly the same as that of the second stage. This indicates that the deterioration caused by water-rock interaction was basically completed in the second stage, and the creep rate of the Yemaomian landslide further decreased compared to the previous stage, with the landslide deformation trend slowing down. Except for AL01, AL02, and AL03, the external influencing factor at the other five monitoring points changed from reservoir water level fluctuations in the second stage to seasonal rainfall. Comparing the proportions of rainfall and reservoir water components, it can be seen that at this time, the main external influencing factor had just changed from reservoir water level fluctuations to rainfall, and the proportion of rainfall was only slightly larger than that of reservoir water. AL01 and AL03 are located at the rear edge and middle of the landslide body, and their influence on reservoir water level fluctuations has a certain lag. Considering all stages, AL02 was less affected by reservoir water level fluctuations than by rainfall, and its deformation was the smallest among the eight monitoring points. Similarly, AL05, located on the east side of the landslide, had a deformation close to zero. Therefore, it can be inferred that the main deformation area of ​​the Yemaomian landslide is on the west side of Baishigou, while the east side is currently in a stable state.

[0141] The transformation relationship of the landslide body under the influence of external factors in the three deformation stages is as follows:

[0142] The influence of reservoir water level has gradually weakened due to the attenuation of the deterioration effect caused by water-rock interaction, and its current impact on landslide deformation is more manifested as a mechanical effect. Meanwhile, the triggering effect of rainfall on landslide deformation is gradually becoming more prominent due to the weakening of the reservoir water level effect. Considering that reservoir water level has a long-term effect on landslides, while rainfall is a short-term, intermittent disturbance, the objective difference in the duration of these two influencing factors puts rainfall at a disadvantage in comparison. This further makes the proportion of rainfall components exceed that of reservoir water level components, which is an important indicative event indicating a stage transformation in the landslide deformation development process. Currently, the most important external influencing factor for the Yemaomian landslide has shifted from reservoir water level to rainfall. This means that the focus of landslide monitoring and prevention should shift to the forecasting of heavy rainfall and the early warning and prevention of potential disasters, which is crucial in the current context of frequent extreme weather events.

[0143] like Figure 5 As shown in the figure, this invention also provides a system for the staged division and main controlling factor identification of landslides in the Three Gorges Reservoir area, the system comprising:

[0144] The preprocessing module 510 is used to collect and preprocess landslide monitoring data in the Three Gorges Reservoir area.

[0145] The first module 520 is used to preliminarily screen factors that may affect landslide stability based on the geological characteristics and monitoring data of the landslide, including rainfall and reservoir water level, and to establish an initial factor set of external influencing factors of the landslide around the two factors of rainfall and reservoir water level.

[0146] The screening module 530 is used to optimize the initial factor set twice through the significance level index Sig and Pearson correlation coefficient analysis to screen out the core factor set that best reflects the changes in landslide stability.

[0147] The second module 540 is used to separate and study the influence of each core factor on landslide deformation based on the optimized core factor set and landslide deformation data, establish a multivariate stepwise regression model that considers time, rainfall and reservoir water level, decompose the landslide deformation process into each component, list the calculation formulas of each component in turn, and obtain the final multivariate stepwise regression model.

[0148] The segmentation module 550 is used to accurately segment the deformation stages of landslides by analyzing the landslide deformation curves at different time periods based on the regression model fitting results and combining evaluation indicators to identify the stage inflection points.

[0149] The determination module 560 is used to analyze the primary and secondary influences of internal and external influencing factors on landslides at different stages by comparing the fitted amplitude values ​​with the measured amplitude values. Based on the changes in the proportion of each component in each stage, the main controlling factors for each stage and the current stage are determined. The system for stage division and main controlling factor identification of landslides in the Three Gorges Reservoir area provided in this embodiment of the invention has a functional structure corresponding to the method for stage division and main controlling factor identification of landslides in the Three Gorges Reservoir area provided in this embodiment of the invention, and will not be described again here.

