Method for determining landslide instability based on water power and pore water pressure and terminal device

Abstract

The application discloses a landslide instability determination method and terminal equipment based on water power and pore water pressure, belongs to the technical field of engineering safety monitoring, and can solve the problem that the existing method cannot realize quantitative prediction of instability probability and trend extrapolation, and is difficult to meet the timeliness requirement of disaster prevention. The method comprises the following steps: S1, determining soil water flow characteristic data of a reservoir bank landslide body and water level fluctuation data of a reservoir according to reservoir bank landslide body monitoring data and reservoir monitoring data; S2, constructing a dynamic coupling evolution model based on water power and pore water pressure; S3, determining deformation characteristic data of the reservoir bank landslide body according to the reservoir bank landslide body monitoring data and the dynamic coupling evolution model, and constructing a linkage early warning model based on water power, pore water pressure and displacement; and S4, performing instability prediction of the reservoir bank landslide body by using an instability early warning index according to the reservoir bank landslide body monitoring data, and obtaining an instability probability prediction result. The application is used for landslide body instability determination.

Description

Technical Field

[0001] This invention relates to a method and terminal equipment for determining landslide instability based on hydrodynamics and pore water pressure, belonging to the field of engineering safety monitoring technology. Background Technology

[0002] Traditional landslide instability assessment techniques are typically characterized by static parameter threshold determination and single-factor monitoring. In practice, pore water pressure sensors, displacement monitoring devices, and rain gauges are usually deployed at key locations within the landslide body. Data on pore water pressure, slope displacement, and rainfall are acquired through manual inspections or periodic data collection. The assessment logic is often based on empirical threshold rules; for example, an early warning signal is triggered when pore water pressure exceeds a preset fixed threshold or displacement rate exceeds a set critical value. Some methods combine limit equilibrium theory or Coulomb earth pressure models, calculating the slope safety factor and comparing it with the critical value to determine the stability state. This technical approach is simple to operate, low in cost, and easy to implement in the field, playing a fundamental supporting role in early landslide monitoring and early warning systems, especially suitable for scenarios with limited data acquisition conditions and relatively stable monitoring needs.

[0003] However, existing technologies suffer from a significant drawback: static threshold judgment models struggle to capture dynamic evolutionary characteristics. Static thresholds, based on empirical values ​​or single-parameter critical values, fail to reflect the temporal correlation between pore water pressure and reservoir water level / rainfall, leading to delayed or missed warnings. Single-factor monitoring fragments the coupling effects of multiple factors; for example, the combined effect of reservoir water level fluctuations and rainfall infiltration is difficult to quantify, reducing judgment accuracy. Low-frequency data acquisition struggles to capture periodic fluctuations, accelerating trends, or precursors of sudden changes in pore water pressure. Single-threshold warnings are prone to false triggering due to brief parameter exceedances or missed warnings due to unrecognized coordinated changes in multiple parameters. The lack of dynamic evolution model support prevents quantitative prediction of instability probability and trend extrapolation, failing to meet the timeliness requirements of disaster prevention and control. Summary of the Invention

[0004] This invention provides a method and terminal device for determining landslide instability based on hydrodynamics and pore water pressure. It can solve the problems of existing methods lacking dynamic evolution model support, being unable to achieve quantitative prediction of instability probability and trend extrapolation, and failing to meet the timeliness requirements of disaster prevention and control.

[0005] On one hand, the present invention provides a method for determining landslide instability based on hydrodynamics and pore water pressure, the method comprising:

[0006] S1. Based on the monitoring data of the reservoir bank landslide and the reservoir monitoring data, determine the soil flow characteristics of the reservoir bank landslide and the water level fluctuation data of the reservoir.

[0007] S2. Based on the soil flow characteristic data and the water level fluctuation data, construct a dynamic coupled evolution model based on hydrodynamics and pore water pressure;

[0008] S3. Based on the monitoring data of the reservoir bank landslide and the dynamic coupling evolution model, determine the deformation characteristic data of the reservoir bank landslide, and construct an early warning model based on the linkage between hydrodynamics, pore water pressure and displacement based on the deformation characteristic data;

[0009] S4. Determine the instability warning index based on the linkage warning model, and use the instability warning index to predict the instability of the reservoir bank landslide based on the monitoring data of the reservoir bank landslide, and obtain the instability probability prediction result.

[0010] Optionally, the monitoring data of the reservoir bank landslide includes pore water pressure data, displacement data, and rainfall data of the reservoir bank landslide; the monitoring data of the reservoir includes reservoir water level data and water flow velocity data.

[0011] The soil flow characteristics data include soil permeability data, water pressure gradient data, and water scour intensity data.

[0012] Optionally, S1 specifically includes:

[0013] Based on the pore water pressure data and the reservoir water level data, the soil permeability data of the reservoir bank landslide body were determined using the improved Darcy's law;

[0014] Based on the pore water pressure data, the water pressure gradient data of the reservoir bank landslide body is determined using the central difference method;

[0015] Based on the water flow velocity data, the water flow scouring intensity data of the reservoir bank landslide body is determined using the Xie Cai formula.

[0016] Based on the reservoir water level data, time series analysis and fluctuation feature extraction methods are used to determine the water level fluctuation data of the reservoir.

[0017] Optionally, S2 specifically includes:

[0018] The soil permeability data, water pressure gradient data, water flow scour intensity data, water level fluctuation data, pore water pressure data, reservoir water level data, water flow velocity data, and rainfall data in the area where the reservoir bank landslide is located are spatiotemporally correlated to obtain spatiotemporal correlation parameters.

[0019] A dynamic coupled evolution model based on hydrodynamics and pore water pressure is constructed based on the spatiotemporal correlation parameters.

[0020] Optionally, the deformation characteristic data includes geological displacement parameters and precursor characteristics of abrupt change points of the reservoir bank landslide, as well as periodic fluctuation characteristics and trend acceleration characteristics of reservoir water level data, water pressure gradient data, and rainfall data; S3, determining the deformation characteristic data of the reservoir bank landslide based on the monitoring data and the dynamic coupling evolution model, specifically includes:

[0021] The geological displacement parameters of the reservoir bank landslide are determined based on the displacement data and the dynamic coupling evolution model.

[0022] Based on the pore water pressure data and the water flow scour intensity data, the precursor characteristics of the sudden change point of the reservoir bank landslide were determined;

[0023] The periodic fluctuation characteristics of the reservoir water level data, the water pressure gradient data, the rainfall data, and the dynamic coupling evolution model are determined based on the water level fluctuation data, the water pressure gradient data, the rainfall data, and the dynamic coupling evolution model.

[0024] The trend acceleration characteristics are determined based on the periodic fluctuation characteristics and the soil permeability data.

