Landslide surge physical and numerical hybrid model construction method, system and application

Abstract

The invention relates to the technical field of hydraulic engineering, geotechnical mechanics and geological disaster prevention and control engineering, and discloses a landslide surge physical and numerical hybrid model construction method and system and application. Analyzing the landslide instability evolution motion process, the landslide warehousing first wave height and the surge propagation process of the landslide surge; performing a numerical simulation test according to a second result of the landslide surge numerical model test; adjusting the particle size and the kernel function in the landslide surge numerical simulation test, so that the first wave height, the opposite-bank arrival height and the surge pressure distribution in the landslide surge numerical simulation test and the landslide surge physical model test are consistent; and taking the adjusted landslide surge numerical model as the obtained landslide surge hybrid model. The system comprises a first result acquisition module, a second result acquisition module, a test parameter adjustment module and a hybrid model acquisition module. According to the method, different calculation working conditions can be conveniently analyzed.

Description

Technical Field

[0001] The present invention relates to the technical fields of water conservancy engineering, rock and soil mechanics and geological disaster prevention and control engineering, and in particular to a method, system and application for constructing a landslide surge physical and numerical hybrid model. Background Art

[0002] Landslide surge is a huge wave phenomenon caused by landslides entering the water body, and its hazards cannot be ignored. The hazards include huge impact force caused by landslide surge, which will cause serious damage to buildings, infrastructure and ecological environment in coastal areas; the huge waves generated can quickly submerge low-lying areas along the coast, affect the navigation of ships, and cause property losses and casualties. In order to accurately reflect the characteristics of landslide surge, the landslide surge physical model test constructs a landslide surge physical model that meets the engineering geological characteristics, and its results have important guiding significance for the prevention and control decision-making of landslide surge. However, due to the limitation of the number of monitoring equipment and the influence of terrain accuracy, the results of the landslide surge physical model test mainly reflect the macroscopic characteristics of landslide surge disasters, and it is difficult to accurately capture the various details of landslide surge disasters. At the same time, the use of numerical simulation methods to study landslide surge problems has the advantages of high efficiency and intuitive results, which can clearly depict the details of landslide surge disasters; however, the calculation results of the numerical simulation method are greatly affected by the calculation parameters, and some parameters are often difficult to determine, resulting in relatively low credibility of the numerical simulation results.

[0003] Prior art 1, Chinese patent, application number: 202410623079.7 discloses a test device and test method for landslide surge with overall adjustment of the bank slope, including a movable lifting test platform, an angle adjustment slope, and a speed measuring pulley; the movable lifting test platform includes a movable frame, a hydraulic lifting test platform, and a load-bearing support plate; the angle adjustment slope includes a slope, an angle adjustment cylinder, a rotating support seat, an angle measuring plate, and a buffer block; the speed measuring pulley includes a slide box and a speed measuring device. It also involves a test method, including the steps of: building a water body model; simulating the process of landslide surge; collecting experimental data and processing experimental data. Changing multiple test conditions to conduct landslide surge physical model tests; adapting to test pools of different heights; being suitable for landslide models with various sliding inclination angles; changing the impact speed of the sliding body; by replacing the slide box, it can be applied to landslide surge tests of different sliding body materials; although it saves the time cost of the test; it is repeatedly used in different test models. However, it has extremely high requirements on the performance and model of the detection equipment, which increases the cost of the test.

[0004] Prior art 2, Chinese patent, application number: 202210105273.7 discloses a normal physical model of a three-dimensional landslide surge test in a river-type reservoir, including a wave-breaking area, and a landslide area, a river area, and a dam area set up by integrating engineering geological map information and hydraulic structure design drawing information; the dam area includes a hydraulic structure with a drainage gate; the landslide area includes landslide body-similar materials and a landslide initiation device; a water flow circulation system for simulating real water flow conditions is provided between the wave-breaking area, the landslide area, the river area, and the dam area, and also includes a data monitoring system for monitoring the occurrence and propagation of landslide surges throughout the process. Although the landslide surge test physical model is a large-scale physical model, the overall normal model designed based on engineering geological map data and hydraulic structure design drawing data can more completely reflect the entire process of landslide-river channel-high dam joint action, and can more accurately reflect the disaster phenomena and data laws of real projects; it can be used to establish a large-scale three-dimensional landslide surge physical model test for real projects. However, it mainly reflects the macroscopic characteristics of landslide surge disasters and is difficult to accurately capture the various details of landslide surge disasters, which increases the difficulty of accurately capturing the various details of landslide surge disasters.

[0005] Prior art three, Chinese patent, application number: 202410555040.6 discloses a method and system for numerical simulation of landslide surge based on nested grids, including obtaining terrain data of the area to be simulated, and establishing a three-dimensional geometric model based on the terrain data; dividing the three-dimensional geometric model into grid nested areas and non-nested areas based on the terrain information of the area to be simulated; performing a first grid division on the grid nested areas and non-nested areas of the three-dimensional geometric model, and then performing a second grid division on the grid nested areas; setting simulation model parameters and initial conditions and performing simulation to obtain simulated values ​​of parameters in landslide surges. Although numerical simulation calculations are realized by dividing the area to be simulated into regions and nesting grids based on regions, the number of grids is greatly reduced, the consumption of computing resources is reduced, and the computing efficiency is improved; and based on the region, grid nesting can set a more refined computing grid to adapt to complex terrain conditions and ensure the accuracy of the calculation results. However, the calculation results are too dependent on the calculation parameters, resulting in relatively low credibility and accuracy of the numerical simulation results.

[0006] At present, the existing technologies 1, 2 and 3 are difficult to accurately capture the various details of landslide surge disasters, and the credibility of numerical simulation results is relatively low. Therefore, the present invention provides a landslide surge physical and numerical hybrid model construction method, system and application, combining the advantages of physical model experiments and numerical simulations, based on large-scale physical model experiments, using physical experimental results to iteratively correct the numerical model, establish a physical-numerical hybrid model, and comprehensively evaluate the data of landslide surge disasters. Summary of the invention

[0007] The main purpose of the present invention is to provide a landslide surge physical and numerical hybrid model construction method, system and application to solve the problem that it is difficult to accurately capture various details of landslide surge disasters in the prior art and the credibility of numerical simulation results is relatively low.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for constructing a landslide surge physical and numerical hybrid model, the method comprising:

[0010] Obtain the contour geological information of the target area, the height of the center of gravity of the sliding body and the water level information; construct the landslide surge physical model test based on the information, which includes constructing the terrain based on the contour information and constructing the position and volume of the sliding body based on the volume and center of gravity of the sliding body; analyze the landslide instability evolution movement process, the height of the first wave entering the landslide reservoir and the surge propagation process of the landslide surge based on the first result of the landslide surge physical model test;

[0011] Constructing a landslide surge numerical model test based on the information, the landslide surge numerical model test includes constructing terrain based on contour information and constructing the position and volume of the sliding body based on the volume and center of gravity of the sliding body; conducting a numerical simulation test based on the second result of the landslide surge numerical model test;

[0012] Taking the first wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, the particle size and kernel function in the numerical simulation test of landslide surge were adjusted to make the first wave height, the height reaching the opposite bank and the surge pressure distribution in the numerical simulation test of landslide surge and the physical model test of landslide surge consistent.

[0013] The adjusted landslide surge numerical model is used as the obtained landslide surge hybrid model to analyze the working conditions of different landslide velocities, landslide volumes and terrain conditions.

[0014] As a further improvement of the present invention, the process of adjusting the particle size and the kernel function in the landslide surge numerical simulation test includes:

[0015] Analyze the experimental data obtained, including extracting the specific values ​​of the head wave height, the height reaching the opposite shore, and the surge pressure distribution, which will be used as the evaluation criteria; establish a relationship diagram between the target value and the current numerical simulation results, and identify the gap between the numerical simulation and the physical experimental results;

[0016] Analyze the particle distribution characteristics in the landslide surge physical model and compare the effects of different particle sizes on the numerical results. Perform numerical simulations by changing the particle diameter and continuously refine the particle size until it matches the results of the landslide surge physical model. Based on the analysis of surge behavior in the landslide surge physical model experiment, confirm the kernel function type and adjust the support domain and attenuation rate of the kernel function.

[0017] After completing the adjustment of the particle size and kernel function, a new round of numerical simulation is carried out: the adjusted numerical model is comprehensively compared with the physical experimental results, including the verification of the first wave height, the detection of the height reaching the opposite shore, and the fitting of the surge pressure distribution.

[0018] As a further improvement of the present invention, the process of establishing a relationship diagram between the target value and the current numerical simulation result includes:

[0019] Obtain multiple key parameters such as the first wave height, the height reaching the opposite shore, and the surge pressure distribution for setting target values; at the same time, obtain the corresponding results from the numerical simulation and record the first wave height, the height reaching the opposite shore, and the corresponding surge pressure distribution data obtained under the same conditions;

[0020] The collected landslide surge physical model experimental data and the second result of numerical simulation are matched with the time point and spatial position of collection, and smoothed; the height of the first wave and the height reaching the opposite bank are normalized according to the maximum value, respectively, to generate a standardized value between 0 and 1;

[0021] A two-dimensional coordinate system is used, with the X-axis representing the simulation result and the Y-axis representing the experimental target value. A scatter plot is made to show the relationship between the simulation result and the target value. The distance of each point is calculated to quantify the gap and obtain the deviation between the numerical simulation result and the target value.

[0022] As a further improvement of the present invention, there are n pairs of simulation results and target values ​​of key parameters, which are respectively expressed as: numerical simulation results S i , where i = 1, 2, ..., n, target value T i ;

[0023] Absolute error refers to the direct difference between the numerical simulation result and the target value, and the calculation formula is:

[0024] E abs,i =|S i -T i |

[0025] Among them, E abs,i represents the absolute error of the i-th key parameter;

[0026] The relative error takes into account the impact of the target value and uses the ratio of the target value to standardize the error. The calculation formula is:

[0027]

[0028] Among them, E rel,i Represents the relative error of the i-th key parameter;

[0029] The root mean square error RMSE is used to comprehensively evaluate the deviations of multiple parameters. The calculation formula is:

[0030]

[0031] The average of the squared errors for all key parameters was calculated and then the square root was taken, providing a more comprehensive error measure;

[0032] The comprehensive deviation evaluation evaluates each key parameter together with the absolute error, relative error and root mean square error to form a new indicator:

[0033]

[0034] Among them, α, β, and γ are weight coefficients, reflecting the importance of different deviation measures in the analysis.

[0035] As a further improvement of the present invention, the process of changing the particle diameter and adjusting the support domain and decay rate of the kernel function includes:

[0036] Extract the distribution characteristics and diameter of the current particles, use the landslide surge numerical model to perform regression analysis on the relationship between the existing particle distribution and surge behavior, and obtain preliminary results; set the range of particle size variation, and select multiple particle diameters for simulation experiments; in the numerical simulation, replace the corresponding particle sizes one by one, and focus on the first wave height, the wave height reaching the shore, and the pressure distribution area in the simulation results;

[0037] After each adjustment of the particle diameter, numerical simulations were repeated to compare the simulation results with the experimental target values. By collecting the simulated data, the accuracy of the corresponding wave height, arrival height and pressure distribution was evaluated, and the particle size was reduced until it matched the target value.

[0038] After the particle size is determined, the support domain of the kernel function is set according to the distance of the particle. According to the actual fluctuations in the simulation results, the decay rate of the kernel function is adjusted, the initial decay rate is set, and fine adjustments are made by comparing the simulation results.

[0039] As a further improvement of the present invention, the process of using the landslide surge numerical model to perform regression analysis on the relationship between the existing particle distribution and surge behavior includes:

[0040] Obtain various data on the particle's trajectory, position, diameter, and water surface fluctuations in the surge, analyze the acquired data, and automatically mark and identify the particle's boundaries; use a tree-based model to evaluate the impact of each variable on surge behavior and select the most influential features; construct cross-features of particle characteristics and wave behavior to generate a new feature combination, namely, wave-particle coupling features;

[0041] Construct a neural network based on deep learning, use adaptive network structure and hierarchical feature extraction mechanism to fit the complex nonlinear relationship between particle characteristics and surge, analyze the characteristics of fluctuating data series, and capture dynamic changes in time series; combine fluid dynamics equations with deep learning models to guide the learning of neural networks through physical constraints;

[0042] The performance of the neural network model is evaluated through k-fold cross-validation, and the hyperparameters are adjusted using Bayesian optimization to achieve the optimal configuration.

[0043] As a further improvement of the present invention, the process of guiding the learning of the neural network by physical constraints includes:

[0044] Collect particle dynamics and water surface fluctuation data, annotate particles to identify the boundaries, shapes and positions of different particles; extract particle characteristics and wave parameters; create time window features to convert all particle characteristics and wave parameters into distributions with a mean of 0 and a variance of 1;

[0045] The particle dynamics and water surface fluctuation data are divided into training sets and validation sets. The loss and accuracy curves of the training set and validation set are plotted to observe the convergence and stability. The effectiveness of the neural network model is evaluated through the scatter plot of the actual wave data and the model prediction results.