[0150] Figure 6 This is a schematic diagram of the structure of an electronic device 600 provided in an embodiment of the present invention. The electronic device 600 may vary considerably due to different configurations or performance. It may include one or more central processing units (CPUs) 601 and one or more memories 602. The memory 602 stores at least one instruction, which is loaded and executed by the processor 601 to implement the steps of the above-mentioned method for staged division and identification of main control factors of landslides in the Three Gorges Reservoir area.

[0151] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including instructions that can be executed by a processor in a terminal to complete the aforementioned method for phased division and identification of controlling factors of landslides in the Three Gorges Reservoir area. For example, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device.

[0152] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0153] 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 within the protection scope of the present invention.

Claims

1. A method for staged classification and identification of main controlling factors of landslides in the Three Gorges Reservoir area, characterized in that, The method includes: S1. Collect and preprocess landslide monitoring data in the Three Gorges Reservoir area; S2. Based on the geological characteristics and monitoring data of the landslide, preliminary screening of factors that may affect the stability of the landslide was conducted, including rainfall and reservoir water level. An initial factor set of external influencing factors of the landslide was established around these two factors. S3. Through the significance level index Sig and Pearson correlation coefficient analysis, the initial factor set was optimized twice to select the core factor set that best reflects the changes in landslide stability. S4. Based on the optimized core factor set and landslide deformation data, separate and study the influence of each core factor on landslide deformation, establish a multivariate stepwise regression model that considers time, rainfall and reservoir water level, decompose the landslide deformation process into each component, list the calculation formulas of each component in turn, and obtain the final multivariate stepwise regression model. S5. Based on the regression model fitting results, by analyzing the landslide deformation curves at different time periods and combining evaluation indicators, the turning points of the stages are identified, and the deformation stages of the landslide are accurately divided. S6. By comparing the fitted amplitude value with the measured amplitude value, analyze the primary and secondary influences of internal and external influencing factors on the landslide at different stages, and determine the main controlling factors at each stage and the current stage based on the changes in the proportion of each component in each stage. S2 specifically includes: (1) Rainfall factors Landslide deformation exhibits a certain lag in its response to rainfall. The rainfall factor is the cumulative rainfall factor and the maximum effective rainfall factor within the previous 30 days, specifically defined as follows: ； ； In the formula: For the front N Daily cumulative rainfall; For the first The cumulative daily rainfall a day ago; N The maximum cumulative number of days of rainfall considered; This represents the maximum effective continuous rainfall within the previous 30 days; The effective rainfall attenuation coefficient is denoted as , where The range of values ​​is (0,1); n The duration of the largest continuous rainfall event within the previous 30 days; (2) Reservoir water level factors The impact of reservoir water level fluctuations on landslide deformation, like rainfall, exhibits a lag. In addition to considering the effect of the reservoir water level fluctuation factor, the influence of reservoir water level elevation on the landslide deformation response is also taken into account. The specific definitions of the reservoir water level fluctuation factor and the reservoir water level elevation factor are as follows: ； ； ； ； ； In the formula: The average reservoir water level elevation over the previous N days; N For the number of days the reservoir water level is considered to affect, N =10,15,30,60; For the first The water level of the reservoir the day before yesterday; For the front N The daily increase in reservoir water level; This represents the rate of decline in reservoir water level over the previous N days. For the front N The relative fluctuation range of reservoir water level within the day; This represents the absolute change in reservoir water level over the previous N days. The initial factor set for external influencing factors of landslides is as follows: Rainfall: X1: Accumulated rainfall in the previous 3 days P 3; X2: Cumulative rainfall in the previous 5 days P 5; X3: Cumulative rainfall in the previous 7 days P 7; X4: Cumulative rainfall in the previous 15 days P 15 X5: Cumulative rainfall in the previous 30 days P 30 X6: Cumulative rainfall within the previous 45 days P 45 X7: Cumulative rainfall in the previous 60 days P 60 X8: Maximum consecutive rainfall within the previous 30 days X9: Maximum consecutive rainfall in the previous 30 days ;X 10 Maximum consecutive rainfall in the previous 30 days ;X 11 Maximum consecutive rainfall in the previous 30 days ;X 12 Maximum consecutive rainfall in the previous 30 days ; Reservoir water level: X 13 The increase in reservoir water level over the past 10 days ;X 14 The increase in reservoir water level over the past 15 days ;X 15 The increase in reservoir water level over the first 30 days ;X 16 The increase in reservoir water level over the past 60 days ;X 17 The percentage decrease in reservoir water level over the first 10 days ;X 18 The percentage decrease in reservoir water level over the first 15 days ;X 19 The percentage decrease in reservoir water level over the first 30 days ;X 20 The percentage decrease in reservoir water level over the first 60 days ;X 21 The relative fluctuation range of the reservoir water level in the first 10 days ;X 22 The relative fluctuation range of the reservoir water level in the first 15 days ;X 23 The relative fluctuation range of reservoir water level in the first 30 days ;X 24 The relative fluctuation range of reservoir water level in the first 60 days ;X 25 Absolute fluctuation range of reservoir water level in the first 10 days ;X 26 Absolute fluctuation range of reservoir water level in the first 15 days ;X 27 Absolute fluctuation range of reservoir water level in the first 30 days ;X 28 Absolute fluctuation range of reservoir water level in the first 60 days ;X 29 Average water level elevation of the reservoir over the previous 10 days ;X 30 Average water level of the reservoir over the previous 15 days ;X 31 Average reservoir water level elevation over the previous 30 days ;X 32 Average reservoir water level elevation over the previous 60 days .

