A method and device for predicting stability of multi-layer slopes with rainfall and vegetation

Abstract

The present invention provides a method and device for predicting the stability of a rainfall-vegetated multi-soil-layer slope, which includes obtaining rainfall parameters and slope parameters of each layer of the slope, wherein the slope includes a vegetation layer and several soil layers; substituting the rainfall parameters and slope parameters into a preset vegetation multi-layer soil slope model to calculate the safety factor of the slope; taking the rainfall parameters, slope parameters and slope safety factor as a group of samples to obtain a data set containing multiple groups of samples; using the data set to train a rainfall-vegetated multi-soil-layer slope stability prediction model; obtaining the slope parameters of a target slope and real-time or predicted rainfall parameters, and inputting the slope parameters and rainfall parameters of each layer of the target slope into the trained rainfall-vegetated multi-soil-layer slope stability prediction model to predict the safety factor of the target slope. The present invention takes into account the influence of vegetation and multiple soil layers on the stability of the soil slope, thereby being able to obtain a more accurate soil slope stability prediction result.

Description

Technical Field

[0001] The invention relates to the technical field of slope stability, and in particular to a method and a device for predicting the stability of a rainfall-vegetated multi-soil-layer slope. Background Art

[0002] As global climate change accelerates, slope stability has become an increasingly prominent issue. Extreme rainfall increases the risk of slope disasters, posing a major challenge to human and infrastructure safety.

[0003] Currently, slope stability assessment relies on experience-based methods, which usually involve the use of historical data and expert knowledge to estimate the stability of the slope. Although these methods are effective in some cases, they often lack comprehensive consideration of complex geological conditions and environmental factors. Although traditional physical simulation methods play an important role in slope stability assessment, they have obvious limitations when facing large-scale and complex terrain slope stability problems. Each slope needs to be modeled in detail, which consumes a lot of computing resources and time. In addition, they are limited by computing power and storage space when dealing with wide areas, which may cause the simulation process to become slow or even impossible to complete.

[0004] Some studies have used statistical analysis methods to assess slope stability and predict the possibility of slope instability by analyzing the relationship between historical landslide events and factors such as rainfall and soil properties. Although these methods can provide certain predictive capabilities, they are usually limited by the quality and quantity of available data. Sensor-based slope stability monitoring and early warning systems have been established in some areas. These systems can issue alarms when slopes show signs of instability, but they usually require a lot of sensor deployment and maintenance work, and are limited in predicting long-term stability issues affected by rainfall and vegetation.

[0005] In addition, although existing technologies have made some progress in slope stability assessment, they are limited in dealing with slope stability assessments with multiple soil layers, vegetation coverage, and rainfall. These models generally lack a deep understanding of the mechanical properties under the combined effects of vegetation and rainfall conditions, which limits their interpretability in practical application scenarios. Vegetation roots can enhance the shear strength of the soil and reduce soil permeability. At the same time, the different moisture and matrix suction states of unsaturated soils directly affect their permeability and shear strength, thereby affecting the stability of the slope. Summary of the invention

[0006] The present invention provides a method and device for predicting the stability of a rainfall-vegetation-multi-soil-layer slope, so as to consider the influence of vegetation, rainfall state and multi-soil-layer factors on the slope stability.

[0007] According to one aspect of the present invention, a method for predicting stability of a rainfall-vegetated multi-layer slope is provided, comprising:

[0008] Obtaining rainfall parameters and slope parameters of each layer of the slope, wherein the slope includes a vegetation layer and several soil layers;

[0009] Substituting rainfall parameters and slope parameters of each layer of the slope into a preset vegetation multi-layer soil slope model to calculate the safety factor of the slope;

[0010] The rainfall parameter, the slope parameter of each layer of the slope and the safety factor of the slope are taken as a group of samples to obtain a data set containing multiple groups of samples; the data set contains multiple groups of samples with different rainfall parameters and different slope parameters;

[0011] The data set is used to train a rainfall-vegetated multi-soil-layer slope stability prediction model to obtain a trained rainfall-vegetated multi-soil-layer slope stability prediction model;

[0012] Obtain the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters, and input the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters into the trained rainfall-vegetation multi-soil layer slope stability prediction model to predict the safety factor of the target slope.

[0013] Optionally, the slope parameters of each layer of the slope include at least the slope height, slope angle, initial groundwater level, saturated permeability coefficient of each layer, residual moisture content of each layer of soil, saturated moisture content of each layer, initial cohesion of each layer, initial internal friction angle of each layer, and VG model parameters;

[0014] The rainfall parameters include at least rainfall intensity, rainfall duration, and rainfall type.