[0046] Establish a set of model equations to describe fluid dynamics, and calculate the loss through the residual of the neural network model output and the equation; after the neural network model training is completed, use physical consistency testing to check whether the wave and particle behaviors output by the model meet the physically based expectations;

[0047] in,

[0048]

[0049] Where ρ is the fluid density, u is the velocity field, p is the pressure, μ is the viscosity coefficient, and f is the body force;

[0050] Calculate the loss function expression:

[0051] Loss = L data +λ·L physics

[0052] Among them, Ldata is the data loss, L physics is the loss of the physical model, and λ is the importance coefficient used to balance the loss.

[0053] As a further improvement of the present invention, the process of reducing the particle size until it matches the target value includes:

[0054] Collect the output results of each numerical simulation, which include three parameters: wave height, arrival height and pressure distribution. The mean square error and absolute error are used as standardized evaluation indicators.

[0055] Set the accuracy threshold and judge the deviation of the three parameters of wave height, arrival height and pressure distribution according to the accuracy threshold; compare the simulation results with the target values, generate error analysis charts, obtain the deviation and trend of the model output under different particle sizes, and judge whether the influence of the current particle size on the surge behavior meets the experimental objectives;

[0056] If the simulation results fail to meet the target value, the particle size is proportionally reduced for the next round of simulation;

[0057] The process of proportionally reducing the particle size for the next round of simulation includes:

[0058] After comparing the simulation results with the target values, the deviation of each parameter is identified;

[0059] Definition symbol: Y sim =(h sim ,z sim ,P sim ) represents the current simulation result vector, h sim represents the simulated wave height, z sim represents the simulated arrival height, P sim represents the simulated pressure distribution, Y target =(h target ,z target ,P target ) represents the target value vector, h target represents the target wave height, z target Indicates the target arrival height, P target represents the target pressure distribution, D i Represents the deviation of the i-th parameter, calculated by:

[0060] where i∈h,z,P

[0061] i = h represents the deviation for wave height, i = z represents the deviation for arrival height, and i = P represents the deviation for pressure distribution;

[0062] Set the comprehensive deviation D to;

[0063] Definition symbol: N represents the number of data points; calculation formula:

[0064]

[0065] ∈The set accuracy threshold, indicating the maximum allowed deviation);

[0066] Reduce the particle size proportionally, and adjust the particle size after identifying that the deviation D exceeds the accuracy threshold;

[0067] Definition symbol: D current Indicates the current particle size value, D new represents the updated new particle size value, r represents the reduction ratio, and the range is r∈(0,1);

[0068] The formula for updating particle size is: D new =D current ×(1-r);

[0069] For example: If the current particle size (D current =0.05) meters, reduce the ratio (r = 0.1), then: D new = 0.05 × (1-0.1) = 0.045 m;

[0070] After adjusting the particle size, a new numerical simulation is performed to generate a new output result Y sim,new ; Return to re-collect data for evaluation.

[0071] To achieve the above object, the present invention also provides the following technical solutions:

[0072] A landslide surge physical and numerical hybrid model construction system, which is applied to the landslide surge physical and numerical hybrid model construction method, and the landslide surge physical and numerical hybrid model construction system includes:

[0073] The first result acquisition module is used to obtain the contour geological information of the target area, the height of the center of gravity of the sliding body and the water level; construct the landslide surge physical model test according to the information, and the landslide surge physical model test includes constructing the terrain according to the contour information and constructing the position and volume of the sliding body according to the volume and center of gravity of the sliding body; according to the first result of the landslide surge physical model test, analyze the landslide instability evolution movement process, the height of the first wave entering the landslide reservoir and the surge propagation process of the landslide surge;

[0074] The second result acquisition module is used to construct a landslide surge numerical model test according to the information, wherein the landslide surge numerical model test includes constructing the terrain according to the contour information and constructing the position and volume of the sliding body according to the volume and center of gravity of the sliding body; and performing a numerical simulation test according to the second result of the landslide surge numerical model test;

[0075] The test parameter adjustment module is used to use the first wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, and adjust the particle size and kernel function in the landslide surge numerical simulation test to make the first wave height, the height reaching the opposite bank and the surge pressure distribution in the landslide surge numerical simulation test and the landslide surge physical model test consistent;

[0076] The hybrid model acquisition module is used to use the adjusted landslide surge numerical model as the obtained landslide surge hybrid model to analyze the working conditions of different landslide velocities, landslide volumes and terrain conditions.

[0077] To achieve the above object, the present invention also provides the following technical solutions:

[0078] The invention discloses an application of a landslide-surge hybrid model. The landslide-surge hybrid model is obtained by the landslide-surge physical and numerical hybrid model construction method. The landslide-surge hybrid model analyzes the working conditions of different landslide velocities, landslide volumes and terrain conditions.

[0079] The construction and preliminary analysis of the physical model test of the present invention provide basic data for the test and model construction by collecting contour geological information, sliding body volume center of gravity height and water level information; based on the collected information, a physical model reflecting the actual terrain and sliding body characteristics is constructed, so that the test results have higher practical significance; physical model tests are carried out to analyze the evolution process of landslide instability, the height of the first wave and the wave propagation process, and obtain a preliminary intuitive understanding of the movement characteristics and mechanical behavior. Significance: Through the physical model, the movement laws and influencing factors of landslide surges can be more intuitively understood, providing a reference for numerical simulation; providing a physical experimental basis for the establishment of numerical models, helping to predict and optimize the parameters and complexity of numerical simulations. The construction and simulation of numerical model tests, constructing numerical models based on contour lines and sliding body characteristics, provides an efficient computing tool for simulating complex landslide surge processes; conducting experiments through numerical models to simulate the dynamic process of landslide surges and provide quantitative data support. Significance: Numerical simulation technology can be applied in a wider range of situations (such as extreme weather, different terrains, etc.) to improve the prediction accuracy of landslide surges; compared with physical experiments, numerical simulation has more obvious advantages in time and cost, and can quickly simulate and analyze a variety of situations. Correction and optimization of model results: By adjusting the particle size and kernel function, the numerical model is optimized to ensure the consistency between the numerical simulation results and the physical experimental results; by setting the evaluation criteria for the first wave height, the height reaching the opposite bank and the surge pressure distribution, the applicability of the model under various conditions is guaranteed. Significance: The calibrated landslide surge numerical model has improved the consistency with the physical experiment, which significantly enhances the reliability and accuracy of the model in practical applications; the calibration and optimization process not only improves the accuracy of the existing model, but also provides a methodological basis for similar studies in the future. Application and analysis of the hybrid model: Using the adjusted hybrid model, the landslide surge under different working conditions is analyzed to evaluate its response to the landslide velocity, volume and terrain conditions; it can simulate the surge behavior under different working conditions and form a comprehensive study of the landslide surge phenomenon. Significance: The analysis results of different working conditions can provide a scientific basis and reference for disaster prevention and mitigation, guide the design of infrastructure and landslide monitoring and early warning; promote the development of research technology on the dynamic behavior of landslide surges, and lay the foundation for the proposal and application of new methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A schematic diagram of the steps of an embodiment of a method for constructing a landslide surge physical and numerical hybrid model of the present invention;

[0081] Figure 2 A schematic diagram of an embodiment of a method for constructing a landslide surge physical and numerical hybrid model of the present invention;

[0082] Figure 3This is a schematic flow chart of the steps of one embodiment of a method for constructing a landslide surge physical and numerical hybrid model according to a first result of a landslide surge physical model test;

[0083] Figure 4 It is a schematic flow chart of the steps of one embodiment of the method for constructing a landslide surge physical and numerical hybrid model according to the second result of the landslide surge numerical model test;

[0084] Figure 5 A schematic flow chart of the steps of adjusting the particle size and kernel function in a landslide surge numerical simulation test in one embodiment of the method for constructing a landslide surge physical and numerical hybrid model of the present invention;

[0085] Figure 6 A schematic flow chart of the steps of obtaining a landslide surge hybrid model according to an embodiment of a method for constructing a landslide surge physical and numerical hybrid model of the present invention;

[0086] Figure 7 A schematic diagram of functional modules of an embodiment of a landslide surge physical and numerical hybrid model construction system of the present invention;

[0087] Figure 8 It is a structural schematic diagram of an embodiment of an electronic device of the present invention;

[0088] Fig. 9 It is a schematic structural diagram of an embodiment of the storage medium of the present invention. DETAILED DESCRIPTION

[0089] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0090] The terms "first", "second" and "third" in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" can explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices.

[0091] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0092] like Figure 1 As shown, this embodiment provides an embodiment of a method for constructing a landslide surge physical and numerical hybrid model. In this embodiment, the method for constructing a landslide surge physical and numerical hybrid model specifically includes the following steps:

[0093] Step S1: obtaining contour geological information of the target area, information on the height of the center of gravity of the sliding body and the water level; constructing a landslide surge physical model test based on the information, the landslide surge physical model test includes constructing the terrain based on the contour information and constructing the position and volume of the sliding body based on the volume and center of gravity of the sliding body; analyzing the landslide instability evolution movement process of the landslide surge, the height of the first wave entering the landslide reservoir and the surge propagation process according to the first result of the landslide surge physical model test;

[0094] Step S2: constructing a landslide surge numerical model test according to the information, the landslide surge numerical model test comprising constructing a terrain according to the contour information and constructing a sliding body position and volume according to the sliding body volume and gravity center; performing a numerical simulation test according to the second result of the landslide surge numerical model test;

[0095] Step S3: taking the first wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, adjusting the particle size and the kernel function in the numerical simulation test of landslide surge, so that the first wave height, the height reaching the opposite bank and the surge pressure distribution in the numerical simulation test of landslide surge and the physical model test of landslide surge are consistent;

[0096] Step S4: using the adjusted landslide surge numerical model as the obtained landslide surge hybrid model to analyze the working conditions of different landslide velocities, landslide volumes and terrain conditions.

[0097] Preferably, in step S1 of this embodiment, the construction and preliminary analysis of the physical model test provide basic data for the test and model construction by collecting contour geological information, the height of the center of gravity of the sliding body and the water level information; based on the collected information, a physical model reflecting the actual terrain and sliding body characteristics is constructed, so that the test results have higher practical significance; a physical model test is carried out to analyze the evolution process of landslide instability, the height of the first wave and the propagation process of the surge wave, and obtain a preliminary intuitive understanding of the motion characteristics and mechanical behavior. Significance: Through the physical model, the motion laws and influencing factors of the landslide surge wave can be more intuitively understood, providing a reference for numerical simulation; providing a physical experimental basis for the establishment of the numerical model, helping to predict and optimize the parameters and complexity of the numerical simulation. Step S2 Construction and simulation of the numerical model test, constructing a numerical model based on contour lines and sliding body characteristics, providing an efficient computing tool for the simulation of complex landslide surge wave processes; conducting experiments through numerical models to simulate the dynamic process of landslide surge waves and provide quantitative data support. Significance: Numerical simulation technology can be applied in a wider range of situations (such as extreme weather, different terrains, etc.) to improve the prediction accuracy of landslide surges; compared with physical experiments, numerical simulation has more obvious advantages in time and cost, and can quickly simulate and analyze a variety of situations. Step S3: Correction and optimization of model results. By adjusting the particle size and kernel function, the numerical model is optimized to ensure the consistency between the numerical simulation results and the physical experimental results; by setting the evaluation criteria for the first wave height, the height reaching the opposite bank and the surge pressure distribution, the applicability of the model under various conditions is guaranteed. Significance: The calibrated landslide surge numerical model has improved the consistency with the physical experiment, which significantly enhances the reliability and accuracy of the model in practical applications; the calibration and optimization process not only improves the accuracy of the existing model, but also provides a methodological basis for similar studies in the future. Step S4: Application and analysis of the hybrid model. Using the adjusted hybrid model, the landslide surge under different working conditions is analyzed to evaluate its response to the landslide velocity, volume and terrain conditions; it can simulate the surge behavior under different working conditions and form a comprehensive study of the landslide surge phenomenon. Significance: The analysis results of different working conditions can provide scientific basis and reference for disaster prevention and mitigation, guide the design of infrastructure and landslide monitoring and early warning; promote the development of landslide surge dynamic behavior research technology, and lay the foundation for the proposal and application of new methods (for specific principles, please refer to the attached Figure 2 ).

[0098] In summary, this embodiment optimizes the accuracy, efficiency and application scope of landslide surge related research through the combination of physical model and numerical model, which is of great value to both theoretical research and practical engineering application. At the same time, this embodiment constructs a landslide surge physical model based on the landslide topographic and geomorphic characteristics, and conducts a landslide surge physical model test according to the landslide characteristics such as the volume of the landslide body, the center of gravity height and other factors; analyzes the landslide instability evolution movement process, the height of the first wave of the landslide entering the reservoir and the surge propagation process of the landslide surge. On the basis of the landslide surge physical model test, a numerical simulation test is carried out to simulate the landslide surge situation under the same geological conditions, and the landslide instability evolution movement process, the height of the first wave of the landslide entering the reservoir, the surge propagation process and the impact on important upstream and downstream residential areas. According to the test results of the landslide surge physical model, the relevant parameters of the landslide surge numerical simulation test are optimized, and multi-angle and multi-variable test results are obtained from the optimized landslide surge numerical simulation test to obtain a landslide surge hybrid model. This embodiment can improve simulation accuracy, enhance calculation efficiency, and expand calculation range; the test cost is low, and different calculation conditions can be easily analyzed, such as different landslide speeds, landslide volumes, and terrain conditions, etc., to provide a scientific basis for disaster prevention and mitigation.