[0025] Optionally, the step S3, which involves constructing a linkage early warning model based on the deformation feature data and the interaction between hydrodynamics, pore water pressure, and displacement, specifically includes:

[0026] Determine the weighting coefficients for the geological displacement parameters, the precursor features of the abrupt change point, the periodic fluctuation features, and the trend acceleration features;

[0027] Based on the geological displacement parameters, the precursor characteristics of the abrupt change point, the periodic fluctuation characteristics, the trend acceleration characteristics, and their corresponding weighting coefficients, a linkage early warning model based on hydrodynamics, pore water pressure, and displacement is constructed.

[0028] Optionally, after S4, the method further includes:

[0029] S5. Based on the pore water pressure data and the displacement data, and using the linkage early warning model, extrapolate the landslide trend of the reservoir bank landslide body to obtain the landslide trend extrapolation result.

[0030] Optionally, after S5, the method further includes:

[0031] S6. Based on the instability probability prediction results and the landslide trend extrapolation results, determine the real-time early warning level of the reservoir bank landslide.

[0032] Optionally, S6 specifically includes:

[0033] Based on the instability probability prediction results and the geological displacement parameters, the instability state of the reservoir bank landslide is determined;

[0034] Based on the landslide trend extrapolation results, abnormal offset signals of the pore water pressure data are identified, and based on the abnormal offset signals and the instability state, the real-time early warning level of the reservoir bank landslide is determined.

[0035] On the other hand, the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the landslide instability determination method based on hydrodynamics and pore water pressure as described above.

[0036] The beneficial effects that this invention can produce include:

[0037] This invention provides a landslide instability determination method based on hydrodynamics and pore water pressure. It collects pore water pressure, displacement, and rainfall data from reservoir bank landslides, as well as reservoir water level and flow velocity data. After preprocessing, it obtains soil permeability, water pressure gradient, water scour intensity, and reservoir water level fluctuation data. The method quantifies the dynamic coupling mechanism between these data, constructs a dynamic coupling evolution model of hydrodynamics and pore water pressure, extracts geological displacement parameters and precursor features of abrupt change points from the reservoir bank landslide, and establishes periodic fluctuation and trend acceleration characteristics of reservoir water level, water pressure gradient, and rainfall data to create a linked early warning model. Combining the collected high-frequency time-series data, it predicts instability probability and extrapolates landslide trends, accurately determining the instability state, clarifying the stability level, and quantifying the instability probability. A real-time early warning system is established based on the precursor features of abrupt change points, enabling real-time instability determination and graded response, providing reliable technical support for reservoir bank landslide disaster prevention and control. Attached Figure Description

[0038] Figure 1 A flowchart of a landslide instability determination method based on hydrodynamics and pore water pressure provided in an embodiment of the present invention;

[0039] Figure 2 The diagram shows the structural block of the landslide instability determination system based on hydrodynamics and pore water pressure provided in this embodiment of the invention. Detailed Implementation

[0040] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0041] This invention provides a method for determining landslide instability based on hydrodynamics and pore water pressure, such as... Figure 1 As shown, the method includes:

[0042] S1. Based on the monitoring data of the reservoir bank landslide and the reservoir monitoring data, determine the soil flow characteristics of the reservoir bank landslide and the water level fluctuation data of the reservoir.

[0043] The monitoring data for the reservoir bank landslide includes pore water pressure data, displacement data, and rainfall data; the monitoring data for the reservoir includes reservoir water level data and water flow velocity data; and the soil flow characteristics data includes soil permeability data, water pressure gradient data, and water erosion intensity data.

[0044] S1 specifically includes:

[0045] Based on pore water pressure data and reservoir water level data, the soil permeability data of the reservoir bank landslide body were determined using the improved Darcy's law;

[0046] Based on pore water pressure data, the central difference method was used to determine the water pressure gradient data of the reservoir bank landslide body;

[0047] Based on the water flow velocity data, the Xie Cai formula was used to determine the water flow scouring intensity data of the reservoir bank landslide body;

[0048] Based on reservoir water level data, time series analysis and fluctuation feature extraction methods are used to determine the reservoir water level fluctuation data.

[0049] Prior to S1, the method also includes: collecting monitoring data of reservoir bank landslides and reservoir monitoring data.

[0050] Specifically, high-frequency time-series data were collected on the displacement of the reservoir bank landslide, the pore water pressure of the reservoir bank landslide, the reservoir water level, the water flow velocity of the reservoir, and the rainfall in the area where the reservoir bank landslide is located.

[0051] In practical applications, high-frequency time-series data of displacement and pore water pressure in reservoir bank landslide areas can be collected using distributed fiber optic grating sensor arrays; high-frequency time-series data of reservoir water levels can be collected using radar level gauges and pressure level gauges, with the radar level gauges deployed at the top of the reservoir bank landslide for non-contact monitoring and the pressure level gauges buried at the bottom of the reservoir for contact monitoring; high-frequency time-series data of water flow velocity can be collected using acoustic current profilers; high-frequency time-series data of rainfall can be collected using rain gauges and microwave link rainfall monitoring systems, with rain gauges arranged in a grid pattern in the reservoir bank landslide area and the upstream catchment area of ​​the reservoir, and antenna pairs of the microwave link rainfall monitoring system set up along the reservoir shoreline; and industrial-grade data acquisition devices can be used to perform time-series alignment of displacement, pore water pressure, reservoir water level, water flow velocity, and rainfall.

[0052] Then, based on the theory of unsaturated soil seepage, combined with high-frequency time-series data of pore water pressure and reservoir water level, the soil permeability data were calculated using the improved Darcy's law; based on the high-frequency time-series data of pore water pressure, the water pressure gradient data were calculated using the central difference method; based on the high-frequency time-series data of water flow velocity, the water flow scour intensity data were calculated using the Chezy formula; and based on the high-frequency time-series data of reservoir water level, the water level fluctuation data were calculated using time series analysis combined with fluctuation feature extraction methods.

[0053] The specific steps of implementation S1 include:

[0054] 1. Data collection.

[0055] 1.1 Displacement data and pore water pressure data.

[0056] 1.1.1 Arrange the grating sensor.

[0057] A high-precision distributed fiber Bragg grating sensor array is used to simultaneously monitor the displacement and pore water pressure of the reservoir bank landslide. The sensor array should be deployed at key locations within the landslide, including potential sliding surfaces, fractured zones, and the boundary between stable bedrock, as well as at different depths covering the landslide.

[0058] 1.1.2 Collect displacement data and pore water pressure data.

[0059] The data acquisition frequency is at least once per minute, and each grating sensor node integrates a temperature compensation module to capture instantaneous fluctuations and long-term trend changes in pore water pressure.

[0060] 1.2 Reservoir water level data.

[0061] 1.2.1 Arrangement of radar level gauges and pressure level gauges.

[0062] High-frequency time-series data of reservoir water level are collected by a combination of radar water level gauges and pressure water level gauges. The radar water level gauges are installed at the top of the landslide body on the reservoir bank, while the pressure sensors are buried at the bottom of the reservoir.

[0063] 1.2.2 Collect reservoir water level data.

