A slope impending sliding prediction method and system based on deep monitoring deformation sequence and a medium

Abstract

The application provides a slope impending slide prediction method and system based on deep monitoring deformation sequence, and a medium, and the method comprises the following steps: obtaining a cumulative displacement-time (S-t) curve at a deep monitoring point of a slope; through coordinate transformation, the horizontal and vertical coordinates of the S-t curve are processed with the same dimension to obtain a T-t curve; a nonlinear creep model capable of reflecting the full-stage creep characteristics of the rock-soil mass of the slope is established by combining a fractional derivative element, a nonlinear viscoplastic body and a Nishihara creep model; when it is determined that the rock-soil mass of the slope is in a constant-speed deformation stage, displacement prediction is performed by using the nonlinear creep model; after it is determined that the rock-soil mass of the slope enters an accelerated deformation stage, the displacement sequence after the start of the accelerated deformation is taken as input, deformation prediction is performed by using the nonlinear creep model, and the slope instability time is predicted by using the rate reciprocal method; the evolution trend of the deep deformation of the slope, the impending slide prediction time and the apparent crack distribution characteristics are analyzed, and the warning level is comprehensively determined.

Description

Technical Field

[0001] This invention relates to the field of geological disaster monitoring and early warning technology during the construction and operation phases of large-scale water conservancy and hydropower projects, specifically to a method, system, and medium for predicting slope slippage based on deep monitoring deformation sequences. Background Technology

[0002] The numerous high slopes formed in the dam site area and near-dam reservoir area are typically massive in scale. Their complex external influencing factors and variable, uncertain geological conditions determine that slope instability has a significant impact and destructive potential, seriously threatening the safety of life and property in the near-dam area and the long-term stable operation of the dam. Therefore, stability assessment and early warning of landslides on high slopes in the dam site area and near-dam reservoir area are key issues in the construction and operation of hydropower projects.

[0003] With the rapid development of new slope monitoring methods and wireless sensing technology, all-weather and multi-level "sky-ground-air" monitoring of unstable slopes and potential landslides has been achieved, providing important technical support for slope deformation prediction and landslide early warning. However, given the uncertainty of slope deformation evolution and the complexity of its inducing mechanisms, simply analyzing the correlation between apparent slope deformation and environmental influencing factors is insufficient for accurate short- to medium-term slope deformation prediction. Furthermore, established instability criteria and landslide timing predictions suffer from unique uncertainties and lack universality. Current research on slope instability prediction and early warning primarily focuses on predicting deformation at observation points, and the prediction models cannot directly output specific landslide times. Moreover, when slopes undergo accelerated deformation, their deformation evolution characteristics differ significantly from those in historical monitoring data. Due to limitations in monitoring information and models, trained displacement prediction models cannot effectively predict these abrupt displacement changes. Therefore, at present, there is no method for predicting the entire process displacement and landslide instability of high slopes with clear physical significance. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a method, system, and medium for predicting landslides based on deep monitoring deformation sequences. This addresses how to consider the evolutionary characteristics of slope deformation based on the creep theory of soil and rock, enabling the use of slope deformation evolution stage characteristics and the rate reciprocal method for landslide instability early warning. This provides strong support for the deformation evolution analysis and landslide prediction system construction of high slopes in water conservancy and hydropower projects.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, the present invention provides a method for predicting slope slippage based on deep monitoring deformation sequences, comprising the following steps:

[0007] S1. Collect cumulative deformation data at the target measuring point deep in the slope using a deep deformation measuring device, and use Origin drawing software to draw the cumulative deformation-time (St) curve. Use coordinate transformation to make the horizontal and vertical coordinates of the deformation curve the same dimension to form the Tt curve.

[0008] S2. Based on the creep theory of soil and rock, the Nishihara creep model is improved by introducing fractional derivative elements and nonlinear viscoplastic bodies. A constitutive model that can reflect the creep characteristics of soil and rock in the slope throughout the whole stage is established, and the model parameters are identified by combining the Tt curve.