[0099] Furthermore, if Figure 3 As shown, the process in step S1 according to the first result of the landslide surge physical model test specifically includes the following steps:

[0100] Step S11: Acquire contour data of the target area, including information such as altitude, slope, soil type and rock properties, and evaluate the volume and center of gravity of the potential sliding body of the landslide. The volume of the sliding body includes the cross-sectional area, volume and shape, etc. The determination of the center of gravity height includes the geometric center of the sliding body and its position on the slope;

[0101] Step S12: constructing a landform model based on the collected contour data; placing the position of the simulated sliding body on the landform model; selecting a fluid medium (such as clean water, salt water, etc.) to simulate the hydrodynamic effect of landslide surges; configuring experimental equipment;

[0102] The experimental equipment includes a water tank, a wave generating device, sensors and a data acquisition system. The sensors can be used to monitor parameters such as wave height, pressure changes and flow velocity.

[0103] Step S13: Set the initial water level and the initial state of the sliding body, use high-speed cameras, pressure sensors, wave height meters and other equipment to record in real time the first result of the landslide entering the water, including the evolution of landslide instability, the height of the first wave, the propagation speed of the surge wave and the wave pressure distribution at each point, process the wave height change diagram and pressure distribution diagram of the first result, and display the propagation process of the landslide surge wave.

[0104] Preferably, in step S11 of this embodiment, data acquisition and landslide assessment can fully understand the characteristics of the terrain, including altitude, slope, etc., by acquiring the contour data of the target area, which is crucial for determining the shape and movement trajectory of the landslide body; assessing the potential volume of the landslide body and its center of gravity position is helpful to establish a more accurate physical model, ensuring that the landslide body reflects the real landslide phenomenon in the model. Significance: Ensure that the constructed physical model is scientific and accurate; predict the stability and potential risks of the landslide by analyzing the volume and center of gravity of the landslide body, and provide a basis for preventing landslide accidents; at the same time, ensure that the model results can be effectively transformed in practical applications (such as disaster assessment, engineering design, etc.). Step S12: Geomorphic model construction and experimental equipment configuration, construct a geomorphic model based on the contour data, ensure that the geometric characteristics of the model are consistent with the actual terrain, and enhance the reality and accuracy of the model; configure a water tank, a wave generating device, a sensor and a data acquisition system to form an integrated experimental platform, which can comprehensively analyze landslide surge events. Significance: Through the actual experimental setting, the hydrodynamic effect caused by the landslide can be simulated in a controllable environment, so as to study the generation and propagation of the surge wave more intuitively; the configured experimental equipment allows real-time monitoring and recording of multiple parameters (wave height, pressure, flow velocity, etc.), providing a rich source of information for subsequent data analysis, which helps to improve the accuracy and depth of the research. Step S13 Initial condition setting and real-time recording, setting the initial water level and the initial state of the sliding body, ensuring that the experiment is carried out in a controllable environment, making the simulation results more valuable for reference; using special equipment such as high-speed cameras, pressure sensors, wave height meters, etc., it is possible to capture the dynamic changes of the landslide entering the water in real time, including the evolution of landslide instability, the propagation speed of the surge wave, etc., and obtain the corresponding wave height change map and pressure distribution map. Significance: Through real-time recording and data processing, the generation mechanism, propagation characteristics and impact of landslide surges on the surrounding environment are observed and analyzed, thus providing empirical support for understanding complex hydrodynamic phenomena; the visualization of results (such as wave height change diagrams and pressure distribution diagrams) provides an important basis for theoretical analysis and model verification, helps to identify the laws of landslide and surge occurrence, and provides scientific basis and suggestions for future disaster prevention and mitigation projects.

[0105] In summary, this embodiment constitutes the core process of the landslide surge physical model test, laying a solid foundation for subsequent numerical simulation, verification and analysis. Through data acquisition, accurate model construction and meticulous experimental observation, we can not only have a deeper understanding of landslides and the surge phenomena caused by them, but also provide reliable theoretical support and practical guidance for related fields (such as water conservancy projects, environmental protection, disaster management, etc.).

[0106] Furthermore, the process of evaluating the volume and center of gravity of the potential sliding body of the landslide in step S11 specifically includes the following steps:

[0107] Step S111: Calculate the slope by the elevation difference and horizontal distance between adjacent contour lines, perform section analysis on the contour lines, calculate the cross-sectional area of ​​the landslide body according to the contours of the landslide body at different heights; calculate the volume of the entire landslide body according to the area of ​​each cross section;

[0108] The calculation formula of slope is:

[0109]

[0110] Where S represents the slope, h i and h j is the altitude of two adjacent contour lines, d ij is the horizontal distance between two adjacent contour lines;

[0111] For each level of cross-section, define a cross-sectional area:

[0112] A(z)=∫ x (y upper -y lower ),dx

[0113] Where A is the cross-sectional area, z is the height, and y is the upper and lower are functions of the upper and lower boundaries respectively, and x is the horizontal coordinate;

[0114] The total volume of the sliding body is obtained by adding up the cross-sectional areas along the height direction:

[0115]

[0116] Where V represents the total volume of the sliding body, z1 and z2 are the heights of the lowest and highest points of the sliding body, respectively;

[0117] Step S112: according to the geometric shape of the sliding body, the sliding body is divided into a plurality of basic geometric bodies such as a triangular prism, a cuboid and a cone, and the center of gravity of each part is calculated respectively; for each divided small volume, the weighted average of the center of gravity of all parts is calculated to comprehensively obtain the center of gravity position of the entire sliding body;

[0118]

[0119] Among them, x c ,y c , z c They represent the coordinates of the center of gravity of the sliding body, V is the overall volume, and x, y, and z are the coordinates of each point in the body;

[0120]

[0121] Among them, the center of gravity position of the sliding body is x g ,y g,z g , V i is the volume of the ith body, is the coordinate of the center of gravity of the ith body;

[0122] Step S113: by converting the height of the center of gravity of the slider into the slope coordinate system;

[0123] The conversion formula is: r =z g -(S·D)

[0124] Among them, z r is the height of the center of gravity on the slope, S is the slope, and D is the horizontal distance from the slope baseline to the center of gravity.

[0125] Preferably, step S111 of this embodiment calculates the slope, cross-sectional area and total volume, determines the slope characteristics of the terrain in the landslide area, and calculates the slope using the elevation difference and horizontal distance between adjacent contour lines, providing basic data for subsequent analysis; by sectioning the contour lines, the outline of the landslide body at different heights is obtained, thereby calculating the cross-sectional area at each height level, ensuring an accurate description of the geometric shape of the sliding body; the total volume of the sliding body is obtained according to the accumulation of each cross-sectional area along the height, thus forming a basis for quantitative analysis of the hydrodynamic impact that the landslide body may cause. Significance: Accurate slope and cross-sectional data provide key parameters for evaluating the stability and sliding risk of the landslide body; the estimation of the total volume can provide an important reference for subsequent hydrodynamic surge calculations. After a landslide occurs, the degree of impact that the water flow may have is closely related to the volume of the sliding body, helping to assess potential disastrous consequences. Step S112: Geometric segmentation and center of gravity calculation. The sliding body is divided into multiple basic geometric bodies (such as triangular prisms, cuboids, cones, etc.) to facilitate the calculation of the center of gravity of each part one by one, which helps to simplify the calculation of the center of gravity of objects with complex shapes; the center of gravity of each small body is calculated, and the weighted average of the centers of gravity of all parts is performed to obtain the overall center of gravity position of the sliding body, which can more accurately reflect the mass distribution characteristics of the sliding body. Significance: The determination of the center of gravity position provides an important basis for evaluating the stability of the sliding body on the slope. A high center of gravity or deviation from the support surface may cause the sliding body to lose its center of gravity, thereby increasing the risk of landslides; through the precise calculation of the sliding body's geometric shape and its center of gravity, a more accurate data basis is provided for the subsequent hydrodynamic model, which improves the authenticity and reliability of the landslide surge simulation. Step S113: The center of gravity height is converted to the slope coordinate system. The center of gravity of the sliding body is transformed from the standard coordinate system to the slope coordinate system through the center of gravity height conversion formula, which ensures the evaluation of the center of gravity position after considering the influence of the slope. Significance: Converting the center of gravity position to the slope coordinate system makes the evaluation result more realistic and can be directly used to judge the stability of the landslide body under specific terrain conditions; it helps to accurately evaluate the potential risk of landslides based on the existing slope and height conditions, guide the monitoring and management of potential landslide areas, and reduce the probability of landslide disasters.

[0126] In summary, this embodiment forms a systematic and scientific landslide volume and center of gravity assessment process, which not only provides solid data support for the physical model of landslide surge, but also provides a scientific basis for the actual landslide risk assessment and formulation of prevention and control measures, which helps to improve safety and reduce disaster risks in engineering practice.

[0127] Furthermore, the construction process of the landform model in step S12 specifically includes the following steps:

[0128] Step S121: arranging grid lines in the target area to form a grid structure, adjusting the vertex coordinates of each grid unit according to the contour line data, calculating and obtaining the heights of all vertices in the grid corresponding to the contour lines; storing the coordinates and height data of each grid in the database;

[0129] Step S122: input the elevation and its corresponding position coordinates into the interpolation model, determine the four corner points of a given grid as input, and perform interpolation calculation on any point in the given grid according to its height value; generate a set of three-dimensional point cloud data through interpolation calculation;

[0130] in:

[0131]

[0132] In the formula, the four corner points of the grid (P1(x1, y1), P2(x2, y3), P3(x3, y3), P4(x4, y4)) are used as input, and the interpolation calculation is performed on any point (x, y) within the given grid according to their height values ​​(Z1, Z2, Z3, Z4);

[0133] Step S123: Use the generated point cloud data to perform triangulation, connect the point cloud data into a triangular mesh, form non-overlapping triangles, and form an irregular mesh; select a point from the point cloud data as the starting point, select the two nearest adjacent points to form the first triangle; among the unconnected points, continue to add adjacent points, and determine whether the newly added points form a triangle, until all points are connected, and display the triangular mesh as a continuous three-dimensional surface.

[0134] Preferably, in step S121 of this embodiment, the grid lines arranged in the target area form a regular grid structure, which can subdivide the area so that each grid unit can reflect the terrain characteristics of the area; the coordinates of the vertices of each grid unit and its corresponding elevation value are calculated and obtained, and these data are stored in the database, providing the necessary data basis for subsequent processing. The significance achieved: by storing the data in the database, the consistency and integrity of the data are guaranteed; the subdivided area enables each grid unit to reflect the subtle changes in the terrain, which is helpful for the subsequent interpolation and modeling process and improves the accuracy of the model. Step S122 interpolation calculation, input the elevation data into the interpolation model, and calculate the elevation value of any point in each grid unit through the interpolation formula (such as bilinear interpolation), thereby generating functional three-dimensional point cloud data. The significance achieved: through interpolation technology, smoother values ​​can be created between known data points, which not only fills the gap in elevation data, but also more accurately reflects the continuity and change characteristics of the terrain; the generated three-dimensional point cloud data lays a solid foundation for subsequent three-dimensional reconstruction and visualization, ensuring that the model can truly present the three-dimensional structure of the terrain. Step S123 triangulation and model reconstruction, using the generated point cloud data for triangulation processing, the non-overlapping triangular mesh formed by connecting the points ensures the connectivity and seamlessness of the model. Significance achieved: The irregular mesh created by the triangulation method can better adapt to complex terrain and express the details of the terrain; the triangular mesh formed by the processing can be represented as a continuous three-dimensional surface, making the landform model more realistic visually and convenient for subsequent analysis and research; the triangular mesh not only provides a basis for subsequent visual display, but also can effectively display the ups and downs, steep slopes, and other important landform features in different areas.

[0135] In summary, each step of this embodiment can eventually construct an accurate, continuous and realistic three-dimensional landform model. These models not only help to understand and analyze actual geographical phenomena, but also provide important basic data and support for related research and engineering applications.

[0136] Furthermore, the propagation process of the landslide surge in step S13 specifically includes the following steps:

[0137] Step S131: Arrange the wave height data obtained from the wave height meter into a time series, draw a wave height change curve, and analyze the wave height change before and after the landslide enters the water; Arrange the data of the pressure sensor to create a time series record of the pressure change of each sensor during the experiment; Extract the motion trajectory of the sliding body and the relevant data of the moment of entering the water from the video recorded by the high-speed camera, including the contact between the sliding body position and the water surface;

[0138] Step S132: import the recorded wave height data, and generate a wave height change graph using a graphical tool; the X-axis represents time, the Y-axis represents wave height, and a wave height change curve is generated to show how the wave height jumps at the moment the landslide enters the water, and the change trend over time; obtain the reading of each pressure sensor, construct a data set of pressure at each point on the water surface that changes with time, map the pressure data to a fictitious two-dimensional plane, generate a pressure distribution graph, and show the change of wave impact force at different positions;

[0139] Step S133: Import the original video obtained by the high-speed camera into the video processing software, extract the key frames related to the natural landslide, water entry and wave formation from the video, and annotate them in combination with the wave height and pressure data; combine the wave height change graph and the pressure distribution graph, superimpose them on the video, generate a visual animation of the surge propagation process, and display the dynamic characteristics of the landslide surge.