[0064] Radar level gauges measure reservoir water level data non-contactly, while pressure level gauges measure water pressure through contact and convert it into water level height. The data obtained from both methods are fused using Kalman filtering to obtain the dynamic change curve of the reservoir water level.

[0065] 1.3 Water flow velocity data.

[0066] 1.3.1 Instrument setup.

[0067] High-frequency time-series data of water flow velocity are acquired using an acoustic Doppler current profiler. The acoustic Doppler current profiler is fixed on the measuring vessel or a fixed platform to ensure that the instrument remains stable during the measurement process.

[0068] 1.3.2 Selection of measurement location and cross-section.

[0069] Based on the reservoir's topography and flow characteristics, select representative measurement sections and locations, including the reservoir inlet, outlet, central area, and waters near the landslide. Ensure the measurement sections are perpendicular to the flow direction to obtain accurate flow velocity distribution.

[0070] 1.3.3 Collect water flow velocity data.

[0071] The data acquisition frequency using the acoustic Doppler current profiler was set to at least once per minute, synchronized with the reservoir water level and displacement. Wind speed, wind direction, and water temperature were also recorded during the measurement process for subsequent data analysis.

[0072] 1.4 Rainfall data.

[0073] 1.4.1 Instrument setup.

[0074] A high-precision rain gauge and a microwave link rainfall monitoring system are used to jointly collect high-frequency time-series rainfall data. The rain gauges are arranged in a grid pattern in the landslide area and the upstream catchment area of ​​the reservoir to ensure comprehensive capture of rainfall distribution; the microwave link antenna pairs are set up along the reservoir shoreline to achieve long-distance, high-precision rainfall monitoring; the installation positions of the rain gauges and microwave link antennas are ensured to be reasonable to avoid obstruction and interference.

[0075] 1.4.2 Collect rainfall data.

[0076] The rainfall data collection frequency is set at at least once per minute, synchronized with reservoir water level, displacement, and water flow velocity data; wireless transmission technology is used to transmit the data to the data center in real time for processing and analysis, ensuring the timeliness and integrity of the data.

[0077] 1.5 The synergistic effect of data.

[0078] The above five types of data are collected synchronously at high frequency to form a multi-parameter monitoring network; displacement directly reflects the state of the landslide body; pore water pressure, reservoir water level, and rainfall data reveal the internal hydrodynamic processes of the landslide body; and water flow velocity data reflect the erosive effect of external hydrodynamics on the slope toe.

[0079] 2. Data preprocessing methods and processes.

[0080] 2.1 Calculation of soil permeability data.

[0081] The soil permeability data mentioned above is the soil permeability coefficient. Based on the theory of unsaturated soil seepage, combined with pore water pressure data and reservoir water level data, the improved Darcy's law is used to calculate soil permeability, and the soil permeability coefficient is obtained. :

[0082] ;

[0083] in, For traffic; This represents the seepage path length. This is the cross-sectional area for seepage. For time; This is the head difference (which can be calculated based on the reservoir water level and pore water pressure).

[0084] 2.2 Calculation of water pressure gradient data.

[0085] Based on pore water pressure data, the water pressure gradient is calculated using the central difference method; firstly, the collected pore water pressure data... By performing time alignment and spatial interpolation, a spatially continuously distributed pore water pressure field is obtained;

[0086] The water pressure gradient is obtained by performing a central difference calculation on the pore water pressure field. The calculation formula is:

[0087] ;

[0088] Each component can be calculated using the central difference formula:

[0089] ;

[0090] in, The spatial step size; Use Cartesian coordinates.

[0091] 2.3 Calculation of water flow scour intensity data.

[0092] The collected water flow velocity data By performing time-aligned and spatial averaging, the water flow velocity distribution at different locations was obtained.

[0093] The Chezy formula is used to calculate the scouring intensity of the water flow. :

[0094] ;

[0095] in, The density of water, It is the acceleration due to gravity. For water flow velocity, This is the coefficient for Xie Cai.

[0096] 2.4 Reservoir water level fluctuation data.

[0097] Water level fluctuation characteristics, including fluctuation amplitude, are extracted from the high-frequency time-series data of the reservoir water level. Fluctuation cycle and long-term trends .

[0098] Based on the above water level fluctuation characteristics, a sequence of water level fluctuation data can be generated.

[0099] S2. Based on soil flow characteristics data and water level fluctuation data, construct a dynamic coupled evolution model based on hydrodynamics and pore water pressure.

[0100] S2 specifically includes:

[0101] Spatiotemporal correlation was performed on soil permeability data, water pressure gradient data, water flow scour intensity data, water level fluctuation data, pore water pressure data, reservoir water level data, water flow velocity data, and rainfall data in the area where the reservoir bank landslide was located to obtain spatiotemporal correlation parameters.

[0102] A dynamic coupled evolution model based on hydrodynamics and pore water pressure is constructed based on spatiotemporal correlation parameters.

[0103] Specifically, based on the theory of unsaturated soil seepage and the principles of fluid mechanics, the calculated soil permeability data, water pressure gradient data, water flow scour intensity data, water level fluctuation data, and high-frequency time-series data of pore water pressure, reservoir water level, water flow velocity, and rainfall are spatiotemporally correlated to establish a theoretical framework for the coupling mechanism and obtain spatiotemporal correlation parameters. Based on the spatiotemporal correlation parameters, the Navier-Stokes equations and a set of partial differential equations considering four-dimensional spatiotemporal characteristics are used for modeling to establish a dynamic coupling evolution model based on hydrodynamics and pore water pressure. The linear and nonlinear coupling strength between parameters is quantified by the Pearson-Spearman joint coefficients to verify the coupling of the dynamic coupling evolution model.

[0104] Step S2 during implementation specifically includes:

[0105] 1. A theoretical framework for the coupling mechanism is constructed based on the theory of seepage in unsaturated soil and the principles of fluid mechanics.

[0106] 1.1 Equations and their effects based on the theory of seepage in unsaturated soil.

[0107] Based on the theory of unsaturated soil seepage, the Richards equation is used to describe the changes in pore water pressure in soil over time and space, quantifying the impact of the seepage field on landslide stability. The Richards equation is as follows:

[0108] ;

[0109] in, This refers to the volumetric water content. The unsaturated permeability coefficient; For pressure head; The location is the water head.

[0110] 1.2 Equations and effects based on fluid mechanics principles.

[0111] A continuity equation is used to ensure the spatial and temporal continuity of water flow quality, reflecting the impact of reservoir water level fluctuations on the water flow distribution around the landslide. The continuity equation is as follows:

[0112] ;

[0113] in, The density of water; Darcy velocity of water; Porosity.

[0114] The momentum equation is used to describe the dynamic relationship between water flow velocity and pore water pressure, revealing the force exerted by water erosion on the soil surface of the landslide. The momentum equation is:

[0115] ;

[0116] in, This is the dynamic viscosity of water.

[0117] The continuity equation and momentum equation describe the hydrodynamic process from the perspective of macroscopic water flow, and clarify the physical mechanism of the interaction between water flow velocity, pressure and landslide body.