[0009] S3. Dynamically identify the acceleration deformation initiation point (AP point) of the Tt curve. When the deep slope deformation is in the uniform deformation stage, the deformation amount is predicted by nonlinear creep constitutive model. When it is determined that the deep slope deformation has entered the accelerated deformation stage, the monitoring deformation sequence after the AP point is used to combine the nonlinear creep model and the reciprocal rate method to obtain the intersection point of the reciprocal rate trend line and the Tt curve on the time axis, that is, the slope instability time t. f ;

[0010] S4. Based on the calculation results of steps S2 and S3 above, analyze the evolution trend of the deep deformation curve of the slope and the instability prediction time t. f The warning level is determined by combining the distribution characteristics of apparent cracks on the slope.

[0011] While adopting the above technical solutions, the present invention may also adopt or combine the following technical solutions:

[0012] As a preferred technical solution of the present invention: In step S1, the dimensionless processing of the deformation curve includes: initially identifying and dividing the initial deformation stage, the constant speed deformation stage and the accelerated deformation stage according to the characteristics of the St curve, and transforming the dimension of the vertical axis of the St curve using the average deformation rate v of the constant speed deformation stage to obtain the calculation formula of the Tt curve with the same dimension of vertical and horizontal axes, as shown in formula (1):

[0013]

[0014] In the formula, T(i) is the ordinate value after dimension conversion, and ΔS(i) is the deformation change within a certain unit time period. The average displacement rate during the constant velocity deformation stage.

[0015] As a preferred technical solution of the present invention: In step S2, the commonly used Riemann-Liouville (RL) type fractional calculus theory is introduced, and under the constant shear stress condition (σ(t)=τ0), the creep equation of the fractional derivative element is obtained, as shown in formula (2):

[0016]

[0017] In the formula, Γ(·) is the gamma function, which is a generalization of integer-order operations in the real and imaginary domains. When Re is the real part of a complex number, it is as shown in equation (3):

[0018]

[0019] The steps to improve the Nishihara model using fractional derivative elements and nonlinear viscoplastic bodies include: replacing the Newton's stick pot in the Kelvin body of the Nishihara model with fractional derivative elements, and replacing the Newton's stick pot in the Bingham body with nonlinear viscous elements. Considering that the viscosity coefficient η(t) of the element is a time-dependent indeterminate constant, after introducing the η(t) viscosity coefficient which decays exponentially with time, the creep equation of the nonlinear viscous element is obtained, as shown in the following formula (4):

[0020]

[0021] In the formula, τ is the shear stress, τ s η represents the long-term shear strength of the soil and rock mass, n is the creep exponent of the nonlinear viscous element, and η is the creep index of the nonlinear viscous element. NV H(τ-τ) is the viscosity coefficient. s ) represents the Heaviside step function. Based on the Nishihara creep model, nonlinear creep model equations (5) and (6) are constructed:

[0022] S1=a+bb•E β,1 (-c•t β (5) (Constant velocity deformation stage)

[0023]

[0024] In the formula, S1 is the displacement of the soil and rock mass in the constant deformation stage, and S2 is the displacement of the soil and rock mass in the accelerated deformation stage; a=τ / G1, b=τ / G2, c=G1 / η1 and M=H(τ-τ s ) / η NV For the fitting parameters, G1 and G2 are the material shear moduli of Hooke and Kelvin bodies, respectively; η1 is the material viscosity coefficient of Kelvin body; β is the order of the fractional derivative; n is the creep exponent of the nonlinear viscous element; and t is the cumulative monitoring time. Wherein, E... β,1 (·) is the Mittag-Leffler function, corresponding to the following equation (7):

[0025]

[0026] In the formula, z and β are both variables in the functional equation.

[0027] As a preferred technical solution of the present invention: In step S3, the method for accurately identifying the AP point of the Tt curve includes: calculating the short-term deformation rate average line (SMA) and the long-term deformation rate average line (LMA), and dynamically identifying the AP(t0,s0) point using the "positive intersection point" of SMA and LMA. The deformation rate average line is calculated as shown in formula (8):

[0028]

[0029] In the formula, v is the moving average rate. t Let n be the velocity value at time t. t It is an integer multiple of the unit monitoring cycle.