[0015] Optionally, the pre-set vegetation multi-layer soil slope model includes:

[0016] An unsaturated seepage control model considering different rainfall parameters and initial groundwater levels, a moisture content model considering any point in each layer of a slope with vegetation and multiple soil layers, a seepage model, a shear strength calculation model for unsaturated soil in each layer of a slope with vegetation and multiple soil layers, and a safety factor calculation model for a slope with vegetation and multiple soil layers under different rainfall conditions.

[0017] Optionally, the pre-set vegetation multi-layer soil slope model includes:

[0018] The moisture content model for any point in each layer of a multi-layer slope including vegetation is:

[0019]

[0020] θ si ,θ ri is the saturated water content and residual water content in the i-th layer of soil; H is the slope height; H i represents the height of a point in the i-th soil layer; represents the VG model parameters related to the soil characteristics of the i-th layer; ni and mi represent the VG model parameters of the i-th layer; h i Indicates the height difference of each layer; Indicates the matrix suction of the vegetation layer; i=1, vegetation layer, h0=0;

[0021] The classic VG saturated-unsaturated seepage model is:

[0022]

[0023] In the formula, k wi represents the permeability coefficient of the i-th soil layer; represents the matrix suction of the i-th soil layer; k 0i is the saturated permeability coefficient of the i-th soil layer;

[0024] The unsaturated seepage control model of soil slope under seepage action under different rainfall parameters and initial groundwater levels is:

[0025]

[0026] H 总 represents the total water head, which is related to rainfall intensity, rainfall duration, initial groundwater level and slope angle; k xi , k yi are the horizontal permeability and vertical permeability of the i-th layer, and the permeability k of the i-th soil layer. wi Correlation; θ w (h i ) is the moisture content of the i-th layer of the soil slope during the seepage process; φ i is the matrix suction of the i-th layer; x is the horizontal direction, y is the vertical direction; t is the rainfall time, that is, the unsaturated seepage control is related to the rainfall duration;

[0027] The shear strength model of each layer of unsaturated soil including vegetation and multi-layer slope is:

[0028]

[0029] is the shear strength of unsaturated soil; c' is the effective cohesion; is the effective internal friction angle; represents the matrix suction of the i-th layer; is the net normal stress, which varies with the moisture content of the soil layer; is the suction parameter, which changes with the moisture content of the soil layer; the effective cohesion of each layer is related to the initial cohesion of each layer, the moisture content of any point in each layer, and the soil quality; the effective internal friction of each layer is related to the initial internal friction of each layer, the moisture content of any point in each layer, and the soil quality;

[0030] The safety factor calculation model of vegetated multi-soil slope under different rainfall conditions is:

[0031]

[0032] C is the cohesion of the slope after rainfall; is the internal friction angle of the slope after rainfall; , are the cohesion and internal friction angle at critical failure of the slope, respectively.

[0033] Optionally, also include:

[0034] Using finite element strength reduction method to obtain the cohesion of slope at critical failure and internal friction angle .

[0035] Optionally, also include:

[0036] A rainfall-vegetation multi-layer slope stability prediction model based on the random forest model was constructed.

[0037] Optionally, the using the data set to train a rainfall-vegetation multi-soil layer slope stability prediction model includes:

[0038] Divide the dataset into training and testing sets;

[0039] For each training sample, the parameters of the random forest model are optimized using a red-billed blue magpie optimization algorithm to obtain the optimal parameter configuration of the random forest model for each training sample;

[0040] The random forest model is tested using a test set to evaluate the performance of the random forest model.

[0041] Optionally, the optimization of the parameters of the random forest model using the red-billed blue magpie optimization algorithm includes:

[0042] Defining a search space for hyperparameters; the hyperparameters include the number of trees, the maximum depth of each tree, the minimum number of node split samples for each tree, the minimum number of leaf node samples for each tree, and the maximum number of features;

[0043] Generate an initial population, each individual in the initial population represents a candidate solution for a set of hyperparameters;

[0044] Find the goal of each individual. The goal formula is:

[0045]

[0046] is the mean of the individual group targets; Q is the group size; is the mth individual in this group; m is a constant greater than or equal to 1 and less than or equal to Q; N is a random number; The target of the current individual calculated for this iteration; The target of the current individual calculated in the previous iteration;