[0118] 1.3 Construction of the theoretical framework of coupling mechanism.

[0119] This coupling mechanism theoretical framework takes hydrodynamic drive-seepage response-soil deformation as the core chain process, integrates Richards equations from unsaturated soil seepage theory with continuity and momentum equations from fluid mechanics, and constructs a multi-scale closed-loop coupling framework of "microscopic seepage-macroscopic hydrodynamics-soil deformation". Reservoir water level fluctuations and rainfall infiltration serve as external drivers. First, the continuity and momentum equations quantify the scouring force and seepage pressure of water flow on the landslide surface, driving the evolution of the seepage field inside the soil. The spatiotemporal distribution of pore water pressure is described by Richards equations. Changes in the seepage field lead to a decrease in effective soil stress and a reduction in shear strength, causing landslide displacement and deformation. The displacement and deformation are then fed back to the hydrodynamic module by adjusting water flow boundary conditions, such as changes in slope roughness, forming a two-way dynamic coupling mechanism, which comprehensively reveals the physical nature of water-soil interaction in landslide instability.

[0120] 2. Parameter association and data assimilation.

[0121] Discrete data such as soil permeability, water pressure gradient, water flow scour intensity, and water level fluctuation are transformed into continuous spatiotemporal parameters through spatiotemporal difference and data assimilation techniques. This ensures the continuity and consistency of the model input parameters in time and space, and provides accurate initial and boundary conditions for the coupled model.

[0122] 2.1 Spatiotemporal interpolation.

[0123] Pore ​​water pressure was analyzed using Kriging interpolation. Perform spatial interpolation to generate a continuous field:

[0124] ;

[0125] in, The weighting coefficients are determined by minimizing the estimated variance. This represents the number of nearest known points within the search neighborhood.

[0126] 2.2 Data assimilation.

[0127] The reservoir water level was fused using the Kalman filter method. Rainfall The model state is dynamically adjusted based on the observed data and model predictions. The calculation formula is as follows:

[0128] , ;

[0129] in, Let be the state vector of pore water pressure and water level. For the observation vectors of pore water pressure and water level, For process noise, To observe the noise, This is the state transition function. For observation functions.

[0130] 3. Establishment of the coupling model.

[0131] Based on the parameters after spatiotemporal correlation, a dynamic coupled evolution model of hydrodynamics and pore water pressure is established using the Navier-Stokes equations and a set of partial differential equations that consider the four-dimensional spatiotemporal characteristics. Through the dynamic coupled evolution model, the interaction between hydrodynamics and seepage field is dynamically simulated, revealing the influence mechanism of water level fluctuations and scour on landslide stability.

[0132] 3.1 Construction of single-physics field governing equations.

[0133] For each physical field, a set of four-dimensional partial differential equations in spacetime is established based on conservation laws, serving as the "basic module" of the coupled model.

[0134] 3.1.1 Construction of the hydrodynamic field module.

[0135] Driven by reservoir water level fluctuations and rainfall infiltration, and based on fundamental principles of fluid mechanics, a hydrodynamic field module is constructed using the Navier-Stokes equations to quantify the velocity field of the water flow. Pressure field The forces acting on the landslide body not only reflect the physical nature of water flow but also provide a basis for coupling with the seepage module. The Navier-Stokes equations are a four-dimensional system of equations obtained by simultaneously solving the continuity and momentum equations from section 1.2.

[0136] 3.1.2 Construction of the seepage field module.

[0137] Based on the theory of unsaturated soil seepage, the Richards equation is used to describe the evolution of pore water pressure inside the soil. Combined with the hydrodynamic module, the coupling of macroscopic water flow and microscopic seepage is realized.

[0138] 3.1.3 Construction of the mechanical deformation field module.

[0139] Based on the effective stress principle, the deformation of the landslide body is calculated using the displacement control equation. :

[0140] ;

[0141] in, For total stress, As volume forces, both are used as inputs. The equation calculates the shear strength attenuation through effective stress changes and finally outputs the displacement distribution, thus achieving the coupling of seepage and mechanics.

[0142] 3.2 Establish a coupling model.

[0143] The individual physics modules are integrated through coupling variables and feedback paths to form a closed-loop coupled system. Specifically, external drivers such as reservoir water level and rainfall are used as inputs to the hydrodynamic module, and calculations are performed. and The system outputs scouring force and seepage pressure; using the hydrodynamic output as the boundary condition, the effective stress field is calculated using the Richards equation; and using the effective stress field as input, the displacement equation is used to calculate... Then Feedback is sent to the hydrodynamic module to adjust the boundary conditions and repeat the above steps until the simulation ends.

[0144] 4. Model validation.

[0145] The coupling strength of the model is verified by quantifying the linear and nonlinear coupling strength between parameters using the Pearson-Spearman joint coefficients.

[0146] First, calculate the Pearson linear correlation coefficient. and Then calculate the Pearson-Spearman joint coefficient. :

[0147] ;

[0148] in, The weighting coefficients, which combine the linear and nonlinear coupling strengths, are used to verify whether the model accurately reflects the actual physical process through the Pearson-Spearman joint coefficients, thus ensuring the reliability of the model.

[0149] 5. Model output and application.

[0150] The key parameters of the dynamic coupling evolution model are output to provide input for the subsequent multi-parameter linkage early warning model.

[0151] S3. Based on the monitoring data and dynamic coupling evolution model of the reservoir bank landslide, determine the deformation characteristic data of the reservoir bank landslide, and construct an early warning model based on the linkage between hydrodynamics, pore water pressure and displacement according to the deformation characteristic data.

[0152] The deformation characteristic data includes geological displacement parameters and precursor characteristics of abrupt change points of the reservoir bank landslide, as well as periodic fluctuation characteristics and trend acceleration characteristics of reservoir water level data, water pressure gradient data, and rainfall data.

[0153] Based on monitoring data and a dynamic coupled evolution model of the reservoir bank landslide, deformation characteristic data of the reservoir bank landslide in S3 are determined, specifically including:

[0154] The geological displacement parameters of the reservoir bank landslide were determined based on displacement data and a dynamic coupled evolution model.

[0155] Based on pore water pressure data and water flow scour intensity data, the precursor characteristics of abrupt change points in reservoir bank landslides were determined.

[0156] The periodic fluctuation characteristics of reservoir water level data, water pressure gradient data, and rainfall data are determined based on water level fluctuation data, water pressure gradient data, rainfall data, and a dynamic coupled evolution model.

[0157] The trend of acceleration was determined based on the periodic fluctuation characteristics and soil permeability data.

[0158] S3 constructs an early warning model based on the linkage between hydrodynamics, pore water pressure, and displacement using deformation characteristic data. Specifically, this includes:

[0159] Determine the weighting coefficients for geological displacement parameters, precursor characteristics of abrupt change points, periodic fluctuation characteristics, and trend acceleration characteristics;

[0160] Based on geological displacement parameters, precursor characteristics of abrupt change points, periodic fluctuation characteristics, trend acceleration characteristics, and their corresponding weighting coefficients, a linkage early warning model based on hydrodynamics, pore water pressure, and displacement is constructed.