[0030] As a preferred technical solution of the present invention: In step S3, the prediction of deep slope deformation and calculation of instability time include: when the current slope deformation is in the isotropic deformation stage, the Tt curve is fitted using equation (5), and the least squares method and Grey Wolf Optimization (GWO) algorithm are used to globally optimize and identify the model parameters. The optimization objective function of the GWO algorithm is shown in equation (9):

[0031]

[0032] In the formula, a, b, c, and β are the parameters of the model to be identified, and t i To calculate the cumulative monitoring time, n s To monitor the length of the deformation sequence, S i and These are the monitored deformation value and the model-predicted deformation value, respectively; E β,1 (·) represents the Mittag-Leffler function, corresponding to formula (7);

[0033] If the current slope deformation is in the accelerated deformation stage after point AP, a new coordinate system is established with point AP as the origin, and the deformation sequence after point AP is used as the model input for deformation curve fitting. The least squares method and GWO optimization algorithm are used to identify the model parameters. The optimization objective function of the GWO algorithm is shown in the following formula (10):

[0034]

[0035] In the formula, M and n are the parameters of the model to be identified, and n s ′ represents the length of the deformation sequence monitored after point AP, S i and These represent the monitored deformation value and the model predicted value, respectively. i′ represents the relative time starting from AP. After the slope deformation enters the accelerated deformation stage, the model parameters M and n are substituted into equation (10), and the relationship curve of the inverse of the deformation rate - time (1 / V ~ t) is established after differentiating with respect to time t, as shown in equation (11):

[0036]

[0037] In the deformation-time coordinate system with AP as the origin, to further simplify the reciprocal rate model, it is assumed that the trend of the reciprocal rate curve is linear. By fitting the 1 / V~t′ curve in equation (11) with a linear function, the intersection of the linear trend line and the t-axis is obtained, which is the slope instability time t. f .

[0038] As a preferred technical solution of the present invention: In step S4, the predicted displacement value and landslide instability time for a future period are calculated based on the nonlinear creep model, and the deep deformation trend of the slope and the instability time t are comprehensively analyzed. f Based on the spatiotemporal evolution of apparent cracks, the slope warning levels were determined as follows: Attention Level, Warning Level, Alert Level, and Alarm Level.

[0039] Secondly, this invention provides a slope slip prediction system based on deep monitoring deformation sequences, comprising the following sub-modules:

[0040] The data acquisition module is used to collect the deep monitoring deformation of the target point and plot the cumulative deep deformation-time (St) curve;

[0041] The data preprocessing module is used to dynamically identify each deformation stage of the deep deformation St curve and calculate the average rate of the constant-rate deformation stage. The Tt curve is obtained by making the St curve dimensionless.

[0042] A creep model construction module is used to introduce fractional derivative elements and nonlinear viscoplastic bodies into the Nishihara creep constitutive model to establish a nonlinear relationship between deep deformation and time at the slope monitoring point.

[0043] The slope instability time prediction module is used to predict the deformation amount at each deformation stage of the slope. When the slope deformation enters the accelerated deformation stage, the module uses the landslide deformation curve after point AP and the inverse rate method to predict the slope instability time t. f ;

[0044] The early warning signal release module is used to comprehensively analyze the deformation trend, instability time, and spatiotemporal characteristics of macroscopic cracks, and to promptly release the corresponding slope early warning level and instability time.

[0045] In another aspect, the present invention provides a computer-readable storage medium including computer-readable instructions, which, when executed, cause a processor to perform the steps of the slope slip prediction method based on deep monitoring deformation sequences as described above.