2. The method according to claim 1, characterized in that, S3 specifically includes: Preliminary optimization: According to different factor categories, calculate the significance level index Sig and Pearson correlation coefficient between each factor in the initial factor set and the deformation event. Based on the calculation results, sort the factors of each type in the factor set according to the magnitude of the correlation coefficient between each factor and the deformation event, and remove factors with low correlation with the deformation event, thereby achieving the initial optimization of the candidate factor set. Secondary optimization: Calculate the Pearson correlation coefficient and significance level index Sig among the candidate factors after the initial optimization, and further screen them according to the correlation between each factor to obtain the final core factor set.

3. The method according to claim 2, characterized in that, The final core factor set includes: X7, X 12 X 32 X 30 X 24 X 22 .

4. The method according to claim 3, characterized in that, S4 specifically includes: S41. Establish a multiple stepwise regression model that considers time, reservoir water level, and rainfall, decomposing the landslide deformation process into its components. The multiple stepwise regression model is expressed as follows: ； In the formula: This is the amount of deformation; This refers to the deformation component caused by changes in aging. The deformation component caused by changes in rainfall; This represents the deformation component caused by reservoir water level fluctuations. D For statistical constants; S42. The rainfall and reservoir water level data are processed by using [0,1] linear normalization to map data from different ranges to a unified interval, so as to ensure consistent data scale and improve model stability. S43. Decompose the landslide deformation process into components and list the calculation formulas for each component in turn. (1) Time-dependent component , The deformation component caused by aging is expressed as: ； In the formula: This is the value obtained by dividing the cumulative number of days from the displacement observation date to the initial measurement date by 100; All of these are regression coefficients of the deformation components caused by aging changes, which can be obtained by regression analysis; (2) Rainfall components , Based on the optimization results of the rainfall pattern factor, the cumulative rainfall within the previous 60 days was used. and the maximum consecutive rainfall in the previous 30 days The formula for calculating rainfall components is as follows: ； In the formula: This refers to the cumulative rainfall over the previous 60 days; The maximum continuous rainfall within the previous 30 days, and the rainfall attenuation coefficient. ; The regression coefficients for the deformation components caused by changes in rainfall can be obtained through regression analysis; (3) Reservoir water level components Based on the optimization results of the reservoir water type factor, the formula for calculating the reservoir water level component is as follows: ； In the formula: These are the average reservoir water level elevations for the previous 60 days and 15 days, respectively. These represent the relative fluctuations in reservoir water levels over the previous 60 days and 15 days, respectively. The regression coefficients of the deformation components caused by reservoir water level fluctuations are given. ; S44. The final multiple stepwise regression model is as follows: ； 。 5. The method according to claim 1, characterized in that, The transformation stages are divided into: Phase 1: Initial deformation phase of the landslide, from June 2003 to September 2006; The second phase: the sensitive period affected by reservoir water levels, from October 2006 to September 2018; The third stage: the landslide stabilization and deformation stage, which lasts from October 2018 to February 2024; Current state: Based on the deformation trend, it is inferred that we are currently in the third stage.