[0161] A dynamic coupled evolution model was used to obtain spatiotemporal dynamic data of reservoir water level fluctuation rate, water pressure gradient data, and infiltration coefficient data of rainfall, and to calculate the periodic fluctuation characteristics of reservoir water level, water pressure gradient data, and rainfall driven by hydrodynamics. The dynamic coupled evolution model was used to obtain the water pressure gradient distribution, and combined with displacement, a Kalman filter algorithm was used for iterative optimization to obtain geological displacement parameters. Based on the periodic fluctuation characteristics, the rate of change of soil permeability data was obtained, and trend acceleration characteristics were quantitatively identified. A cumulative deviation monitoring algorithm was used to continuously track abnormal offset signals of pore water pressure, and based on the abnormal offset signals and water flow scour intensity data, precursor characteristics of abrupt change points were obtained. The periodic fluctuation characteristics, trend acceleration characteristics, precursor characteristics of abrupt change points, and geological displacement parameters were spatiotemporally correlated and multi-parameter fused to construct a linkage early warning model based on hydrodynamics, pore water pressure, and displacement.

[0162] During implementation, step S3 specifically includes:

[0163] 1. Use a dynamic coupling evolution model to obtain feature parameters.

[0164] 1.1 Obtaining periodic fluctuation characteristics.

[0165] Calculate the rate of fluctuation of reservoir water level using model output data. This rate reflects how quickly the water level changes over time; it also acquires spatiotemporal dynamic data of the water pressure gradient. This data describes the spatial and temporal distribution and variation of water pressure; it also analyzes the infiltration coefficient data of rainfall. This coefficient reflects the efficiency with which rainfall is converted into groundwater or soil water; the dominant fluctuation period of the reservoir is determined using Fourier transform or wavelet analysis methods. This cycle reflects the main frequency of water level changes. Calculate the reservoir water level. Water pressure gradient and rainfall The periodic fluctuation characteristics, including the fluctuation amplitude Phase These features quantify the periodic changes of each parameter.

[0166] 1.2 Obtaining geological displacement parameters.

[0167] Combining the long-range distribution data of water pressure gradient output by the model Compared with the measured displacement These data provide direct observations and model predictions of landslide deformation; the Kalman filter gain algorithm is used to minimize the predicted displacement through iterative optimization. The error between the measured displacement and the actual displacement is reduced, thus improving the accuracy of the displacement parameters; Kalman filter gain. :

[0168] ;

[0169] in, It is the prediction error covariance. It is the observation noise covariance.

[0170] Based on the prediction results, the geological displacement parameters of the reservoir bank landslide body were obtained, including the total displacement. Displacement rate These parameters are important indicators for assessing landslide stability.

[0171] 1.3 Trend-based accelerated feature recognition.

[0172] Based on the analysis results of the periodic fluctuation characteristics, the soil permeability data were further calculated. rate of change :

[0173] ;

[0174] This rate reflects the dynamic changes in soil permeability.

[0175] Quantitative analysis was used to identify the trend of acceleration characteristics of the landslide, including the onset time of the acceleration phase. acceleration These characteristics help predict the future development of landslides.

[0176] 1.4 Capturing precursor features of mutation points.

[0177] An accumulated deviation monitoring algorithm is used to continuously track abnormal shift signals in pore water pressure. :

[0178] ;

[0179] in, These are cumulative deviation values. These signals may be precursors to landslide instability. Combined with water scour intensity data... By setting a threshold or using a pattern recognition method, when When the threshold is exceeded, it is determined to be a precursor to a sudden change, providing a basis for timely early warning.

[0180] 2. Construction of an early warning model based on the linkage of hydrodynamics, pore water pressure, and displacement.

[0181] 2.1 Model Architecture Design

[0182] The model employs a hierarchical design, comprising a data layer, a feature layer, an evaluation layer, and a decision layer. The data layer receives and stores real-time monitoring data from various sensors; the feature layer preprocesses the data and extracts key feature parameters; the evaluation layer uses these feature parameters to conduct risk assessment and probability prediction; and the decision layer generates early warning signals and emergency response recommendations based on the evaluation results.

[0183] 2.2 Model feature preprocessing and spatiotemporal correlation.

[0184] 2.2.1 Feature standardization processing.

[0185] Input features of different dimensions and scales are uniformly transformed to the [0,1] interval using the min-max normalization method to ensure that the weights of each feature are balanced in the coupled calculation.

[0186] 2.2.2 Spatiotemporal correlation calibration.

[0187] Missing data were completed using linear interpolation, and timestamps of different features were aligned using a sliding time window to eliminate temporal deviations caused by data transmission and acquisition delays. Then, spatiotemporal correlation was performed; based on the three-dimensional spatial grid division of the reservoir bank landslide body, regional features and point features were spatially matched, with each grid cell corresponding to a complete set of multi-dimensional features, achieving three-dimensional correlation between space, time, and features.

[0188] 2.3 Multi-feature coupling of the model and model establishment.

[0189] 2.3.1 Feature weight allocation.

[0190] The weights of each feature are determined using the analytic hierarchy process (AHP). Based on geological engineering experience and the degree of influence of the features on instability, weight coefficients are set as follows: For example, the weight of periodic fluctuation features is 0.2, the weight of trend acceleration features is 0.3, the weight of geological displacement parameters is 0.3, and the weight of precursor features of abrupt change points is 0.2.

[0191] 2.3.2 Establish a linkage early warning model based on hydrodynamics, pore water pressure, and displacement.

[0192] By integrating standardized features, weighting coefficients, and nonlinear coupling logic, a linkage early warning model is established. The core formula is as follows:

[0193] ;

[0194] in, The risk quantification value output by the model. For the first Weight coefficients of class features For the first Standardized values ​​of class features This is a nonlinear correction function.

[0195] The model achieves dynamic coupling of three parameters—hydrodynamics, pore water pressure, and displacement—through this formula, and directly outputs risk quantification results that can be used for early warning and judgment.

[0196] 3. Model optimization and validation.

[0197] After the model is built, it needs to be rigorously validated and optimized to ensure its early warning performance. First, the model is backtested using known historical landslide event data. By comparing the model's predictions with actual landslide events, the accuracy and reliability of the early warning system can be evaluated. For any problems or deficiencies discovered during validation, the model is optimized and adjusted accordingly. This includes improving feature extraction methods, optimizing network structure, and adjusting parameter settings. Simultaneously, in practical applications, it is crucial to collect new monitoring data and landslide event feedback information to enable the model to learn online and self-optimize.

[0198] S4. Determine the instability warning indicators based on the linkage early warning model, and use the instability warning indicators to predict the instability of the reservoir bank landslide based on the monitoring data of the reservoir bank landslide, and obtain the instability probability prediction results.