[0140] Preferably, in step S131 of this embodiment, data collation and trajectory extraction are used to collate the continuous wave height data from the wave height meter into a time series, so that it is comparative and analytical, and is convenient for observing the change of wave height over time; the reading of the pressure sensor is converted into a pressure data set that changes over time, and the pressure fluctuation of the waves recorded by each sensor during the experiment is identified; the image provided by the high-speed camera is used to extract the motion trajectory of the sliding body entering the water, such as the position change of the sliding body and the moment of contact with the water surface, which provides the necessary dynamic information for subsequent analysis. Significance: By comprehensively collating and reviewing the data, a preliminary understanding of the landslide and its interaction with the water medium can be obtained, which provides the necessary foundation for subsequent analysis and can lay a good data foundation for subsequent dynamic analysis and visualization. Step S132 is graphical and data mapping, which uses graphical tools to present the time series wave height data in graphical form, so that the observer can intuitively see the change of wave height over time. This graph can quickly convey the change trend of wave height and the impact caused by the landslide event; the data recorded by the pressure sensor is mapped to a two-dimensional plane to form a pressure distribution map, which allows for spatial understanding of the wave impact force at different locations. For example, the intensity and impact range of waves at different locations can be observed. Significance: Through the visualization of wave height change graphs and pressure distribution graphs, the key dynamic changes of landslide surges can be quickly identified and explained. This not only improves the comprehensibility of the data, but also provides obvious and easy-to-understand visual data support for further demonstration, report writing or scientific communication. Step S133 video processing and animation generation, extracting key frames related to landslide, water entry and wave formation from the original video to provide visual presentation for making dynamic demonstrations; superimposing the wave height change graph and pressure distribution graph with the video to form a dynamic visualization including data and real scenes, so that the audience can see the real dynamics of the landslide event and related data changes at the same time. Significance: The generated visualization animation not only vividly shows the propagation process of landslide surges, making the dynamic characteristics clearly visible, but also provides a powerful means of display for teaching, publicity, and scientific research; while intuitively understanding, it will also have a deeper understanding of the complex engineering background and hydrodynamic behavior of landslide surges, which is of great significance to promoting the formulation and implementation of corresponding disaster prevention and mitigation measures.

[0141] In summary, each step of this embodiment not only improves the analysis efficiency and comprehensibility of data through the integration, visualization and animation of landslide surge data, but also provides key support for understanding the dynamic characteristics of landslide surge. It not only assists the development of scientific research, but also provides important foundation and materials for academic exchanges and practical applications in related fields.

[0142] Furthermore, if Figure 4 As shown, the process in step S2 according to the second result of the landslide surge numerical model test specifically includes the following steps:

[0143] Step S21: obtaining contour geological information of the target area, analyzing and determining the volume, center of gravity height and affected water level of the sliding body according to the geological survey data of the source of the landslide; constructing a corresponding contour terrain model based on the obtained contour information, constructing and placing the sliding body in the contour model according to the calculated center of gravity position and volume;

[0144] Step S22: Based on the same geological data and gravity center characteristics, the sliding body is simulated in the landslide surge numerical model to present the sliding body movement process;

[0145] Step S23: In the numerical model test of the sliding body on the landslide surge, the height of the wave formed for the first time after the landslide enters the reservoir is accurately measured using an altimeter and recorded as a reference value; a numerical simulation is run to automatically extract the height of the first wave, the height reaching the opposite bank, and the distribution value of the surge pressure.

[0146] Preferably, in step S21 of this embodiment, the contour geological information acquisition and model construction are carried out by high-precision geological survey means to obtain the contour lines and related geological information of the target area, which provides a terrain basis for landslide simulation and ensures the authenticity of the model; based on the survey data, the volume, center of gravity height and affected water level of the sliding body are determined to provide necessary parameter support for the performance of the sliding body in the model; the acquired contour information is used to construct an accurate contour terrain model, and the simulation scene is ensured to be consistent with the actual environment by accurately drawing the terrain and landforms. Significance: It is ensured that the constructed numerical model has real geological characteristics; through the combination of accurate geological information and sliding body characteristics, the formation and evolution process of landslide surge can be simulated and analyzed more scientifically. Step S22 realizes the motion simulation of the sliding body in the numerical model according to the previously obtained volume, center of gravity position, etc. This process may use a fluid dynamics model to calculate the motion trajectory and water entry time of the sliding body under the action of gravity and water flow; during the simulation process, the system will track the motion state of the sliding body in real time, and dynamically update the interaction between the sliding body and the water surface through numerical algorithms, providing basic data for subsequent wave height and pressure changes. Significance: It can deeply understand the complex process of the sliding body entering the water, so as to predict the initial formation of the landslide surge and its impact on the surrounding environment; the accurate simulation of the sliding body movement is the premise of the subsequent surge characteristic analysis, which improves the effectiveness and reliability of the model. Step S23: Measurement and extraction of wave height and pressure distribution. Using precision measuring tools such as altimeters, the height of the first wave after the landslide enters the reservoir is accurately recorded in the physical model. This data has a benchmark role in subsequent analysis; through numerical simulation, the height of the first wave after the model runs, the height to the opposite bank and the surge pressure distribution value are automatically extracted, and the specific numerical model analysis software is used for rapid processing to ensure the efficiency and accuracy of the results. Significance: It not only provides data support for quantifying the initial formation process of landslide surge, but also establishes a direct comparison basis for the relationship between numerical models and physical models; by recording and extracting various parameters in detail, it can effectively verify the accuracy of the numerical model, and make necessary corrections and optimizations to the model; it helps to improve the credibility of the results and provide a theoretical basis for actual disaster prevention and early warning and engineering design.

[0147] In summary, this embodiment is an indispensable part of the landslide surge numerical model experiment process. By gradually constructing a terrain model, simulating the movement of the sliding body, and accurately measuring the surge parameters, it provides a scientific basis and reliable data support for the study of the dynamic characteristics of landslide surges. It not only enhances the applicability of the model, but also provides rich experience and data references for future related research and practice.

[0148] Furthermore, the process of automatically extracting the head wave height, the height reaching the opposite shore and the surge pressure distribution value in step S23 specifically includes the following steps:

[0149] Step S231: meshing is performed, and the mesh size is matched to key areas such as the interface between the sliding body and the water surface; the sliding body is injected into the water under the action of gravity to form an initial wave, simulating the surge generated by the interaction between the sliding body and the water; the movement of the sliding body is simulated by considering the fluid dynamics equation;

[0150] Among them, the grid generation standard is set, and the gradient change of the flow field is greater than a certain threshold Δx adaptive When the mesh is refined, the following standard is expressed: is the adaptation function according to the change of velocity field, Δx min Indicates the minimum grid size;

[0151] The form of the equation in the initial wave simulation is:

[0152]

[0153] Where h0 is the still water depth and η is the wave height, which is used for high-order correction;

[0154] Step S232: The numerical model continuously calculates the motion state of the sliding body in the water and optimizes the flow field in each time step; monitors and records the change of the water surface height in real time, and the monitoring points are set in the downstream waters and the opposite bank area where the sliding body enters the water to capture the propagation of the first wave; extracts the height of the first wave of the ripple and records it as the reference value;

[0155] Among them, the complex interaction between the water-entering sliding body and the water surface is tracked, and the relevant calculation format is as follows: Where φ is the interface function, identifying the interface between water and air;

[0156] Step S233: After the sliding body enters the water, the landslide surge numerical simulation model will continue to track the propagation process of the surge, and at the end of the operation, extract the height of the first wave, the height reaching the opposite bank, and the surge pressure distribution;

[0157] Among them, the first wave height records the water surface height when it reaches the highest point of the wave; the arrival height at the opposite bank measures the water level change when the first wave reaches the opposite bank and is recorded at a specific point on the opposite bank; the surge pressure distribution automatically calculates the pressure distribution caused by the surge by monitoring the water depth change, fluid velocity and water surface height, combined with the fluid dynamics model; the expression of the surge propagation process is:

[0158]

[0159] f i represents the distribution function of the particle, τ is the relaxation time e i is the direction of particle flow.

[0160] Preferably, in step S231 of this embodiment, mesh division and initial wave simulation use adaptive mesh technology to finely divide key areas in the flow field (such as the interface between the sliding body and the water surface), ensuring calculation accuracy and numerical stability; by dynamically adjusting the mesh size, the calculation efficiency is improved, especially using smaller meshes in places where the flow field changes drastically; the introduction of high-order corrections to the equation can more realistically reflect the wave characteristics and their propagation laws, providing a more scientific basis for the setting of initial conditions. Significance: Through precise mesh division and the application of high-order equations, the reliability of surge simulation is guaranteed, and the waves caused by the sliding body entering the water can be effectively captured, which is helpful for a deeper understanding of the surge characteristics. Step S232 motion state calculation and real-time monitoring, the model continuously calculates the motion state of the sliding body in the water, and updates the flow field state in real time, so that the dynamic evolution of the surge can be quickly reflected in the simulation, improving the response capability; the interface function is used to track the interface between the water body and the air, ensuring that the complex interaction between the sliding body and the water surface can be accurately captured, providing necessary information support for subsequent wave height extraction. Significance: Real-time monitoring of changes in water surface height enables the propagation process of the head wave to be captured, improving the extraction accuracy of the head wave height, which is of great significance for understanding the propagation mechanism of surge waves; it provides forward-looking warning information for actual disaster prevention and mitigation work, and can predict possible impacts in advance. Step S233 Surge propagation tracking and parameter extraction, the model can continuously monitor the dynamic evolution of surge waves, and the use of relevant expressions makes the calculation of surge behavior more accurate. At the same time, through the particle distribution function, it can better simulate complex fluid behavior; by combining a variety of fluid dynamics models and monitoring results, it can automatically calculate the pressure distribution caused by surge waves, providing a basis for quantifying the potential threat of surge waves. Significance: In the final parameter extraction, the head wave height, the height reaching the opposite bank, and the surge pressure distribution can be clearly recorded, so that a comprehensive assessment of the surge effect can be achieved; it has important application value in the fields of engineering design, facility safety assessment, etc., and can provide a scientific basis for actual flood control, facility design and emergency response, and reduce potential risks.

[0161] In summary, through the complete execution of the above steps in this embodiment, the model can not only efficiently extract key parameters related to surge waves, but also plays an important positive role in optimizing hydrodynamic simulation models, improving the accuracy of parameter extraction, and supporting practical applications and decision-making processes; the improvement of the degree of automation not only makes the prediction of surge behavior more accurate, but also provides strong support for predicting the impact of surge waves on structures in actual engineering, which helps to enhance the society's ability to respond to natural disasters and the safety of engineering facilities.

[0162] Furthermore, if Figure 5 As shown, the process of adjusting the particle size and kernel function in the landslide surge numerical simulation test in step S3 specifically includes the following steps:

[0163] Step S31: Analyze the experimental data obtained, including extracting the specific values ​​of the head wave height, the height reaching the opposite shore and the surge pressure distribution, which will be used as the evaluation criteria; establish a relationship diagram between the target value and the current numerical simulation result, and identify the gap between the numerical simulation and the physical experimental results;

[0164] Step S32: Analyze the particle distribution characteristics in the landslide surge physical model and compare the influence of different particle sizes on the numerical results; perform numerical simulation by changing the particle diameter and continuously refine the particle size until it matches the results of the landslide surge physical model; based on the analysis of surge behavior in the landslide surge physical model experiment, confirm the kernel function type and adjust the support domain and attenuation rate of the kernel function;

[0165] Step S33: After the particle size and kernel function are adjusted, a new round of numerical simulation is performed: the adjusted numerical model is comprehensively compared with the physical experimental results, including the verification of the first wave height, the detection of the height reaching the opposite shore, and the fitting of the surge pressure distribution;

[0166] Among them, the first wave height verification: observe whether the height of the first wave reaches the peak value of the physical experiment, record and evaluate the difference between the two; the arrival height detection on the other side: check the time and height of the surge reaching the other side in the new simulation results to ensure the consistency between the two in time and space; surge pressure distribution fitting: analyze the pressure changes at different positions to ensure that the pressure distribution of the surge is consistent with the physical experiment, especially the pressure characteristics in key areas (such as the junction of the water body and the shore).