[0046] This invention provides a method, system, and medium for predicting slope slippage based on deep monitoring deformation sequences. The method includes: acquiring the cumulative displacement-time (St) curve at deep monitoring points of the slope using a deformation monitoring device; transforming the horizontal and vertical coordinates of the St curve to achieve dimensionless processing to obtain the Tt curve; introducing the creep theory of soil and rock mass, and combining fractional derivative elements, nonlinear viscoplastic bodies, and the Nishihara creep model to establish a nonlinear creep model that reflects the creep characteristics of the slope soil and rock mass throughout the entire process; when the slope soil and rock mass is determined to be in the isotropic deformation stage, displacement prediction is performed using the nonlinear creep model; after the slope soil and rock mass is determined to have entered the accelerated deformation stage, the displacement sequence after the accelerated deformation start point (AP) is used as input to predict deformation using the nonlinear creep model, and the slope instability time t is predicted using the reciprocal rate method. f The analysis of the deep deformation evolution trend, the prediction time of landslide, and the distribution characteristics of apparent cracks in the slope is used to comprehensively determine the early warning level and issue slope instability warning signals through the early warning system.

[0047] This invention is based on the deep deformation monitoring sequence of slope geological bodies, effectively avoiding invalid data caused by sudden changes in human engineering activities and environmental factors in the apparent deformation sequence. Therefore, the deep deformation monitoring data of slopes has greater value for early warning analysis than surface deformation information. Furthermore, this invention fully integrates the nonlinear creep theory of soil and rock masses and the method for predicting landslides. It utilizes fractional derivative elements and nonlinear viscoplastic bodies to establish the nonlinear creep relationship of soil and rock masses, which can better describe the full-stage creep characteristics of slope soil and rock masses, and the physical meaning of the model parameters is clear. Combining deep deformation measurement devices, nonlinear creep models, the reciprocal rate method, and the macroscopic crack distribution characteristics of slopes, this invention integrates the needs for deep deformation information acquisition, evolution process analysis, and landslide prediction, providing strong support for the deformation evolution analysis and landslide prediction system construction of high slopes in water conservancy and hydropower projects. Attached Figure Description

[0048] Figure 1 The flowchart shows the slope slip prediction method based on deep monitoring deformation sequence provided by this invention.

[0049] Figure 2 This is the cumulative deformation-time (St) curve at a typical monitoring point on a slope.

[0050] Figure 3 This is the deformed time (Tt) curve after the vertical and horizontal axes are converted to the same dimension.

[0051] Figure 4 This is a schematic diagram of the Nishihara creep constitutive model.

[0052] Figure 5 This is a schematic diagram of the nonlinear creep model structure.

[0053] Figure 6 The nonlinear creep model is used to perform curve fitting and model parameter identification diagrams after the slope enters the accelerated deformation stage.

[0054] Figure 7 This is a time relationship diagram for slope displacement prediction and instability prediction based on a nonlinear creep model. Detailed Implementation

[0055] To more clearly illustrate the technical solution of the present invention, the method and system of the present invention are further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. After reading the present invention, any modifications of the present invention in various equivalent forms by those skilled in the art fall within the scope defined by the appended claims.

[0056] Numerous studies have shown that the deformation evolution and instability characteristics of slopes are controlled by the creep mechanical properties of the internal soil and rock mass. The deformation evolution process typically includes three stages: initial deformation, isochronous deformation, and accelerated deformation, which can be considered a standard for landslide prediction. Therefore, based on the deep deformation monitoring sequence of slopes, this study introduces soil and rock creep theory to establish a slope deformation prediction and instability time prediction model with clearly defined physical meanings and concepts. Please refer to [link / reference]. Figure 1 This embodiment uses a slope in an open-pit coal mine as an example for calculation and analysis. The deformation of the slope geological body is continuously monitored with a monitoring cycle of 10 minutes / time. The displacement sequence of monitoring points at a typical slope location is selected to construct a prediction model and calculate the time to slippage and instability.

[0057] S1. Based on the deformation monitoring data of this typical monitoring point, the cumulative deformation-time (St) curve was plotted using Origin plotting software. (See attached image) Figure 2 Based on the characteristics of the St curve, it is divided into the initial deformation stage, the constant-rate deformation stage, and the accelerated deformation stage. The dimensions of the ordinate of the St curve are transformed using the average deformation rate v of the constant-rate deformation stage, resulting in the calculation formula for the Tt curve, which has the same dimensions for both the ordinate and abscissa, as follows:

[0058]

[0059] In the formula, T(i) is the ordinate value after dimension conversion, and ΔS(i) is the deformation change within a certain unit time period. The average displacement rate during the constant velocity deformation stage.