[0199] Furthermore, after S4, the method further includes:

[0200] S5. Based on pore water pressure data and displacement data, and using a linkage early warning model, extrapolate the landslide trend of the reservoir bank landslide body to obtain the landslide trend extrapolation results.

[0201] The temporal characteristics of pore water pressure and geological displacement parameters are analyzed using nonlinear dynamic phase space reconstruction technology. Early warning indicators based on the nonlinear characteristics of the linkage early warning model based on hydrodynamics-pore water pressure-displacement are extracted, and these indicators are used to predict the instability probability of reservoir bank landslides. The temporal characteristics of abnormal fluctuations in pore water pressure and deformation of displacement driven by hydrodynamics are analyzed and coupled using deep learning algorithms. The landslide trend is extrapolated by combining the linkage early warning model based on hydrodynamics-pore water pressure-displacement.

[0202] During implementation, steps S4 and S5 specifically include:

[0203] 1. Prediction of the instability probability of landslide bodies.

[0204] 1.1 Nonlinear dynamic phase space reconstruction.

[0205] Construct a multidimensional phase space vector using the delayed coordinate embedding method. , where the embedding dimension The delay time is determined using the pseudo-nearest neighbor method. The optimal value is obtained using the autocorrelation function method. The Lyapunov exponent is calculated based on the reconstructed phase space:

[0206] ;

[0207] Using the calculated Lyapunov index The chaotic characteristics of the quantified system are used as an indicator for early warning of instability.

[0208] 1.2 Calculation of instability probability.

[0209] An LSTM network was constructed to analyze the temporal characteristics of abnormal fluctuations in pore water pressure and geological displacement parameters; the network input layer received high-frequency time-series data. Through nonlinear mapping of the forget gate, input gate, and output gate, a coupled feature vector is output. By combining dynamic Bayesian networks, periodic fluctuation characteristics and trend acceleration characteristics are probabilistically coupled to generate an initial distribution of instability probability. , Dynamic correction based on Bayesian update mechanism, the formula is:

[0210]

[0211] in, It is in an unstable state. For a safe state, To observe feature vectors in real time.

[0212] Based on the initial distribution of instability probability This will quantify the instability risk of landslides and provide a scientific basis for early prevention and control.

[0213] 2. Extrapolation of landslide trends.

[0214] 2.1 Quantitative analysis of uncertainty.

[0215] Uncertainty analysis was performed using Monte Carlo simulation, and an early warning model based on a multi-parameter linkage of hydrodynamics, pore water pressure, and displacement was used to assess pore water pressure. Reservoir water level fluctuations Random sampling of key parameters is performed to generate Parameter combinations; calculate the instability probability of each parameter combination using the above method, and statistically obtain the probability distribution. Quantify the uncertainty of forecasting.

[0216] 2.2 Adjustment of critical mutation points and trend extrapolation.

[0217] Calculate the displacement rate by combining the precursor characteristics of the abrupt change point. Second derivative:

[0218] ;

[0219] in, These are geological displacement parameters.

[0220] When displacement rate The second derivative exceeds the threshold or Lyapunov exponent. Exceeding the critical value When the warning signal is upgraded, the extrapolation trend curve is dynamically adjusted to extrapolate the landslide trend, accurately predict the deformation trend and potential risks of the landslide body, and support the optimization of long-term risk assessment and emergency response strategies.

[0221] Following S5, the method further includes:

[0222] S6. Determine the real-time early warning level of the reservoir bank landslide based on the instability probability prediction results and the landslide trend extrapolation results.

[0223] S6 specifically includes:

[0224] Based on the instability probability prediction results and geological displacement parameters, the instability state of the reservoir bank landslide was determined;

[0225] Based on the extrapolation results of landslide trends, abnormal offset signals of pore water pressure data are identified, and the real-time early warning level of the reservoir bank landslide is determined according to the abnormal offset signals and the instability state.

[0226] Based on the dynamic coupling evolution model, the probability of instability is quantified and the instability state is predicted and determined. Based on the instability state, the geological displacement parameters caused by hydrodynamic mutations are identified by extrapolating the landslide trend, and the abnormal offset signal of pore water pressure is identified by the precursor features of the mutation point. Based on the instability state and the abnormal offset signal of pore water pressure, a real-time early warning system is established to conduct real-time determination and graded response to the instability of the reservoir bank landslide.

[0227] Specifically, the real-time determination and graded response to instability states include: when the instability probability is ≥70% and the geological displacement parameter reaches the threshold, it is determined to be an impending instability state; when 50% ≤ instability probability < 70% and the geological displacement parameter does not reach the threshold, it is determined to be a potential instability state; when the instability probability is < 50% and there is no abnormal offset signal and the geological displacement parameter does not reach the threshold, it is determined to be a stable state. Real-time early warning rules are established by combining abnormal offset signals and instability states: if a stable state is determined, no warning is triggered; if a potential instability state is determined and no abnormal offset signal is detected, a yellow warning is triggered; if a potential instability state is determined and an abnormal offset signal is detected, an orange warning is triggered; if an impending instability state is determined, a red warning is triggered, and a real-time response is performed according to the real-time early warning rules.

[0228] During implementation, step S6 specifically includes:

[0229] 1. Determination of the instability state of the landslide body.

[0230] Based on the instability probability distribution obtained in S4 Combined with displacement rate Real-time and displacement acceleration Multi-parameter coupling is achieved through dynamic Bayesian networks. When the instability probability... Instability ≥70% and When the instability is imminent, it is determined to be in an "imminent unstable state"; if 50% ≤ P instability < 70% and If so, it is determined to be a "potentially unstable state".

[0231] 2. Correlation analysis between landslide instability and abnormal pore water pressure offset signals.

[0232] Based on the abnormal shift signal of pore water pressure, combined with the intensity of water flow scouring, it is directly used as a precursor feature of the critical abrupt change point, combined with displacement acceleration. To verify the correlation between the abnormal offset signal and the instability probability, the instability state determination result and the abnormal offset signal of pore water pressure were dynamically coupled and analyzed. The instability probability was updated through a dynamic Bayesian network. This enables closed-loop verification of the decision logic.

[0233] 3. Real-time early warning system and tiered response execution.

[0234] Based on different instability states and different abnormal shift signals in pore water pressure, a real-time early warning system is established to implement graded responses, specifically including:

[0235] Yellow alert: Potential instability state and no abnormal offset signal exceeding the threshold has been detected;

[0236] Orange alert: Potential instability accompanied by abnormal offset signals exceeding the threshold;

[0237] Red alert: The state of instability is imminent or signs of a critical mutation point have been detected.

[0238] The warning level is based on a dynamic comparison between real-time data and preset thresholds, combined with the correlation analysis of abnormal offset signals, to achieve accurate classification of the warning level and ensure that the response measures are highly matched with the risk level.