[0167] Preferably, the experimental data analysis and initial evaluation criteria of step S31 of this embodiment are established, and by extracting the specific values ​​of the head wave height, the height to the opposite shore and the surge pressure distribution, a clear quantitative basis is provided for the subsequent numerical model adjustment; by establishing a relationship diagram between the target value and the current numerical simulation result, the gap between the numerical analysis and the physical experimental results can be intuitively identified, helping engineers and researchers understand the current performance of the model to narrow the gap between the model and the actual situation. The significance achieved: clear and quantifiable goals are set for subsequent model adjustment and verification to ensure the repeatability and reliability of numerical simulation; it is helpful to deeply understand the specific error sources in the model, and provide a practical basis for adjusting parameters through quantitative analysis, so as to form an iterative improvement idea in model development. Step S32 particle size and kernel function adjustment, experiments have shown that different particle sizes have a significant impact on fluid flow and wave characteristics; by comparing the numerical simulation results of different particle diameters, the particle settings can be optimized to ensure the accuracy of fluid momentum and energy transfer; confirm and adjust the applicable kernel function type, support domain and attenuation rate to ensure that the model better simulates surge behavior and fluid interaction. Significance achieved: Reasonable adjustment of particle size and kernel function characteristics can help enhance the performance of numerical models under complex flow conditions and improve the model's ability to predict actual phenomena; by adjusting specific model parameters (such as kernel function configuration), higher flexibility and personalization can be achieved, so that the model can adapt to different landslide surge scenarios and improve comprehensive analysis capabilities. Step S33 A new round of numerical simulation and result verification. After comparing the adjusted model with the physical experimental results, the improvement effect of the numerical model can be evaluated, and key projects such as first wave height verification, arrival height detection on the opposite bank, and surge pressure distribution fitting can be completed; through consistency analysis of key parameters (such as wave height, time, pressure), comprehensive feedback on the accuracy and reliability of the model is provided. Significance achieved: Through systematic comparative verification, it can not only support the effective adjustment of the current numerical model, but also help to confirm which parameters are crucial for future surge prediction; ensure that the iteratively adjusted numerical model has good applicability, which can provide a scientific basis for safety assessment and engineering design under similar landslide or water wave conditions, and provide support for subsequent practical applications.

[0168] In summary, through the above steps, the entire process of this embodiment not only provides a systematic methodology for the optimization and calibration of the landslide surge numerical model, but also ensures the effectiveness and usability of the model through quantitative experimental data and cross-validation. It has important practical significance for improving the safety assessment of water conservancy projects and reducing the potential losses caused by landslides.

[0169] Furthermore, the process of establishing a relationship diagram between the target value and the current numerical simulation result in step S31 specifically includes the following steps:

[0170] Step S311: obtaining multiple key parameters such as the head wave height, the height reaching the opposite shore and the surge pressure distribution, for setting the target value; at the same time, obtaining corresponding results from the numerical simulation, and recording the head wave height, the height reaching the opposite shore and the corresponding surge pressure distribution data obtained by simulation under the same conditions;

[0171] Step S312: The collected landslide surge physical model experimental data and the second result of the numerical simulation are matched with the time point and spatial position of collection, and smoothed; the height of the first wave and the height reaching the opposite bank are normalized according to the maximum value respectively, and a standardized value between 0 and 1 is generated;

[0172] Step S313: using a two-dimensional coordinate system, with the X-axis representing the simulation result and the Y-axis representing the experimental target value; making a scatter plot to display the relationship between the simulation result and the target value; and obtaining the deviation between the numerical simulation result and the target value by calculating the distance quantification gap of each point;

[0173] Among them, there are n pairs of simulation results and target values ​​of key parameters, which are expressed as follows: Numerical simulation results S i , where i = 1, 2, ..., n, target value T i ;

[0174] Absolute error refers to the direct difference between the numerical simulation result and the target value, and the calculation formula is:

[0175] E abs,i =|S i -T i |

[0176] Among them, E abs,i represents the absolute error of the i-th key parameter;

[0177] The relative error takes into account the impact of the target value and uses the ratio of the target value to standardize the error. The calculation formula is:

[0178]

[0179] Among them, E rel,i represents the relative error of the i-th key parameter (expressed as a percentage);

[0180] The root mean square error RMSE is used to comprehensively evaluate the deviations of multiple parameters. The calculation formula is:

[0181]

[0182] The average of the squared errors for all key parameters was calculated and then the square root was taken, providing a more comprehensive error measure;

[0183] The comprehensive deviation evaluation evaluates each key parameter together with the absolute error, relative error and root mean square error to form a new indicator:

[0184]

[0185] Among them, α, β, and γ are weight coefficients, reflecting the importance of different deviation measures in the analysis.

[0186] Preferably, in step S311 of this embodiment, data acquisition and recording is performed by collecting key parameters such as the head wave height, the height reaching the opposite shore, and the surge pressure distribution in the physical experiment to establish a comprehensive data foundation, and the data becomes an important basis for subsequent evaluation and comparison; the corresponding parameters are obtained from the numerical simulation to ensure the consistency of the two data sets in terms of conditions, time and space, and to ensure the scientificity and effectiveness of the subsequent comparison. The significance achieved is: clear evaluation criteria are set for the accurate evaluation of the model to ensure that the comparison between the physical experimental data and the numerical simulation results is feasible and effective; the original data is provided for the subsequent construction of the relationship diagram to ensure the comprehensiveness and systematicness of the analysis. Step S312 data sorting and normalization processing is performed by corresponding the acquisition time points and spatial positions of the experimental data and the numerical simulation data to establish the consistency of the data and eliminate the errors caused by time delay or spatial offset; the input data is smoothed to remove noise and outliers, improve the stability and reliability of the data, and enhance the accuracy of the analysis results; the head wave height and the height reaching the opposite shore are normalized according to the maximum value, so that the data range is between 0 and 1, avoiding the distortion of the analysis results due to the difference in order of magnitude. Significance achieved: Through normalization, it is convenient to put parameters of different categories and scales in the same coordinate system for comparison, which improves the visualization and comprehensibility of the results; smoothing and normalized data processing can make the comparative analysis clearer, facilitate the identification and quantification of the deviation between the model and the actual situation, and provide an effective basis for subsequent adjustments. Step S313 establishes a relationship diagram and gap quantification. By constructing a two-dimensional coordinate system, in which the X-axis represents the numerical simulation results and the Y-axis represents the target value, it provides an intuitive way to show the relationship between the data; the relationship between the simulation results and the target value is displayed using a scatter plot. Through the intuitive graphical display, the correspondence and difference between the two can be clearly observed; by calculating the distance of each point, the deviation between the simulation results and the target value can be quantitatively identified. Significance achieved: Through the scatter plot, the relationship between the numerical simulation results and the experimental target value can be quickly judged, so as to clearly identify where there is a large deviation; the quantitative deviation analysis provides a scientific basis for subsequent model adjustments, so that researchers can optimize parameters in a targeted manner and improve the prediction ability of the model; it provides visual support for model verification, and can intuitively display the analysis results, which is conducive to scientific decision-making and further research and discussion.

[0187] In summary, this embodiment plays an important role in the overall goal of establishing a relationship diagram between the target value and the current numerical simulation results. It not only helps to accurately evaluate the accuracy of the numerical model, but also provides a solid data foundation and clear path guidance for the optimization and improvement of the model. Through the above steps, establishing a relationship diagram between the target value and the current numerical simulation results is not only a scientific and technological analysis process, but also an important tool for effectively identifying and evaluating the accuracy of the model. Steps such as data sorting, normalization, coordinate system construction and error identification ensure the accuracy and reliability of the model, and lay a solid foundation for subsequent adjustment and optimization. Through such quantitative analysis, the numerical model can be improved more specifically, thereby improving the accuracy of landslide surge simulation and providing support for research in related fields.

[0188] Furthermore, the process of changing the particle diameter and adjusting the support domain and decay rate of the kernel function in step S32 specifically includes the following steps:

[0189] Step S321: extract the distribution characteristics and diameter of the current particles, use the landslide surge numerical model to perform regression analysis on the relationship between the existing particle distribution and surge behavior, and obtain preliminary results; set the range of particle size variation, and select multiple particle diameters for simulation experiments; in the numerical simulation, replace the corresponding particle sizes one by one, and pay attention to the first wave height, the wave height reaching the shore, and the pressure distribution area in the simulation results;

[0190] Step S322: After each adjustment of the particle diameter, numerical simulation is repeatedly performed to compare the simulation results with the experimental target values; by collecting the simulated data, the accuracy of the corresponding wave height, arrival height and pressure distribution is evaluated, and the particle diameter is reduced until it matches the target value;

[0191] Step S323: After the particle size is determined, the support domain of the kernel function is set according to the particle's action distance; according to the actual fluctuations in the simulation results, the decay rate of the kernel function is adjusted, the initial decay rate is set, and fine adjustments are made by comparing the simulation results.

[0192] Preferably, step S321 of this embodiment is particle distribution feature extraction and preliminary simulation experiment, by extracting the diameter and distribution characteristics of the particles, obtaining a quantitative understanding of the particle system, such as average particle size, particle concentration, particle spacing, etc.; establishing a mathematical model between the current particle distribution and surge behavior, identifying the relationship between the two, which may involve linear regression or other regression techniques to reveal the dependency between particle characteristics and wave characteristics (such as wave height, pressure); on the basis of setting the particle size variation range, selecting multiple particle diameters for subsequent simulation experiments. Thereby ensuring the coverage analysis of surge behavior by different particle diameters. Significance achieved: This step provides a baseline understanding of existing numerical models, and provides a basis for theoretically analyzing how particles affect the formation and propagation of waves; through regression analysis, it can guide subsequent particle size adjustments, so as to more effectively explore the impact of particle size changes on surge behavior. Step S322: Particle diameter adjustment and simulation verification. By replacing particles of different diameters one by one, multiple sets of numerical simulations are performed to systematically evaluate the impact of different particle sizes on wave characteristics, including key data collection for multiple indicators such as the first wave height, the wave height reaching the shore, and pressure distribution; the simulation results are compared with the target values, and relevant error indicators (e.g., root mean square error) are calculated to evaluate the model matching degree under each particle size to ensure the accuracy of the data and the effectiveness of the model; after repeated simulations, the particle size range is gradually narrowed to try to find the actual particle size closest to the experimental target value. Significance achieved: This step ensures that the matching degree between the numerical simulation results and the physical experiment is gradually improved, and the accuracy and credibility of the model are improved through gradual debugging; through comparison and feedback, the optimal particle size range can be identified, thereby providing accurate parameter configuration for subsequent research and reducing unnecessary waste of experimental resources. Step S323: Adjustment of kernel function support domain and attenuation rate. According to the effective action distance of the particles, the kernel function support domain is reasonably set so that the interaction can be accurately captured, thus having higher accuracy. The kernel function attenuation rate is preliminarily set and finely adjusted. Through the feedback results of numerical simulation, the influence of the attenuation rate on the simulation results is quantified, in order to achieve output closer to the experimental value. Significance achieved: It can effectively capture the interaction between particles and improve the simulation accuracy of the model, especially when dealing with complex flow conditions. By adjusting the attenuation rate, the model shows better adaptability when dealing with different wave conditions and particle characteristics, thereby improving the reliability and application value of the simulation results.

[0193] In summary, the three steps of this embodiment form an effective feedback loop, from the extraction and preliminary verification of particle characteristics to the optimization and adjustment of specific particle size and kernel function, forming a systematic methodology. It can not only be used for the optimization of landslide surge models, but also can be extended to the research of other complex fluid dynamics phenomena, enhancing the application potential of numerical models in practical engineering and scientific research.

[0194] Further, in step S321, the process of performing regression analysis on the relationship between the existing particle distribution and the surge behavior using the landslide surge numerical model includes the following steps:

[0195] Step S3211: Obtain various data on the particle's motion trajectory, position, diameter, and water surface fluctuation in the surge, analyze the acquired data, automatically mark and identify the particle's boundary; evaluate the influence of each variable on the surge behavior based on the tree model, and select the most influential features; construct cross-features of particle characteristics (diameter, number, shape) and wave behavior (wave height, wave energy, wave period), and generate a new feature combination, namely, wave-particle coupling feature;

[0196] Step S3212: construct a neural network based on deep learning, use the adaptive network structure and hierarchical feature extraction mechanism to fit the complex nonlinear relationship between particle characteristics and surges, analyze the characteristics of the fluctuating data series, and capture the dynamic changes in the time series; combine the equations of fluid dynamics with the deep learning model, and guide the learning of the neural network through physical constraints;

[0197] Step S3213: Evaluate the performance of the neural network model through k-fold cross-validation, and use Bayesian optimization to adjust the hyperparameters to achieve the optimal configuration.