[0060] The slope deformation-time (St) curve is converted into a deformation-time (Tt) relationship curve with the same dimensions for both the vertical and horizontal axes using calculation formula (1). See [link to relevant documentation]. Figure 3 ;

[0061] S2, the Nishihara model, is a classic element model describing the rheological properties of rock and soil masses. It consists of Hooke, Kelvin, and Bingham bodies connected in series. See [link to relevant documentation]. Figure 4 The Nishihara model can effectively describe the linear, non-decaying creep behavior of soil and rock masses. However, since the basic elements in the Nishihara model are all ideal linear elements, the combined element model is difficult to reflect the nonlinear accelerated creep behavior of soil and rock materials. To better describe the creep characteristics of soil and rock masses throughout the entire process, based on the Nishihara model, fractional derivative soft elements and nonlinear viscous bodies were introduced to replace the Newtonian sticky pots in the Kelvin and Bingham bodies, respectively, to construct the following nonlinear creep model (see [reference]). Figure 5 :

[0062] S1=a+bb·E β,1 (-c·t β (Constant velocity deformation stage)

[0063]

[0064] In the formula, S1 is the displacement of the soil and rock mass in the constant deformation stage, and S2 is the displacement of the soil and rock mass in the accelerated deformation stage; a=τ / G1, b=τ / G2, c=G1 / η1 and M=H(τ-τ s ) / η NV For the fitting parameters, G1 and G2 are the material shear moduli of Hooke and Kelvin bodies, respectively; η1 is the material viscosity coefficient of Kelvin body; β is the order of the fractional derivative; n is the creep exponent of the nonlinear viscous element; t is the cumulative monitoring time; and E β,1 (·) is the Mittag-Leffler function.

[0065] Based on the transformed Tt curve, the short-term deformation rate average (SMA) and long-term deformation rate average (LMA) are calculated respectively. Then, the AP(t0,s0) point is dynamically identified using the "positive intersection point" of the SMA and LMA. The formula for calculating the deformation rate average is as follows:

[0066]

[0067] In the formula, v is the moving average rate. t Let n be the velocity value at time t. tIt is an integer multiple of the unit monitoring period. In this example, the moving average of SMA is n=3, the moving average of LMA is n=7, and the time corresponding to the AP point is identified as t0=76.07h.

[0068] If the slope deformation is in the constant-rate deformation stage before point AP, the Tt curve is fitted, and the least squares method and Grey Wolf Optimization (GWO) algorithm are used to optimize and identify the model parameters. The optimization objective function of the GWO algorithm is shown in the following formula:

[0069]

[0070] In the formula, a, b, c, and β are the parameters of the model to be identified, and t i To calculate the cumulative monitoring time, n s To monitor the length of the deformation sequence, S i and These are the monitored deformation values ​​and the model-predicted deformation values, respectively.

[0071] If the current slope deformation is in the accelerated deformation stage after point AP, establish a new coordinate system with point AP as the origin, and use the deformation sequence after point AP as the model input for deformation curve fitting. See [link to relevant documentation]. Figure 6 The least squares method and the GWO optimization algorithm are used to identify the model parameters. The optimization objective function of the GWO algorithm is shown below:

[0072]

[0073] In the formula, M and n are the parameters of the model to be identified, and S i and These represent the monitored deformation value and the model-predicted deformation value, respectively. i ′ represents the relative time starting from AP.

[0074] Taking the slope entering the accelerated deformation stage as an example, the technical solution is illustrated, and the displacement sequence (t) after the displacement curve enters the accelerated deformation (AP) stage is presented. i ,S i Using a set of wolves (i = 1, 2, ..., n) as model input for fitting, the least squares method and the GWO optimization algorithm are used to identify model parameters M and n. The model parameter identification problem is transformed into a global optimization problem using GWO to solve the extrema in the above equation. In this example, the parameters of the GWO algorithm are set as follows: the population size N of gray wolves is 100, the maximum number of update iterations is 200, and the parameter search space dimension is 2. After iterative optimization using the GWO algorithm, the objective function f is obtained. i The coordinates (x, y) of the global minimum value are the model parameters (M, n).