[0047] Get the optimal solution found in the population of this iteration. The formula for the optimal solution is:

[0048]

[0049] is the optimal solution currently found; M is the adjustment coefficient; is a normally distributed random number;

[0050]

[0051] The adjustment factor M is:

[0052]

[0053] t is the current iteration number; T is the maximum iteration number;

[0054] The fitness of individuals in each iteration is evaluated by minimizing the loss function. The fitness evaluation formula is:

[0055]

[0056] In the formula, is the safety factor of each sample slope; The machine learning modeling predicted value for each sample; n p Indicates the number of samples;

[0057] When the number of iterations reaches the maximum number of iterations or the individual fitness is less than or equal to the fitness threshold, the iteration is stopped and the optimal hyperparameters are output. The optimal hyperparameters define the optimal parameter configuration of the random forest model.

[0058] According to a second aspect of the present invention, there is provided a device for predicting stability of a rainfall-vegetated multi-soil-layer slope, comprising:

[0059] A parameter acquisition unit, used to acquire rainfall parameters and slope parameters of each layer of the slope, wherein the slope includes a vegetation layer and several soil layers;

[0060] A calculation unit, used for substituting the rainfall parameter and the slope parameter of each layer of the slope into the preset vegetation multi-layer soil slope model to calculate the safety factor of the slope;

[0061] A data construction unit is used to take the rainfall parameter, the slope parameter of each layer of the slope and the safety factor of the slope as a group of samples to obtain a data set containing multiple groups of samples; the data set contains multiple groups of samples with different rainfall parameters and different slope parameters;

[0062] A training unit, used for training a rainfall-vegetated multi-soil-layer slope stability prediction model using the data set to obtain a trained rainfall-vegetated multi-soil-layer slope stability prediction model;

[0063] A prediction unit is used to obtain the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters, and input the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters into the trained rainfall-vegetation multi-soil layer slope stability prediction model to predict the safety factor of the target slope.

[0064] Optionally, the training unit is further used to divide the data set into a training set and a test set;

[0065] For each training sample, the parameters of the random forest model are optimized using a red-billed blue magpie optimization algorithm to obtain the optimal parameter configuration of the random forest model for each training sample;

[0066] The random forest model is tested using a test set to evaluate the performance of the random forest model.

[0067] The technical solution of the embodiment of the present invention is to obtain rainfall parameters and slope parameters of each layer of the slope, the slope includes a vegetation layer and several soil layers, and establish a vegetation multi-layer soil slope model that takes into account the combined effects of vegetation, rainfall status and multiple soil layers to analyze the slope stability. The present invention can more realistically reflect the actual state of the slope containing multiple soil layers of vegetation under rainfall conditions, and can obtain a more accurate slope safety factor. By introducing a machine learning model for predicting the stability of a rainfall vegetation multi-soil layer slope, a rainfall vegetation multi-soil layer slope stability prediction model with good generalization ability is obtained through training with a large number of soil slope data samples, and a stability analysis result that is closer to the actual situation is predicted. On the other hand, the present invention can also quickly predict the safety of the slope in real time or in advance and issue an alarm based on the slope parameters of the transplant target slope and the real-time or predicted rainfall parameters.

[0068] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0070] Figure 1 It is a method flow chart of a method for predicting stability of a rainfall-vegetated multi-soil layer slope provided by the present invention;

[0071] Figure 2 A schematic diagram of an embodiment of a vegetation multi-layer soil slope model provided by the present invention;

[0072] Figure 3 is a schematic diagram of a finite element model established based on COMSOL in one embodiment of the present invention;

[0073] Figure 4 A curve showing the change of soil water saturation with rainfall time under uniform rainfall conditions in one embodiment of the present invention;

[0074] Figure 5 (a) is a distribution diagram of soil saturation in a vegetation multi-layer soil slope model at the rainfall start time 0h in one embodiment of the present invention;

[0075] Figure 5 (b) is a distribution diagram of soil saturation in a vegetated multi-layer soil slope model 2 hours after the start of rainfall in one embodiment of the present invention;

[0076] Figure 5 (c) is a distribution diagram of soil saturation in a vegetated multi-layer soil slope model 6 hours after the start of rainfall in one embodiment of the present invention;

[0077] Figure 5 (d) is a distribution diagram of soil saturation in a vegetated multi-layer soil slope model 9 hours after the start of rainfall in one embodiment of the present invention;