[0198] Preferably, in step S3211 of this embodiment, data acquisition and feature construction are performed to obtain the motion trajectory, position, diameter and other features of the particles, as well as the data of the water surface fluctuations; high-frequency and accurate data are obtained through sensor technology, which helps to fully understand the interaction between particles and water waves; computer vision technology, such as image processing and machine learning algorithms (such as convolutional neural networks in deep learning), is applied to automatically extract particle boundaries. This reduces the need for manual annotation and improves efficiency and accuracy; by evaluating the influence of each variable on the surge behavior through a tree model (such as a random forest), the preliminary relationship between particle characteristics and surge behavior can be obtained, which helps to identify the most influential features and reduce noise in subsequent analysis; wave-particle coupling features are generated, and by combining particle characteristics with wave behavior to generate new feature combinations, it is expected to reflect more complex physical phenomena and improve the output effect and accuracy of the model. The significance achieved: through systematic data acquisition and feature construction, a solid data foundation is laid, providing useful features for subsequent modeling; automatic labeling and boundary recognition realize efficient and accurate particle analysis, which helps to reduce human errors and improve data processing efficiency; by quantifying the influence of variables, the direction for model selection and feature engineering is pointed out, thereby establishing a more accurate prediction model. Step S3212: Deep learning neural network construction. The constructed neural network can capture the complex nonlinear relationship between particle characteristics and wave behavior. The adaptive network structure can realize automatic feature extraction. The application of time series models such as LSTM can analyze the data sequence characteristics of surge fluctuations and capture the dynamic changes in the time series, so that the model can consider the influence of historical data, which helps to improve the accuracy of prediction. The basic equations of fluid dynamics are combined with the deep learning model, so that the output results are reasonable in physical laws. Constraints help limit the learning scope of the model, avoid overfitting, and improve the interpretability of the model. Significance achieved: Through deep learning methods, the model's ability to capture complex nonlinear relationships is enhanced, and the accuracy of prediction is improved; combined with time series modeling, the model can accurately reflect the dynamic behavior that changes over time, and provide a more appropriate simulation for landslide surge phenomena; physical constraints are used to improve the rationality and interpretability of the model, ensuring that the prediction results are not only mathematically valid, but also physically understandable. Step S3213: Model evaluation and optimization. K-fold cross validation can be used to comprehensively evaluate model performance, ensure the robustness of the model on different segmentations, prevent overfitting and underfitting, and effective validation results can increase confidence in model performance. Bayesian optimization methods are used to automatically adjust model hyperparameters, such as learning rate, number of layers, etc., so that the model can find a better parameter configuration during training and improve the overall performance of the model. Significance achieved: Cross validation is used to improve the robustness of the model and ensure feasibility and reliability in practical applications. The hyperparameter adjustment strategy using Bayesian optimization is more efficient than the traditional grid search method, and can find the best hyperparameter configuration in a shorter time, improving model performance.

[0199] In summary, this embodiment achieves a deep understanding and accurate modeling of the relationship between particle distribution and surge behavior in the numerical model of landslide surge through a systematic process from data acquisition, feature engineering to model construction and optimization. Each step provides a scientific basis and decision-making support for the management and prediction of landslide surge. It not only improves the accuracy and effectiveness of the research, but also lays a foundation for practical applications, thereby promoting the theoretical and practical development of fluid dynamics and related fields.

[0200] Furthermore, the process of guiding the learning of the neural network by physical constraints in step S3212 specifically includes the following steps:

[0201] Step S32121: Collect particle dynamics and water surface fluctuation data, annotate particles to identify the boundaries, shapes and positions of different particles; extract particle characteristics (such as particle size, shape, velocity vector in the fluid, etc.) and wave parameters (such as wave height, energy spectrum, frequency, wavelength, etc.); create time window features, and convert all particle characteristics and wave parameters into distributions with a mean of 0 and a variance of 1;

[0202] Among them, the neural network architecture includes:

[0203] Input layer: The input dimension should be consistent with the number of features. For example, for data containing particle characteristics and wave characteristics at 5 time points, the input dimension of each sample is 15 (assuming 3 indicators for each characteristic);

[0204] LSTM layer: construct three layers of LSTM, with the number of hidden units increasing layer by layer (e.g. 100-150-200) to capture different levels of temporal features; apply a Dropout layer after each layer of LSTM, with the dropout ratio set to 0.2 to prevent overfitting;

[0205] Fully connected layer: The LSTM layer is followed by a fully connected layer. The number of nodes in the output layer should be the same as the surge characteristics to be predicted (e.g. wave height, wave speed), and a linear activation function is used;

[0206] Kuang Design: Use adaptive learning rate algorithm, such as Adam optimizer, with the initial learning rate set to 0.001, and further optimized through learning rate decay strategy; the output layer is the surge parameter;

[0207] Step S32122: Particle dynamics and water surface fluctuation data are divided to obtain training sets and validation sets, and loss and accuracy curves of the training sets and validation sets are plotted to observe convergence and stability; the effectiveness of the neural network model is evaluated through a scatter plot of actual wave data and model prediction results;

[0208] Step S32123: Establish a model equation group to describe the fluid dynamics, and calculate the loss through the residual of the neural network model output and the equation; after the neural network model training is completed, use the physical consistency test to check whether the wave and particle behaviors output by the model meet the physical-based expectations;

[0209] in,

[0210]

[0211] Where ρ is the fluid density, u is the velocity field, p is the pressure, μ is the viscosity coefficient, and f is the body force;

[0212] Calculate the loss function expression:

[0213] Loss = L data +λ·L physics

[0214] Among them, L data is the data loss, L physics is the loss of the physical model, and λ is the importance coefficient used to balance the loss.

[0215] Preferably, in step S32121 of this embodiment, data acquisition and preprocessing are performed to collect dynamic information of particles and water surface fluctuation data, and to mark particles, so that the boundaries, shapes and positions of different particles can be identified. This marking lays the foundation for subsequent data processing and feature extraction; particle characteristics (such as particle size, shape, velocity vector) and wave parameters (such as wave height, energy spectrum, frequency, wavelength) are extracted, so that the model can be trained based on rich features, and the features provide the model with necessary information to capture the interaction between particles and waves; all features are converted into distributions with a mean of 0 and a variance of 1, which can ensure that the model converges faster during training and is less sensitive to differences in feature scales, which helps to improve the learning efficiency and accuracy of the model; a neural network architecture including an input layer, an LSTM layer, a fully connected layer, etc. is designed to capture complex dependencies in time series data; setting the number of hidden units of the LSTM layer and the discard ratio of the Dropout layer helps to prevent overfitting and further enhance the generalization ability of the model. The significance achieved: rich and high-quality data features are provided for model input. This is the key prerequisite to ensure the learning effect of the subsequent model; the network structure has been carefully designed to fully mine the time series information in the data and improve the accuracy and reliability of the prediction. Step S32122 Dataset division and model evaluation, the particle dynamics and wave data are divided into training sets and validation sets, providing an independent data source for model training and evaluation, allowing the model to perform performance tests on unseen data; observing the loss and accuracy curves of the training set and validation set can effectively monitor the training process of the model, including convergence and stability, which helps to adjust the training strategy early or in time to avoid overfitting; by comparing the actual wave data with the model prediction results through a scatter plot, the performance and generalization ability of the model can be intuitively evaluated, which is an intuitive test of the model output results and can discover the limitations of the model. Significance achieved: By evaluating the model performance, verification information on the model learning effect is provided, and the transparency and credibility of model development are improved; the comparison of the actual wave data with the model prediction results provides direction for subsequent model improvements and can better guide model optimization. Step S32123 Physical consistency test and equation establishment, using the basic equations of fluid dynamics (such as the continuity equation and momentum equation) to describe the dynamic characteristics of the fluid, provides physical constraints for the output of the neural network. This ensures that the model not only has a good fit on the data, but also follows the laws of physics; by calculating the residual between the output of the neural network model and the physical equation, the two parts of the loss function (data loss and physical loss) are formed. The design of the loss function can effectively integrate data-driven learning with physics-based learning; after the model training is completed, a physical consistency test is performed to ensure that the wave and particle behavior output by the model meets the theoretical expectations, avoiding the model only conforming to the data and losing the ability to explain the physics.Significance achieved: Combining physical constraints to improve the scientific nature of the model, so that the prediction is not only based on data features, but also guided by physical laws; making the output of the model more interpretable and trustworthy; through the balanced design of the loss function, the application ability of the model in real scenarios is optimized, which is helpful for the needs of actual engineering and scientific research.

[0216] In summary, through the synergy of these three steps, this embodiment realizes the effective construction and application of the deep learning model based on physical constraints. The data collection and processing provide a solid foundation for the model, the reasonable design of the network structure ensures the effectiveness of learning, and the physical consistency test enables the model to accurately predict and follow the laws of physics. It will promote the in-depth development of landslide surge research and provide solutions to more complex fluid dynamics problems.

[0217] Furthermore, the process of reducing the particle size until it matches the target value in step S322 specifically includes the following steps:

[0218] Step S3221: Collect the output results of each numerical simulation, which include three parameters: wave height, arrival height and pressure distribution, and use mean square error and absolute error as standardized evaluation indicators;

[0219] Step S3222: Set the accuracy threshold, and judge the deviation of the three parameters of wave height, arrival height and pressure distribution according to the accuracy threshold; compare the simulation results with the target values, generate an error analysis chart, obtain the deviation and trend of the model output under different particle sizes, and judge whether the influence of the current particle size on the surge behavior meets the experimental objectives;

[0220] Step S3223: If the simulation results fail to meet the target value, the particle size is proportionally reduced for the next round of simulation.

[0221] Preferably, step S3221 of this embodiment collects simulation output results. By systematically collecting simulation results, including key parameters such as wave height, arrival height and pressure distribution, a comprehensive understanding of the simulation behavior is ensured; mean square error (MSE) and absolute error (MAE) are used as evaluation standardization indicators to provide a data basis for accuracy judgment; statistics of these indicators can quantify the gap between the simulation results and the experimental target values, so as to effectively compare the parameter performance under different conditions. The significance achieved: providing basic data for analysis, ensuring that all simulations are evaluable, and giving clear quantitative standards so that decisions in the optimization process are based on measurable results; by collecting multiple simulation results, the trend of the impact of particle size adjustment on surge behavior can be identified, so as to follow in subsequent iterations. Step S3222 sets the accuracy threshold. By setting the accuracy threshold, the deviation of wave height, arrival height and pressure distribution is effectively judged to ensure that each simulation has a reasonable success standard; the analysis results are visualized so that the output deviation under different particle size settings is clear at a glance, which is convenient for analyzing and identifying the specific situation of particle size influence. Significance achieved: Setting the accuracy threshold can provide a limit for the model output, ensure that the research direction is clear, and fully focus on finding the appropriate particle size range; error analysis can clearly point out the problem, guide the next particle size adjustment decision, avoid invalid parameter attempts, and improve the efficiency of the entire process. Step S3223 proportionally reduces the particle size for the next round of simulation. According to the results of the previous step, the particle size is systematically adjusted, and the parameter is proportionally reduced for a new test; the progressive adjustment not only reduces the simulation range, but also ensures that the target value is gradually approached; each adjustment is made based on the feedback of the results to ensure that the model parameters are always within a reasonable range, effectively promoting the improvement of accuracy. Significance achieved: By comparing the simulation results with the target values, an effective feedback mechanism can be formed, so that the particle size adjustment is no longer arbitrary, but a scientific decision based on actual data; each fine adjustment can make the model closer to the target, thereby improving the simulation accuracy of the actual surge phenomenon, so that the results can not only achieve the experimental goals, but also have good physical significance.

[0222] In summary, this embodiment shows a closed loop from data collection, error assessment to scientific optimization through these three steps. This makes the entire parameter adjustment process more efficient and targeted, and gradually achieves the preset goals. It not only effectively improves the output accuracy of the simulation model, but also enhances the understanding of the nature of surge behavior, and ultimately promotes research progress in this field.

[0223] Furthermore, the process of proportionally reducing the particle size for the next round of simulation in step S3223 specifically includes the following steps:

[0224] Step S32231: After comparing the simulation results with the target values, identifying the deviation of each parameter;

[0225] Definition symbol: Y sim =(h sim ,z sim ,P sim ) represents the current simulation result vector, h sim represents the simulated wave height, z sim represents the simulated arrival height, P sim represents the simulated pressure distribution, Y target =(h target ,z target ,P target ) represents the target value vector, h target represents the target wave height, z target Indicates the target arrival height, P target represents the target pressure distribution, D i Represents the deviation of the i-th parameter, calculated by:

[0226] where i∈h,z,P

[0227] i = h represents the deviation for wave height, i = z represents the deviation for arrival height, and i = P represents the deviation for pressure distribution;

[0228] Step S32232: Set the comprehensive deviation D to;

[0229] Definition symbol: N represents the number of data points; calculation formula:

[0230]

[0231] ∈The set accuracy threshold, indicating the maximum allowed deviation (for example, the maximum deviation of wave height is 0.1 m);

[0232] Step S32233: reduce the particle size proportionally, and adjust the particle size after recognizing that the deviation D exceeds the accuracy threshold;

[0233] Definition symbol: D current Indicates the current particle size value, D new represents the updated new particle size value, r represents the reduction ratio, and the range is r∈(0,1);

[0234] The formula for updating particle size is: D new =D current ×(1-r);

[0235] For example: If the current particle size (D curren t=0.05) meters, reduce the ratio (r=0.1), then: D new = 0.05 × (1-0.1) = 0.045 m;

[0236] Step S32234: After adjusting the particle size, perform a new numerical simulation to generate a new output result Y sim,new ; Return to step S3221 to re-collect data for evaluation.