[0075] Substituting the identified model parameters M and n into the above equation, and taking the derivative with respect to time t, we establish the relationship curve between the reciprocal of the deformation rate and time (1 / V ~ t), as shown below:

[0076]

[0077] A deformation-time coordinate system is established with point AP as the origin, and a 1 / V ~ t curve is fitted using a linear function. The intersection of the fitted line and the t-axis is then determined to be the slope instability time t. f .

[0078] The GWO algorithm was applied to optimize and identify model parameters based on displacement data following the AP (Advanced Perception) event. The model was then progressively updated and corrected using monitoring data that was closer to the landslide time, resulting in predicted displacement values ​​and instability times. (See [link to relevant documentation]). Figure 7 The model parameters and prediction errors in the implementation cases are shown in Table 1.

[0079] Table 1. Statistical Table of Model Identification Parameters and Predicted Landslide Time

[0080]

[0081] Slope deformation monitoring data indicates the actual landslide time was 94.23 hours. As shown in Table 1, as the monitoring data input into the model gradually approached the moment of the dramatic landslide, the predicted time of the nonlinear creep model became increasingly closer to the actual landslide time, and the predicted time gradually stabilized. The prediction error rate was within 5% in the 10 hours leading up to the actual landslide. The final predicted time calculated using the creep model and the reciprocal rate method was 94.33 hours, which is basically consistent with the actual landslide time, verifying the effectiveness of the short-term landslide prediction based on the nonlinear creep model in this invention.

[0082] This invention also provides a slope slip prediction system based on deep monitoring deformation sequences, comprising the following sub-modules:

[0083] The data acquisition module is used to automatically collect the deep deformation at the target monitoring points of the slope and call Origin to draw the cumulative deep deformation-time (St) curve;

[0084] The data preprocessing module is used to automatically identify each deformation stage of the deformation-time (St) curve, calculate the average rate v of the constant-rate deformation stage, and convert the St curve to a dimensionless form to obtain the Tt curve.

[0085] The creep model construction module is used to introduce fractional derivative elements and nonlinear viscoplastic bodies into the Nishihara creep constitutive model, establish the nonlinear relationship between deep deformation and time at the slope measuring points, and combine the actual slope deformation-time (Tt) curve to call the GWO optimization algorithm to identify the model parameters of the nonlinear creep model.

[0086] The slope instability time prediction module is used to calculate and predict the deformation at each stage of deformation. For the slope deformation after point AP (accelerated deformation stage), it uses the reciprocal rate method to calculate and predict the slope instability time t. f ;

[0087] The early warning signal release module is used to comprehensively analyze the deformation evolution trend, instability time, and spatiotemporal characteristics of macroscopic cracks, and to promptly release the corresponding slope early warning level and instability time.

[0088] The present invention also provides a computer-readable storage medium including computer-readable instructions, which, when executed, cause a processor to perform the steps of the slope slip prediction method based on deep monitoring deformation sequences as described above.

[0089] The slope slip prediction system based on deep monitoring deformation sequences and the computer-readable storage medium including computer-readable instructions described above share the same inventive concept as the aforementioned method. Therefore, for any aspects not described here, please refer to the relevant descriptions in the foregoing method embodiments, which will not be repeated here. This invention, based on the monitoring deformation sequences of deep slope rock and soil, fully integrates the nonlinear creep theory of rock and soil with slope slip prediction methods. It can achieve integrated automation of deep slope deformation information acquisition, dynamic analysis of the evolution process, deformation prediction, and instability time prediction. It can provide strong support for the deformation evolution analysis and slip prediction early warning system construction of high slopes in dam areas of water conservancy and hydropower projects, and has significant engineering significance and good application prospects.