[0239] Another embodiment of the present invention provides a reservoir bank landslide instability determination system based on hydrodynamic effects and dynamic evolution of pore water pressure, such as... Figure 2 As shown, the system includes:

[0240] The data acquisition and processing module is used to collect high-frequency time-series data on the displacement of the reservoir bank landslide, the pore water pressure of the reservoir bank landslide, the reservoir water level, the water flow velocity of the reservoir, and the rainfall in the area where the reservoir bank landslide is located. It also preprocesses the collected high-frequency time-series data to obtain soil permeability data, water pressure gradient data, water flow scour intensity data, and reservoir water level fluctuation data of the reservoir bank landslide.

[0241] The coupled evolution model construction module is connected to the data acquisition and processing module. It is used to quantify the dynamic coupling mechanism between soil permeability data, water pressure gradient data, water flow scour intensity data and water level fluctuation data of the reservoir bank landslide body, and to construct a dynamic coupled evolution model based on hydrodynamics and pore water pressure.

[0242] The early warning model construction module is connected to the coupled evolution model construction module. It is used to obtain the geological displacement parameters, precursor characteristics of abrupt change points, periodic fluctuation characteristics of reservoir water level, water flow velocity and rainfall, and trend acceleration characteristics of reservoir bank landslides driven by hydrodynamics using the dynamic coupled evolution model. It establishes a linkage early warning model based on hydrodynamics-pore water pressure-displacement.

[0243] The instability prediction module, connected to the early warning model construction module, is used to predict the instability probability and extrapolate the landslide trend of the reservoir bank landslide based on the linkage early warning model and by combining the collected high-frequency time series data. Based on the instability probability prediction and landslide trend extrapolation results, the module determines the instability state of the reservoir bank landslide, quantifies the corresponding instability probability, and combines the precursor characteristics of abrupt change points to make real-time judgments and graded responses to the instability state.

[0244] Specifically, the instability prediction module includes;

[0245] The instability probability prediction submodule is used to analyze the temporal characteristics of pore water pressure and geological displacement parameters based on nonlinear dynamic phase space reconstruction technology, extract early warning indicators that reflect the nonlinear characteristics of the system, and use the early warning indicators to predict the instability probability of the landslide body;

[0246] The landslide trend extrapolation submodule is used to analyze and couple the temporal characteristics of abnormal fluctuations in pore water pressure and displacement deformation under hydrodynamic driving through deep learning algorithms, and to extrapolate landslide trends by combining a linkage early warning model based on hydrodynamics-pore water pressure-displacement.

[0247] The landslide instability determination submodule is used to determine the instability state of the reservoir bank landslide based on the instability probability prediction and landslide trend extrapolation results, simultaneously quantify the corresponding instability probability, and combine the precursor characteristics of abrupt change points to perform real-time determination and graded response of the instability state.

[0248] The specific implementation includes:

[0249] 1. Connection and implementation of the data acquisition and processing module.

[0250] 1.1 Physical connection.

[0251] The distributed fiber optic grating sensor array is connected via FC / APC fiber optic interface; the radar level gauge and pressure level gauge are connected via RS485 and 4-20mA analog interfaces respectively to the general-purpose I / O port; the acoustic current profiler is connected via RJ45 Ethernet interface and TCP / IP protocol; all the above instruments are connected to an industrial-grade data acquisition unit, which achieves timing alignment through NTP clock synchronization.

[0252] 1.2 Function Implementation.

[0253] Data such as displacement, pore water pressure, and reservoir water level are collected at a frequency of ≥1 time / minute, and the sensor node integrates a temperature compensation module; based on the unsaturated soil seepage theory and the central difference method, soil permeability and water pressure gradient data are calculated.

[0254] 2. Connection and implementation of the coupled evolution model construction module.

[0255] 2.1 Physical connection.

[0256] The model is built using an edge computing server, which is connected to the coupled evolution model building module via a gigabit Ethernet RJ45 interface.

[0257] 2.2 Function Implementation.

[0258] Coupling Mechanism Construction: Based on the theory of unsaturated soil seepage and the principles of fluid mechanics, a coupling framework of "microscopic seepage-macroscopic hydrodynamics-soil deformation" is established.

[0259] Modeling: The model is constructed using the Navier-Stokes equations and a set of four-dimensional spatiotemporal partial differential equations, and the coupling strength is quantified using the Pearson-Spearman joint coefficients.

[0260] 3. Connection and implementation of the early warning model construction module.

[0261] 3.1 Physical connection.

[0262] The early warning model building module is connected to the GPU computing node of the edge computing server via PCIe 4.0 direct board connection.

[0263] 3.2 Function Implementation.

[0264] Feature extraction: Obtain multi-dimensional features such as geological displacement parameters and precursor features of abrupt change points;

[0265] Data verification: Data integrity is verified using CRC32 checksum, and retransmission is triggered in case of an anomaly.

[0266] 4. Instability prediction module.

[0267] 4.1 Instability Probability Prediction Submodule.

[0268] 4.1.1 Physical connection.

[0269] The instability probability prediction submodule is directly connected to the early warning model construction module on the PCIe 3.0 high-speed interface board. It directly reads the nonlinear early warning indicators and multi-feature fusion results output by the linkage early warning model through memory mapping.

[0270] 4.1.2 Function implementation.

[0271] Temporal Feature Analysis: Based on nonlinear dynamic phase space reconstruction technology, the temporal data of pore water pressure and geological displacement parameters are transformed into high-dimensional phase space trajectories, and the chaotic characteristics of the Lyapunov exponential quantization system are calculated.

[0272] Probability calculation: By combining dynamic Bayesian networks, periodic fluctuation features and trend acceleration features are probabilistically coupled to generate an initial distribution of instability probability, and the probability values ​​are dynamically corrected through a Bayesian update mechanism.

[0273] 4.2 Landslide Trend Extrapolation Submodule.

[0274] 4.2.1 Physical connection.

[0275] The landslide trend extrapolation submodule is directly connected to the instability probability prediction submodule within the PCIe 3.0 high-speed interface board, and directly reads the initial distribution of instability probability output by the instability probability prediction submodule through memory mapping.

[0276] 4.2.2 Function implementation.

[0277] Feature Coupling: The time-series features of abnormal fluctuations in pore water pressure and displacement deformation driven by hydrodynamics are analyzed and coupled using an LSTM network to extract key coupling factors.

[0278] Trend extrapolation: Combining the linkage early warning model based on hydrodynamics-pore water pressure-displacement, and using Monte Carlo simulation to quantify the uncertainty of prediction, a landslide trend curve is generated.

[0279] Dynamic adjustment: The prediction results are updated in real time using the Kalman filter algorithm, when the second derivative of the displacement rate... When the threshold is exceeded, the curve is corrected, triggering an upgraded warning signal.

[0280] 4.3 Landslide Instability Judgment Submodule.

[0281] 4.3.1 Physical connection.

[0282] This module uses dual redundant Ethernet interfaces to communicate with the instability probability prediction submodule, the landslide trend extrapolation submodule, and the core decision server, supporting automatic switching between primary and backup links to ensure communication continuity. During communication, the primary link uses the TCP / IP protocol, and the backup link uses the UDP protocol.