[0237] Preferably, step S32231 of this embodiment identifies the deviation of each parameter, and by calculating the deviation, accurately quantifies the gap between the simulation result and the target value of each parameter (wave height, arrival height, pressure distribution), providing a clear basis for particle size adjustment. The significance achieved: It can clearly identify which specific parameters of the simulation results deviate from the target, thereby pointing to the direction that needs to be focused on and optimized, which is crucial in the optimization process to ensure that the improvement measures are targeted at practical problems. Step S32232 sets the comprehensive deviation, and by calculating the comprehensive deviation, the deviations of multiple parameters are incorporated into a unified metric, making the evaluation of the model output quality more comprehensive and scientific. The significance achieved: It provides a comprehensive evaluation standard for the entire model output, which is easy to compare and judge the model performance; it helps to efficiently determine when the next step is required and when the expected goal can be achieved when the particle size is adjusted. Step S32233 Proportionally reduce the particle size by setting a specific strategy for proportionally reducing the particle size, ensuring that each particle size adjustment is carried out within a scientific and reasonable range. Fine control is achieved through formulas to ensure that the new particle size setting is more suitable for actual simulation needs. Significance achieved: It avoids simulation distortion that may be caused by overly drastic adjustments, ensures that the particle size gradually approaches the optimal range, and makes the simulation results gradually approach the target value during the adjustment process, thereby improving the accuracy and reliability of the model. Step S32234: Execute a new round of simulation and collect data. After adjusting the particle size, execute a new round of numerical simulation, generate new output results, and then use the new data for feedback. This forms an effective iterative process that can continuously optimize model parameters. Significance achieved: Through continuous simulation and data collection, it is possible to quickly obtain results and adjust strategies to form a closed-loop feedback mechanism. The ultimate goal is to make the model output continuously close to the experimental goal, greatly improve the effectiveness and scientificity of the simulation, and promote research progress.

[0238] In summary, this embodiment can accurately measure and adjust parameters to ensure that the accuracy of simulation results is gradually improved. This not only helps to improve the performance of the model, but also can significantly improve research efficiency and ensure that the experimental goals are achieved in a shorter time.

[0239] Furthermore, the process of adjusting the decay rate of the kernel function in step S323 specifically includes the following steps:

[0240] Step S3231: estimate the diameter and density of the particle, check the basic principles of fluid dynamics, and determine the influence range of the particle under fluid conditions. The influence range, i.e., the action distance of the particle, refers to the range of influence exerted by the particle on nearby media (such as surges, fluid motion, etc.); determine the support domain of the function according to the range;

[0241] Step S3232: Based on historical experimental data, an empirical decay rate is selected as a basis to set the effect at the beginning of the simulation; during the simulation process, the accuracy of the simulation output is evaluated by comparing it with the actual observation value (such as the wave height and pressure change measured in the field);

[0242] Step S3233: By repeating the simulation experiments multiple times and performing various tests according to different environmental parameters and conditions, the sensitivity of the kernel function to the simulation results under different decay rates is evaluated, and the optimal parameter settings are identified.

[0243] Preferably, in step S3231 of this embodiment, the determination of the influence range and the setting of the support domain are carried out, and the action distance of the particles is estimated by the diameter and density of the particles, so as to clarify the ability and range of the particles to affect the surrounding fluid medium; the evaluation can be based on the basic principles of fluid dynamics, combined with equations related to particle motion (such as Navier-Stokes equations); by reasonably setting the support domain of the kernel function, it is coordinated with the action distance of the particles to ensure that the simulation can accurately reflect the impact of the particles on the surrounding environment. Significance: Through the appropriate setting of the support domain, it is ensured that the numerical values ​​in the region are close to the impact of particles on the fluid in real situations; numerical calculation provides a more scientific basis, making the simulation results more accurate; theoretically, limiting the range of the support domain can reduce the calculation area, thereby reducing the calculation complexity and time, so as to efficiently perform multiple simulations. Step S3232 is the setting and evaluation of the empirical decay rate, and an appropriate empirical decay rate is selected based on historical experimental data so that it can reflect the impact of particles in the current fluid environment; the basic parameter setting provides a starting point for the simulation; during the simulation, the accuracy of the model output is evaluated by comparing with the wave height and pressure changes measured on site, so as to detect the degree of consistency of the simulation. Significance: By comparing with the actual observations, the deviations and inaccuracies in the model can be identified in time, providing a basis for further adjustments and enhancing the adaptability of the model to complex fluid environments; the empirical decay rate setting reflects the actual physical situation to a certain extent. Through this combination, the model has greater practical significance and applicability in theory. Step S3233 simulates and tests parameters multiple times. Through repeated simulations, the influence of the kernel function on the simulation results under different decay rates is tested for different environmental variables (such as flow rate, particle size, density, etc.), and its sensitivity to the simulation output results is clarified; through continuous experimental analysis and data comparison, the optimal decay rate parameter setting is identified to ensure the best simulation effect. Significance: By finding the best parameter settings, guidance is provided for future experimental designs and numerical models, helping to obtain more ideal results under similar conditions; through the simulation of different environmental parameters, not only the decay rate can be adjusted, but also a more comprehensive understanding of the performance of the kernel function itself is obtained, which greatly enhances the adaptability of the model under uncertain conditions in the future.

[0244] In summary, the decay rate of the kernel function in this embodiment can not only be reasonably set and continuously adjusted, but also provide reproducible experimental conditions for numerical simulation, while ensuring high accuracy and reliability of simulation results. Ensuring that the kernel function can effectively perform in complex fluid environments (such as surges, etc.) is a key link, and plays an important role in understanding the interaction between fluid and particle behavior and its predictive ability in actual environments.

[0245] Furthermore, if Figure 6 As shown, the process of the landslide surge hybrid model obtained in step S4 specifically includes the following steps:

[0246] Step S41: combining the modified landslide surge numerical model with the landslide surge physical model to form a hybrid model of landslide surge;

[0247] Step S42: using the obtained hybrid model to analyze various working conditions with different landslide speeds, volumes and terrain conditions; including comparing the transformation dynamics and surge characteristics under actual working conditions, and deriving the mechanism of surge and its influencing factors;

[0248] Step S43: By simulating different parameter scenarios, the hybrid model's resilience to changing conditions is evaluated, and the possible consequences of landslide surges under different circumstances are predicted, including wave height, wave width, and potential dangers caused by surges.

[0249] Preferably, step S41 of this embodiment forms a hybrid model of landslide surge, integrates the modified numerical model of landslide surge with the physical model of landslide surge, and creates a comprehensive hybrid model of landslide surge; ensures that the model has both the depth of theoretical analysis and the reliability of experimental verification; by parameterizing the numerical model and the physical model, verifies the consistency and complementarity between the two, and strengthens the understanding of the landslide surge process. Significance: The hybrid model utilizes the actual data of the physical model and the computational advantages of the numerical model, making the prediction of the dynamics of landslide surge more accurate; by integrating different models, researchers can analyze the landslide surge phenomenon more comprehensively within the theoretical framework, providing more powerful support for practical applications. Step S42 analyzes multiple working conditions of different landslide parameters, uses the hybrid model, and systematically simulates the surge behavior under different landslide speeds, volumes and terrain conditions to obtain a large amount of simulation data; by comparing the simulation results with the actual working conditions, explores the dynamic transformation and surge characteristics in the landslide process, and reveals the mechanism of surge and its influencing factors, such as landslide angle, volume change and water level influence. Significance: It can deeply analyze the specific conditions and environmental factors of landslide surge, thus providing a scientific basis to better understand the laws of landslide surge; understanding the characteristics and mechanisms of surge will help to formulate corresponding monitoring and early warning strategies, improve the effectiveness of natural disaster response, and provide reference for environmental protection and urban planning. Step S43 evaluates the resilience of the hybrid model and predicts potential results. By inserting different parameters (such as landslide velocity, volume and terrain changes), a variety of scenario simulations are carried out to observe the response of the hybrid model under changing conditions; identify and predict the surge characteristics that may be triggered under different landslide conditions, such as wave height, wave width and other potential hazards. Significance: Through the prediction of various scenarios, researchers can evaluate the risks that landslide surges may bring, provide scientific basis for local governments and relevant agencies, and promote the efficiency of emergency management; enhance the stability and applicability of the model in a changeable and unpredictable environment, so that it can be widely used in various landslide surge research and practice.

[0250] In summary, the landslide surge hybrid model obtained through three steps in this embodiment not only strengthens the theoretical research on the landslide surge phenomenon, but also provides strong support for practical applications. The establishment of a comprehensive model enables the scientific community to analyze the interaction between water bodies and landslides from a more comprehensive and systematic perspective, providing a strong theoretical basis and practical guidance for the fields of natural disaster prevention and control, urban planning, and ecological and environmental protection.

[0251] like Figure 7 As shown, this embodiment also provides an embodiment of a landslide surge physical and numerical hybrid model construction system. In this embodiment, the landslide surge physical and numerical hybrid model construction system is applied to the landslide surge physical and numerical hybrid model construction method in the above-mentioned embodiment. The landslide surge physical and numerical hybrid model construction system includes a first result acquisition module 1, a second result acquisition module 2, a test parameter adjustment module 3, and a hybrid model acquisition module 4 which are electrically connected in sequence.

[0252] Among them, the first result acquisition module 1 is used to obtain the contour geological information of the target area, the height of the center of gravity of the sliding body and the water level information; construct a landslide surge physical model test based on the information, and the landslide surge physical model test includes constructing the terrain according to the contour information and constructing the position and volume of the sliding body according to the volume and center of gravity of the sliding body; according to the first result of the landslide surge physical model test, analyze the landslide instability evolution movement process, the height of the first wave entering the landslide reservoir and the surge propagation process of the landslide surge; the second result acquisition module 2 is used to construct a landslide surge numerical model test based on the information, and the landslide surge numerical model test includes constructing the terrain according to the contour information and constructing the position and volume of the sliding body according to the sliding body volume and center of gravity. The volume and the center of gravity are used to construct the position and volume of the sliding body; a numerical simulation test is carried out according to the second result of the landslide surge numerical model test; the test parameter adjustment module 3 is used to take the first wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, and adjust the particle size and the kernel function in the landslide surge numerical simulation test to make the first wave height, the height reaching the opposite bank and the surge pressure distribution in the landslide surge numerical simulation test and the landslide surge physical model test consistent; the hybrid model acquisition module 4 is used to use the adjusted landslide surge numerical model as the obtained landslide surge hybrid model to analyze the working conditions of different landslide velocities, landslide volumes and terrain conditions.

[0253] Preferably, the first result acquisition module 1 of this embodiment obtains the contour geological information, the center of gravity height of the sliding body and the water level information of the target area, which provides important basic information for model construction; based on the information obtained, a physical model of landslide surge is created, which can intuitively experiment and observe the landslide phenomenon. Significance: It provides necessary basic data for model experiments to ensure the authenticity and relevance of the model; through the testing of the physical model, it helps to deeply understand the dynamic process of landslide instability and its consequences, and provide a basis for analysis. The second result acquisition module 2 constructs a numerical model of landslide surge based on the information, conducts quantitative analysis and simulation; and conducts numerical simulation experiments to obtain specific numerical data of the landslide surge process. Significance: The numerical model can be theoretically extended to multiple different cases, and more accurate predictions of landslide surge behavior can be obtained through calculations; compared with physical model experiments, numerical simulations can quickly obtain results under different working conditions, saving time and resources. The test parameter adjustment module 3 uses the head wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, adjusts the particle size and kernel function in the numerical model, and ensures the consistency between the numerical simulation and the physical experiment; through the systematic adjustment of the model parameters, the numerical results are more consistent with the actual observation. Significance: The calibration process ensures the reliability of the numerical model, so that it can more realistically reflect the landslide surge process and provide higher simulation accuracy; the adjusted model can be better applied to actual projects and provide a scientific basis for disaster prevention and mitigation. The hybrid model acquisition module 4 combines the adjusted numerical model with the physical model to form a comprehensive landslide surge hybrid model; conducts a comprehensive analysis of different landslide velocities, volumes and terrain conditions, and penetrates a wider range of application scenarios and conditions. Significance: Through the establishment of the hybrid model, the understanding of landslide and surge behavior is more profound, and the influence of multiple factors can be comprehensively considered; based on the results of the model analysis, it can provide a scientific basis for related engineering design, safety assessment and emergency plans, and promote the implementation of practical applications.

[0254] In summary, through the collaborative work of various modules, the landslide surge physical and numerical hybrid model construction system in this embodiment realizes the complete process from data acquisition, model establishment, parameter adjustment to result analysis. This makes the research on landslide surge not only more scientific but also more practical; it helps to improve the ability of landslide monitoring and early warning, reduce potential risks, and protect life and property safety.