[0090] The above specific embodiments are used to explain and illustrate the present invention, and are only preferred embodiments of the present invention, not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made to the present invention within the spirit and scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A method for predicting slope sliding based on deep monitoring deformation sequence, characterized in that: The method includes the following steps: S1, collecting accumulated deformation data of a target measuring point in a deep part of a slope by a deep deformation measuring device, drawing an accumulated deformation-time curve, and performing same-dimension processing on the horizontal and vertical coordinates of the deformation curve by coordinate change into Tt curve; S2. Based on the creep theory of soil and rock masses, a fractional derivative element and nonlinear viscoplastic body are introduced to improve the Nishihara creep model, establishing a constitutive model that can reflect the creep characteristics of slope soil and rock masses throughout the entire process. Tt Curve identification model parameters; S3, Dynamic Recognition Tt The acceleration deformation starting point of the curve is predicted by using a creep constitutive model when the deep deformation of the slope is in the uniform deformation stage. When it is determined that the deep deformation of the slope has entered the accelerated deformation stage, the inverse rate trend line is obtained by using the monitoring deformation sequence after the point, combining the nonlinear creep model and the rate inverse method. Tt The intersection of the curves along the horizontal axis represents the time of slope instability. ; S4. Based on the calculation results of steps S2 and S3 above, analyze the evolution trend of the deep deformation curve of the slope and the instability prediction time. The warning level is determined by comprehensively considering the macroscopic crack distribution characteristics of the slope. In step S2, Introducing the commonly used Riemann-Liouville (RL) type fractional calculus theory, under the condition of constant shear stress, i.e. The creep equation for the fractional derivative element is obtained as shown in formula (2): （2） In the formula, For the gamma function, this function is a generalization of integer-order operations in the real and imaginary domains. When Re is the real part of a complex number, it is as shown in equation (3): （3） The steps for improving the Nishihara model using fractional derivative elements and nonlinear viscoplastic bodies include: replacing the Newton's gluepot in the Kelvin body of the Nishihara model with fractional derivative elements; replacing the Newton's gluepot in the Bingham body with nonlinear viscous elements to reflect the nonlinear accelerated creep characteristics of the soil and rock mass; and considering the viscosity coefficient of the elements. It is a time-dependent non-stationary constant, which introduces an exponentially decaying relationship with time. After determining the viscosity coefficient, the creep equation for the nonlinear viscous element is obtained, as shown in formula (4) below: （4） In the formula, For shear stress, The long-term shear strength of the soil and rock mass. n The creep exponent of a nonlinear viscous element. The viscosity coefficient of the nonlinear Bingham material is... The Heaviside step function is used; based on the Nishihara creep model, nonlinear creep model equations (5) and (6) are constructed: Constant velocity deformation stage: (5) Accelerated Deformation Stage: (6) In the formula, This represents the displacement of the soil and rock mass during the constant-rate deformation stage. Displacement of the rock and soil mass during the accelerated deformation stage; and For the fitting parameters, and The material shear moduli of Hooke bodies and Kelvin bodies, respectively. The viscosity coefficient of Kelvin material; Let be the order of the fractional derivative. n The creep exponent of a nonlinear viscous element. t This refers to the cumulative monitoring time; among which, The Mittag-Leffler function corresponds to the following equation (7): （7） In the formula, z and All of these are variables in the functional equation. .

2. The slope slip prediction method based on deep monitoring deformation sequence according to claim 1, characterized in that: In step S1, the dimensionless processing of the deformation curve includes: according to St The curve features were initially identified and divided into the initial deformation stage, the constant-rate deformation stage, and the accelerated deformation stage. The average deformation rate of the constant-rate deformation stage was then used to... right St The dimensions of the ordinate of the curve are transformed to obtain a curve with ordinates and abscissas of the same dimensions. Tt The formula for calculating the curve is shown in formula (1): （1） In the formula, The ordinate value is the dimensional transformation value. The change in deformation within a certain unit of time period. The average displacement rate during the constant velocity deformation stage.