[0283] 4.3.2 Function implementation.

[0284] State determination: Based on the instability probability and landslide trend extrapolation results, combined with geological displacement parameters and pore water pressure anomaly migration signals, three stability levels are defined. When the instability probability is less than 50% and there are no abnormal signals, it is determined to be a stable state. When the instability probability is between 50% and 70% and the geological displacement parameters have not reached the threshold, it is determined to be a potential instability state. When the instability probability is not less than 70% and the geological displacement parameters have reached the threshold, it is determined to be an impending instability state.

[0285] Warning generation: Match the corresponding warning signal according to the stability level. No warning is triggered in a stable state. A yellow warning is triggered when no abnormal offset signal is detected in a potential unstable state. An orange warning is triggered when an abnormal offset signal is detected in a potential unstable state. In other words, a red warning is triggered when an unstable state is in a stable state.

[0286] Tiered Response: Yellow alert triggers encrypted monitoring, with data collection frequency increased to once every 30 seconds; Orange alert triggers on-site inspections, sets up warning signs, and adjusts reservoir scheduling plans to mitigate water level fluctuations; Red alert immediately activates personnel evacuation plans, closes traffic in landslide-affected areas, and coordinates emergency rescue forces to stand by.

[0287] This system is designed for monitoring the instability of reservoir bank landslides. It utilizes multiple sensors to collect high-frequency data on displacement and pore water pressure. After preprocessing, a dynamic coupling model of hydrodynamics and pore water pressure is constructed, and multi-dimensional features are extracted to establish a coordinated early warning system. The instability prediction module performs probability prediction and trend extrapolation, while the instability state determination and graded response module classifies stability levels, triggering yellow, orange, and red level early warnings and corresponding response measures. The entire process is automated and highly accurate, providing reliable technical support for landslide disaster prevention and control.

[0288] Another embodiment of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the landslide instability determination method based on hydrodynamics and pore water pressure as described above.

[0289] The above descriptions are merely a few embodiments of the present invention and are not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any modifications or alterations made by those skilled in the art without departing from the scope of the technical solution of the present invention using the disclosed technical content are equivalent to equivalent implementation cases and fall within the scope of the technical solution.

Claims

1. A method for determining landslide instability based on hydrodynamics and pore water pressure, characterized in that, The method includes: S1. Based on the monitoring data of the reservoir bank landslide and the reservoir monitoring data, determine the soil flow characteristics of the reservoir bank landslide and the water level fluctuation data of the reservoir. S2. Based on the soil flow characteristic data and the water level fluctuation data, construct a dynamic coupled evolution model based on hydrodynamics and pore water pressure; S3. Based on the monitoring data of the reservoir bank landslide and the dynamic coupling evolution model, determine the deformation characteristic data of the reservoir bank landslide, and construct an early warning model based on the linkage between hydrodynamics, pore water pressure and displacement based on the deformation characteristic data; S4. Determine the instability warning index based on the linkage warning model, and use the instability warning index to predict the instability of the reservoir bank landslide based on the monitoring data of the reservoir bank landslide, and obtain the instability probability prediction result.

2. The method according to claim 1, characterized in that, The monitoring data of the reservoir bank landslide includes pore water pressure data, displacement data, and rainfall data of the reservoir bank landslide; the monitoring data of the reservoir includes reservoir water level data and water flow velocity data. The soil flow characteristics data include soil permeability data, water pressure gradient data, and water scour intensity data.

3. The method according to claim 2, characterized in that, S1 specifically includes: Based on the pore water pressure data and the reservoir water level data, the soil permeability data of the reservoir bank landslide body were determined using the improved Darcy's law; Based on the pore water pressure data, the water pressure gradient data of the reservoir bank landslide body is determined using the central difference method; Based on the water flow velocity data, the water flow scouring intensity data of the reservoir bank landslide body is determined using the Xie Cai formula. Based on the reservoir water level data, time series analysis and fluctuation feature extraction methods are used to determine the water level fluctuation data of the reservoir.

4. The method according to claim 2, characterized in that, S2 specifically includes: The soil permeability data, water pressure gradient data, water flow scour intensity data, water level fluctuation data, pore water pressure data, reservoir water level data, water flow velocity data, and rainfall data in the area where the reservoir bank landslide is located are spatiotemporally correlated to obtain spatiotemporal correlation parameters. A dynamic coupled evolution model based on hydrodynamics and pore water pressure is constructed based on the spatiotemporal correlation parameters.

5. The method according to claim 2, characterized in that, The deformation characteristic data includes geological displacement parameters and precursor characteristics of abrupt change points of the reservoir bank landslide, as well as periodic fluctuation characteristics and trend acceleration characteristics of reservoir water level data, water pressure gradient data, and rainfall data; The step S3, which determines the deformation characteristic data of the reservoir bank landslide based on the monitoring data and the dynamic coupling evolution model, specifically includes: The geological displacement parameters of the reservoir bank landslide are determined based on the displacement data and the dynamic coupling evolution model. Based on the pore water pressure data and the water flow scour intensity data, the precursor characteristics of the sudden change point of the reservoir bank landslide were determined; The periodic fluctuation characteristics of the reservoir water level data, the water pressure gradient data, the rainfall data, and the dynamic coupling evolution model are determined based on the water level fluctuation data, the water pressure gradient data, the rainfall data, and the dynamic coupling evolution model. The trend acceleration characteristics are determined based on the periodic fluctuation characteristics and the soil permeability data.

6. The method according to claim 5, characterized in that, The step S3, which involves constructing a linkage early warning model based on the deformation feature data and the interaction between hydrodynamics, pore water pressure, and displacement, specifically includes: Determine the weighting coefficients for the geological displacement parameters, the precursor features of the abrupt change point, the periodic fluctuation features, and the trend acceleration features; Based on the geological displacement parameters, the precursor characteristics of the abrupt change point, the periodic fluctuation characteristics, the trend acceleration characteristics, and their corresponding weighting coefficients, a linkage early warning model based on hydrodynamics, pore water pressure, and displacement is constructed.

7. The method according to claim 5, characterized in that, Following S4, the method further includes: S5. Based on the pore water pressure data and the displacement data, and using the linkage early warning model, extrapolate the landslide trend of the reservoir bank landslide body to obtain the landslide trend extrapolation result.

8. The method according to claim 7, characterized in that, Following S5, the method further includes: S6. Based on the instability probability prediction results and the landslide trend extrapolation results, determine the real-time early warning level of the reservoir bank landslide.

9. The method according to claim 8, characterized in that, S6 specifically includes: Based on the instability probability prediction results and the geological displacement parameters, the instability state of the reservoir bank landslide is determined; Based on the landslide trend extrapolation results, abnormal offset signals of the pore water pressure data are identified, and based on the abnormal offset signals and the instability state, the real-time early warning level of the reservoir bank landslide is determined.

10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the landslide instability determination method based on hydrodynamics and pore water pressure as described in any one of claims 1 to 9.

Images