[0255] The landslide surge physical model test of this embodiment constructs a landslide surge physical model that meets the engineering geological characteristics, which can accurately reflect the characteristics of landslide surge, and its results have important guiding significance for the prevention and control decision-making of landslide surge. However, due to the limitation of the number of monitoring equipment and the influence of terrain accuracy, its results mainly reflect the macro characteristics of landslide surge disasters, and it is difficult to accurately capture the various details of landslide surge disasters. The use of numerical simulation methods to study landslide surge problems has the advantages of high efficiency and intuitive results, and can clearly depict the details of landslide surge disasters. The calculation results of the numerical simulation method are greatly affected by the calculation parameters, and some parameters are often difficult to determine, resulting in relatively low credibility of the numerical simulation results. To this end, this embodiment combines the advantages of physical model tests and numerical simulations, based on large-scale physical model tests, uses the results of physical model experiments to iteratively correct the numerical model, establishes a physical-numerical hybrid model, and comprehensively evaluates landslide surge disasters.

[0256] like Figure 8 As shown, this embodiment provides an embodiment of an electronic device. In this embodiment, the electronic device 4 includes a processor 41 and a memory 42 coupled to the processor 41.

[0257] The memory 42 stores program instructions for implementing the method for constructing a landslide surge physical and numerical hybrid model according to any of the above embodiments.

[0258] The processor 41 is used to execute the program instructions stored in the memory 42 to construct a landslide surge physical and numerical hybrid model.

[0259] The processor 41 may also be referred to as a CPU (Central Processing Unit). The processor 41 may be an integrated circuit chip having signal processing capabilities. The processor 41 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0260] Further, Fig. 9The schematic diagram of the structure of the storage medium of an embodiment of the present application is that the storage medium 5 of the embodiment of the present application stores program instructions 51 that can implement all the above methods, wherein the program instructions 51 can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0261] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0262] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional units. The above is only an implementation mode of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present invention.

[0263] The specific implementation methods of the invention are described in detail above, but they are only examples, and the invention is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the invention, therefore, the equalization, modification, improvement, etc. made without departing from the spirit and principle of the invention should be included in the scope of the invention.

Claims

1. A method for constructing a landslide surge physical and numerical hybrid model, characterized in that: The method for constructing a landslide surge physical and numerical hybrid model includes: Obtain the contour geological information of the target area, the height of the center of gravity of the sliding body and the water level information; construct the landslide surge physical model test based on the information, which includes constructing the terrain based on the contour information and constructing the position and volume of the sliding body based on the volume and center of gravity of the sliding body; analyze the landslide instability evolution movement process, the height of the first wave entering the landslide reservoir and the surge propagation process of the landslide surge based on the first result of the landslide surge physical model test; Constructing a landslide surge numerical model test based on the information, the landslide surge numerical model test includes constructing terrain based on contour information and constructing the position and volume of the sliding body based on the volume and center of gravity of the sliding body; conducting a numerical simulation test based on the second result of the landslide surge numerical model test; Taking the first wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, the particle size and kernel function in the numerical simulation test of landslide surge were adjusted to make the first wave height, the height reaching the opposite bank and the surge pressure distribution in the numerical simulation test of landslide surge and the physical model test of landslide surge consistent. The adjusted landslide surge numerical model is used as the obtained landslide surge hybrid model to analyze the working conditions of different landslide velocities, landslide volumes and terrain conditions.

2. The method for constructing a landslide surge physical and numerical hybrid model according to claim 1, characterized in that: The process of adjusting the particle size and kernel function in the numerical simulation of landslide surge includes: Analyze the experimental data obtained, including extracting the specific values ​​of the head wave height, the height reaching the opposite shore, and the surge pressure distribution, which will be used as the evaluation criteria; establish a relationship diagram between the target value and the current numerical simulation results, and identify the gap between the numerical simulation and the physical experimental results; Analyze the particle distribution characteristics in the landslide surge physical model and compare the effects of different particle sizes on the numerical results. Perform numerical simulations by changing the particle diameter and continuously refine the particle size until it matches the results of the landslide surge physical model. Based on the analysis of surge behavior in the landslide surge physical model experiment, confirm the kernel function type and adjust the support domain and attenuation rate of the kernel function. After completing the adjustment of the particle size and kernel function, a new round of numerical simulation is carried out: the adjusted numerical model is comprehensively compared with the physical experimental results, including the verification of the first wave height, the detection of the height reaching the opposite shore, and the fitting of the surge pressure distribution.

3. The method for constructing a landslide surge physical and numerical hybrid model according to claim 2, characterized in that: The process of establishing a relationship diagram between the target value and the current numerical simulation result includes: Obtain multiple key parameters such as the first wave height, the height reaching the opposite shore, and the surge pressure distribution for setting target values; at the same time, obtain the corresponding results from the numerical simulation and record the first wave height, the height reaching the opposite shore, and the corresponding surge pressure distribution data obtained under the same conditions; The collected landslide surge physical model experimental data and the second result of numerical simulation are matched with the time point and spatial position of collection, and smoothed; the height of the first wave and the height reaching the opposite bank are normalized according to the maximum value, respectively, to generate a standardized value between 0 and 1; A two-dimensional coordinate system is used, with the X-axis representing the simulation result and the Y-axis representing the experimental target value. A scatter plot is made to show the relationship between the simulation result and the target value. The distance of each point is calculated to quantify the gap and obtain the deviation between the numerical simulation result and the target value.

4. The method for constructing a landslide surge physical and numerical hybrid model according to claim 3, characterized in that: in, There are n pairs of simulation results and target values ​​of key parameters, which are expressed as follows: Numerical simulation results S i , where i = 1, 2, ..., n, target value T i ; Absolute error refers to the direct difference between the numerical simulation result and the target value, and the calculation formula is: E abs,i =|S i -T i | Among them, E abs,i represents the absolute error of the i-th key parameter; The relative error takes into account the impact of the target value and uses the ratio of the target value to standardize the error. The calculation formula is: Among them, E rel,i Represents the relative error of the i-th key parameter; The root mean square error RMSE is used to comprehensively evaluate the deviations of multiple parameters. The calculation formula is: The average of the squared errors for all key parameters was calculated and then the square root was taken, providing a more comprehensive error measure; The comprehensive deviation evaluation evaluates each key parameter together with the absolute error, relative error and root mean square error to form a new indicator: Among them, α, β, and γ are weight coefficients, reflecting the importance of different deviation measures in the analysis.

5. The method for constructing a landslide surge physical and numerical hybrid model according to claim 2, characterized in that: The process of changing the particle diameter and adjusting the support domain and decay rate of the kernel function includes: Extract the distribution characteristics and diameter of the current particles, use the landslide surge numerical model to perform regression analysis on the relationship between the existing particle distribution and surge behavior, and obtain preliminary results; set the range of particle size variation, and select multiple particle diameters for simulation experiments; in the numerical simulation, replace the corresponding particle sizes one by one, and focus on the first wave height, the wave height reaching the shore, and the pressure distribution area in the simulation results; After each adjustment of the particle diameter, numerical simulations were repeated to compare the simulation results with the experimental target values. By collecting the simulated data, the accuracy of the corresponding wave height, arrival height and pressure distribution was evaluated, and the particle size was reduced until it matched the target value. After the particle size is determined, the support domain of the kernel function is set according to the distance of the particle. According to the actual fluctuations in the simulation results, the decay rate of the kernel function is adjusted, the initial decay rate is set, and fine adjustments are made by comparing the simulation results.

6. The method for constructing a landslide surge physical and numerical hybrid model according to claim 5, characterized in that: The process of regression analysis of the relationship between the existing particle distribution and surge behavior using the landslide surge numerical model includes: Obtain various data on the particle's trajectory, position, diameter, and water surface fluctuations in the surge, analyze the acquired data, and automatically mark and identify the particle's boundaries; use a tree-based model to evaluate the impact of each variable on surge behavior and select the most influential features; construct cross-features of particle characteristics and wave behavior to generate a new feature combination, namely, wave-particle coupling features; Construct a neural network based on deep learning, use adaptive network structure and hierarchical feature extraction mechanism to fit the complex nonlinear relationship between particle characteristics and surge, analyze the characteristics of fluctuating data series, and capture dynamic changes in time series; combine fluid dynamics equations with deep learning models to guide the learning of neural networks through physical constraints; The performance of the neural network model is evaluated through k-fold cross-validation, and the hyperparameters are adjusted using Bayesian optimization to achieve the optimal configuration.

7. The method for constructing a landslide surge physical and numerical hybrid model according to claim 6, characterized in that: The process of guiding the learning of a neural network through physical constraints includes: Collect particle dynamics and water surface fluctuation data, annotate particles to identify the boundaries, shapes and positions of different particles; extract particle characteristics and wave parameters; create time window features to convert all particle characteristics and wave parameters into distributions with a mean of 0 and a variance of 1; The particle dynamics and water surface fluctuation data are divided into training sets and validation sets. The loss and accuracy curves of the training set and validation set are plotted to observe the convergence and stability. The effectiveness of the neural network model is evaluated through the scatter plot of the actual wave data and the model prediction results. Establish a set of model equations to describe fluid dynamics, and calculate the loss through the residual of the neural network model output and the equation; after the neural network model training is completed, use physical consistency testing to check whether the wave and particle behaviors output by the model meet the physically based expectations; in, Where ρ is the fluid density, u is the velocity field, p is the pressure, μ is the viscosity coefficient, and f is the body force; Calculate the loss function expression: Loss=L data +λ·L physics Among them, L data is the data loss, L physics is the loss of the physical model, and λ is the importance coefficient used to balance the loss.

8. The method for constructing a landslide surge physical and numerical hybrid model according to claim 5, characterized in that: The process of reducing the particle size until it matches the target value includes: Collect the output results of each numerical simulation, which include three parameters: wave height, arrival height and pressure distribution. The mean square error and absolute error are used as standardized evaluation indicators. Set the accuracy threshold and judge the deviation of the three parameters of wave height, arrival height and pressure distribution according to the accuracy threshold; compare the simulation results with the target values, generate error analysis charts, obtain the deviation and trend of the model output under different particle sizes, and judge whether the influence of the current particle size on the surge behavior meets the experimental objectives; If the simulation results fail to meet the target value, the particle size is proportionally reduced for the next round of simulation; The process of proportionally reducing the particle size for the next round of simulation includes: After comparing the simulation results with the target values, the deviation of each parameter is identified; Definition symbol: Y sim =(h sim ,z sim ,P sim ) represents the current simulation result vector, h sim represents the simulated wave height, z sim Indicates the simulated arrival height, P sim represents the simulated pressure distribution, Y target =(h target ,z target ,P target ) represents the target value vector, h target represents the target wave height, z target Indicates the target arrival height, P target represents the target pressure distribution, D i Represents the deviation of the i-th parameter, calculated by: where i∈h,z,P i = h represents the deviation for wave height, i = z represents the deviation for arrival height, and i = P represents the deviation for pressure distribution; Set the comprehensive deviation D to; Definition symbol: N represents the number of data points; calculation formula: ∈The set accuracy threshold, indicating the maximum allowed deviation); Reduce the particle size proportionally, and adjust the particle size after identifying that the deviation D exceeds the accuracy threshold; Definition symbol: D current Indicates the current particle size value, D new represents the updated new particle size value, r represents the reduction ratio, and the range is r∈(0,1); The formula for updating particle size is: D new =D current ×(1-r); For example: If the current particle size (D current =0.05) meters, reduce the ratio (r = 0.1), then: D new = 0.05 × (1-0.1) = 0.045 m; After adjusting the particle size, a new numerical simulation is performed to generate a new output result Y sim,new ; Return to re-collect data for evaluation.

9. A landslide surge physical and numerical hybrid model construction system, which is applied to the landslide surge physical and numerical hybrid model construction method according to any one of claims 1 to 8, characterized in that: The landslide surge physical and numerical hybrid model construction system includes: The first result acquisition module is used to obtain the contour geological information of the target area, the height of the center of gravity of the sliding body and the water level; construct the landslide surge physical model test according to the information, and the landslide surge physical model test includes constructing the terrain according to the contour information and constructing the position and volume of the sliding body according to the volume and center of gravity of the sliding body; according to the first result of the landslide surge physical model test, analyze the landslide instability evolution movement process, the height of the first wave entering the landslide reservoir and the surge propagation process of the landslide surge; The second result acquisition module is used to construct a landslide surge numerical model test according to the information, wherein the landslide surge numerical model test includes constructing the terrain according to the contour information and constructing the position and volume of the sliding body according to the volume and center of gravity of the sliding body; and performing a numerical simulation test according to the second result of the landslide surge numerical model test; The test parameter adjustment module is used to use the first wave height, the height reaching the opposite bank and the surge pressure distribution as the evaluation criteria, and adjust the particle size and kernel function in the landslide surge numerical simulation test to make the first wave height, the height reaching the opposite bank and the surge pressure distribution in the landslide surge numerical simulation test and the landslide surge physical model test consistent; The hybrid model acquisition module is used to use the adjusted landslide surge numerical model as the obtained landslide surge hybrid model to analyze the working conditions of different landslide velocities, landslide volumes and terrain conditions.

10. An application of a landslide surge hybrid model, characterized in that: The landslide surge hybrid model is obtained by the landslide surge physical and numerical hybrid model construction method according to any one of claims 1 to 8, and the landslide surge hybrid model analyzes the working conditions of different landslide velocities, landslide volumes and terrain conditions.

Images