3. The slope slip prediction method based on deep monitoring deformation sequence according to claim 1, characterized in that: In step S3, accurate identification Tt curve AP The point-based method includes: calculating the short-term deformation rate moving average (SMA) and the long-term deformation rate moving average (LMA), and using the "positive intersection point" of the SMA and LMA to dynamically identify... The average deformation rate of the point is calculated as shown in formula (8): （8） In the formula, The moving average rate, for t Time rate value, It is an integer multiple of the unit monitoring cycle.

4. The slope slip prediction method based on deep monitoring deformation sequence according to claim 1, characterized in that: In step S3, the prediction of deep slope deformation and the calculation of instability time include: combining AP The deformation stage of the slope is determined by the point. If the current slope deformation is within the constant rate deformation stage, the fitting is performed. Tt The curve is obtained, and the least squares method and the Grey Wolf Optimization (GWO) algorithm are used to optimize and identify the model parameters. The optimization objective function of the GWO algorithm is shown in formula (9): （9） In the formula, a , b , c and For the parameters of the model to be identified, To accumulate monitoring time, To monitor the length of the deformation sequence, and These are the monitored deformation values ​​and the model-predicted deformation values, respectively. For Mittag-Leffler functions; If the current slope deformation is at AP The accelerated deformation stage after the point, with AP Establish a new coordinate system with point as the origin, and... AP The deformation sequence after the point is used as the model input for deformation curve fitting. The least squares method and GWO optimization algorithm are used to identify the model parameters. The optimization objective function of the GWO algorithm is shown in the following formula (10): （10） In the formula, M and n For the parameters of the model to be identified, for AP After the point is reached, monitor the length of the deformation sequence. and These are the monitored deformation values ​​and the model-predicted deformation values, respectively. For AP The relative time starting from the slope deformation stage; after the slope deformation enters the accelerated deformation stage, the model parameters are... M and n Substitute it into equation (10) to adjust the time. t Differentiate and establish the reciprocal of deformation rate - time ( The relationship curve is shown in formula (11): （11） In AP In the deformed-time coordinate system with the origin at point A, to further simplify the reciprocal rate model, it is assumed that the trend of the reciprocal rate curve is linear; the linear function is used to fit equation (11) The curve is further used to obtain the linear trend line and... t The intersection of the axes represents the time of slope instability. .

5. The slope slip prediction method based on deep monitoring deformation sequence according to claim 1, characterized in that: In step S4, the predicted displacement and the time to imminent slippage are calculated based on the nonlinear creep model, and the deep deformation trend and the time to imminent slippage of the slope are comprehensively analyzed. Based on the spatiotemporal evolution of apparent cracks, the slope warning levels were determined as follows: Attention Level, Warning Level, Alert Level, and Alarm Level.

6. A slope slip prediction system based on deep monitoring deformation sequences, characterized in that, The system is based on the slope slip prediction method based on deep monitoring deformation sequence as described in claim 1, and includes the following modules: The data acquisition module is used to collect the deep monitoring deformation of the target point and plot the cumulative deep deformation-time (DTD) curve. St )curve; The data preprocessing module is used to dynamically identify deep deformations. St Calculate the average rate of the constant-rate deformation stage for each deformation stage of the curve. ,Will St Curve dimensionless processing yields Tt curve; A creep model construction module is used to introduce fractional derivative elements and nonlinear viscoplastic bodies into the Nishihara creep constitutive model to establish a nonlinear relationship between deep deformation and time at the slope monitoring point. A slope instability time prediction module is used to predict the deformation amount at each deformation stage of the slope. When the slope deformation enters the accelerated deformation stage, it utilizes... AP The landslide deformation curve after the point is used to predict the slope instability time using the reciprocal rate method. ; The early warning signal release module is used to comprehensively analyze deformation trends, instability time, and the spatiotemporal characteristics of macroscopic cracks, and promptly release corresponding slope early warning information.

7. A computer-readable storage medium comprising computer-readable instructions, characterized in that, When executed, the computer-readable instructions cause the processor to perform the steps of the slope slip prediction method based on deep monitoring deformation sequences as described in any one of claims 1-5.

Images