6. The method according to claim 1, characterized in that, S6 specifically includes: In the first stage, the time-dependent component accounts for the main part of the deformation. The influence of reservoir water level on the landslide front edge is significantly greater than that of rainfall. According to the data from the initial stage, the landslide was generally in a stable state before water impoundment. The first phase of experimental water impoundment led to a significant rise in the water level in front of the slope, resulting in significant physical and mechanical effects, which caused the landslide to revive through front edge traction deformation. In the second stage, the time-dependent component of landslide deformation is the main component. Compared with the initial deformation stage, the proportion of the time-dependent component in this stage has increased significantly. The main external influencing factor is the fluctuation of reservoir water level, while the proportion of rainfall component is relatively small. The deformation mode of the landslide changes from the leading edge traction in the early stage of water storage to overall creep. In the third stage, the deformation pattern of the landslide did not change substantially, but its overall creep rate decreased significantly. This indicates that the deterioration effect caused by large-scale water-rock interaction tended to weaken. After years of reservoir water level fluctuations, the landslide gradually adapted to the changes in the groundwater environment caused by reservoir water level fluctuations. The time-dependent component was the main deformation component. Compared with the second stage, the proportion of the time-dependent component in this stage was roughly the same. This indicates that the deterioration caused by water-rock interaction was basically completed in the second stage. The creep rate of the landslide further decreased compared with the previous stage, and the landslide deformation trend slowed down. The external influencing factor changed from reservoir water level fluctuations in the second stage to seasonal rainfall. Comparing the proportions of rainfall and reservoir water components, it can be seen that at this time, the main external influencing factor changed from reservoir water level fluctuations to rainfall. Currently, we are in the third stage. The main controlling factor, which is the most important external influencing factor of landslides, has changed from reservoir water level to rainfall. This means that the focus of landslide monitoring and prevention should shift to the forecasting of heavy rainfall and the early warning and prevention of disasters that may be caused by it.

7. A system for phased classification and identification of main controlling factors of landslides in the Three Gorges Reservoir area, characterized in that, The system includes: The data collection and preprocessing module is used to collect and preprocess landslide monitoring data in the Three Gorges Reservoir area; The first module is used to preliminarily screen factors that may affect landslide stability based on the geological characteristics and monitoring data of the landslide, including rainfall and reservoir water level, and to establish an initial factor set of external influencing factors of the landslide around the two factors of rainfall and reservoir water level. The screening module is used to optimize the initial factor set twice through significance level index Sig and Pearson correlation coefficient analysis to screen out the core factor set that best reflects the changes in landslide stability. The second module is used to separate and study the influence of each core factor on landslide deformation based on the optimized core factor set and landslide deformation data, establish a multivariate stepwise regression model that considers time, rainfall and reservoir water level, decompose the landslide deformation process into each component, list the calculation formulas of each component in turn, and obtain the final multivariate stepwise regression model. The segmentation module is used to accurately segment the deformation stages of landslides by analyzing the landslide deformation curves at different time periods based on the regression model fitting results and combining evaluation indicators to identify the stage inflection points. The determination module is used to analyze the primary and secondary impacts of internal and external influencing factors on landslides at different stages by comparing the fitted amplitude values ​​with the measured amplitude values. Based on the changes in the proportion of each component in each stage, the main controlling factors for each stage and the current stage are determined. The first module is specifically used for: (1) Rainfall factors Landslide deformation exhibits a certain lag in its response to rainfall. The rainfall factor is the cumulative rainfall factor and the maximum effective rainfall factor within the previous 30 days, specifically defined as follows: ； ； In the formula: For the front N Daily cumulative rainfall; For the first The cumulative daily rainfall a day ago; N The maximum cumulative number of days of rainfall considered; This represents the maximum effective continuous rainfall within the previous 30 days; The effective rainfall attenuation coefficient is denoted as , where The range of values ​​is (0,1); n The duration of the largest continuous rainfall event within the previous 30 days; (2) Reservoir water level factors The impact of reservoir water level fluctuations on landslide deformation, like rainfall, exhibits a lag. In addition to considering the effect of the reservoir water level fluctuation factor, the influence of reservoir water level elevation on the landslide deformation response is also taken into account. The specific definitions of the reservoir water level fluctuation factor and the reservoir water level elevation factor are as follows: ； ； ； ； ； In the formula: The average reservoir water level elevation over the previous N days; N For the number of days the reservoir water level is considered to affect, N =10,15,30,60; For the first The water level of the reservoir the day before yesterday; For the front N The daily increase in reservoir water level; This represents the rate of decline in reservoir water level over the previous N days. For the front N The relative fluctuation range of reservoir water level within the day; This represents the absolute change in reservoir water level over the previous N days. The initial factor set for external influencing factors of landslides is as follows: Rainfall: X1: Accumulated rainfall in the previous 3 days P 3; X2: Cumulative rainfall in the previous 5 days P 5; X3: Cumulative rainfall in the previous 7 days P 7; X4: Cumulative rainfall in the previous 15 days P 15 X5: Cumulative rainfall in the previous 30 days P 30 X6: Cumulative rainfall within the previous 45 days P 45 X7: Cumulative rainfall in the previous 60 days P 60 X8: Maximum consecutive rainfall within the previous 30 days X9: Maximum consecutive rainfall in the previous 30 days ;X 10 Maximum consecutive rainfall in the previous 30 days ;X 11 Maximum consecutive rainfall in the previous 30 days ;X 12 Maximum consecutive rainfall in the previous 30 days ; Reservoir water level: X 13 The increase in reservoir water level over the past 10 days ;X 14 The increase in reservoir water level over the past 15 days ;X 15 The increase in reservoir water level over the first 30 days ;X 16 The increase in reservoir water level over the past 60 days ;X 17 The percentage decrease in reservoir water level over the first 10 days ;X 18 The percentage decrease in reservoir water level over the first 15 days ;X 19 The percentage decrease in reservoir water level over the first 30 days ;X 20 The percentage decrease in reservoir water level over the first 60 days ;X 21 The relative fluctuation range of the reservoir water level in the first 10 days ;X 22 The relative fluctuation range of the reservoir water level in the first 15 days ;X 23 The relative fluctuation range of reservoir water level in the first 30 days ;X 24 The relative fluctuation range of reservoir water level in the first 60 days ;X 25 Absolute fluctuation range of reservoir water level in the first 10 days ;X 26 Absolute fluctuation range of reservoir water level in the first 15 days ;X 27 Absolute fluctuation range of reservoir water level in the first 30 days ;X 28 Absolute fluctuation range of reservoir water level in the first 60 days ;X 29 Average water level elevation of the reservoir over the previous 10 days ;X 30 Average water level of the reservoir over the previous 15 days ;X 31 Average reservoir water level elevation over the previous 30 days ;X 32 Average reservoir water level elevation over the previous 60 days .

8. An electronic device comprising a processor and a memory, wherein the memory stores at least one instruction, characterized in that, The processor loads and executes at least one instruction to implement the method for phased division and identification of main control factors of landslides in the Three Gorges Reservoir area as described in any one of claims 1-6.

9. A computer-readable storage medium storing at least one instruction, characterized in that, The at least one instruction is loaded and executed by the processor to implement the method for phased division and identification of main control factors of landslides in the Three Gorges Reservoir area as described in any one of claims 1-6.

Images