[0078] Figure 6 A comparison diagram of the predicted value and the true value of the vegetation multi-layer soil slope model under rainfall environment using the red-billed blue magpie optimized random forest model of the present invention;

[0079] Figure 7It is a structural schematic diagram of a rainfall vegetation multi-soil layer slope stability prediction device in an embodiment of the present invention. DETAILED DESCRIPTION

[0080] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0081] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof 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 necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0082] Figure 1 The present invention provides a method for predicting the stability of a rainfall-vegetated multi-soil slope. Figure 1 As shown, the method includes:

[0083] S101. Obtain rainfall parameters and slope parameters of each layer of the slope, where the slope includes a vegetation layer and several soil layers.

[0084] It should be noted that the actual slope usually includes a vegetation layer and multiple soil layers. Since the vegetation root system can enhance the shear strength of the soil and reduce the soil permeability, it will have a positive effect on the strength of the slope. In addition, soil layers of different soil types will also have a greater impact on the strength of the slope.

[0085] The present invention can adopt an on-site collection method to obtain soil samples of any slope, and determine parameters such as slope height, slope angle, groundwater level, etc. through on-site geological survey; determine the slope parameters of each layer in the slope soil through geotechnical tests. It can be clearly seen that different soil layers may include sand layers, clay layers, rock and soil layers, etc., and the slope parameters of different soil layers are different.

[0086] In one embodiment, the slope parameters include at least slope height, slope angle, initial groundwater level, saturated permeability coefficient of each layer, residual moisture content of each layer, saturated moisture content of each layer, initial cohesion of each layer, initial internal friction angle of each layer, VG model parameters, and may also include elastic modulus, Poisson's ratio and density of each layer.

[0087] The rainfall parameters can be obtained by obtaining the historical rainfall information of the area where the corresponding slope is located. In one embodiment, the rainfall parameters include rainfall intensity, rainfall duration, and rainfall type. Among them, the rainfall types include uniform type, front type, back peak type, intermittent type, etc. Uniform rainfall refers to rainfall in which the rainfall intensity remains basically unchanged during the entire rainfall period. Frontal rainfall is usually accompanied by a long period of continuous rainfall, and the intensity may gradually change and weaken. Back peak rainfall has a low intensity at the beginning, gradually increases with the passage of time, and then gradually weakens after reaching a peak value. It is more common in some convective rain or frontal rainfall. Intermittent rainfall refers to interruptions in the rainfall process, that is, the rainfall is not continuous, but sometimes there is no rainfall, and there may be multiple rainfalls and pauses alternating.

[0088] In addition, the acquired rainfall parameters may include historical rainfall parameters of different regions, and the slope parameters of each layer of the slope may include slope parameters of multiple slopes in the region where each rainfall parameter is located, so that the rainfall parameters and the slope parameters of each layer of the slope corresponding to the rainfall parameters constitute a set of data.

[0089] S102, substituting rainfall parameters and slope parameters of each layer of the slope into a preset vegetation multi-layer soil slope model to calculate the safety factor of the slope.

[0090] The pre-set vegetation multi-layer soil slope model in this embodiment is a slope model constructed by considering factors including vegetation and multi-layer soil. By substituting the obtained rainfall parameters of each group and the slope parameters of each layer of the slope into the pre-set vegetation multi-layer soil slope model, an accurate slope safety factor can be obtained because the model is more in line with the actual slope structure.

[0091] S103: taking the rainfall parameter, the slope parameter of each layer of the slope, and the safety factor of the slope as a group of samples, and obtaining a data set including a plurality of groups of samples.

[0092] A set of rainfall parameters, slope parameters of each layer of the slope, and the calculated safety factor of the slope are combined into a set of samples, wherein the safety factor of the slope can be used as a label of the sample. The data set includes multiple sets of samples with different rainfall parameters and different slope parameters.

[0093] S104, using the data set to train a rainfall-vegetated multi-soil-layer slope stability prediction model to obtain a trained rainfall-vegetated multi-soil-layer slope stability prediction model.

[0094] It should be noted that the rainfall-vegetation multi-soil layer slope stability prediction model can adopt a machine learning model, such as a random forest model, to train a data set containing different rainfall and different slope samples.

[0095] S105, obtaining slope parameters of each layer of the target slope and real-time or predicted rainfall parameters, and inputting the slope parameters of each layer of the target slope and real-time or predicted rainfall parameters into the trained rainfall-vegetation multi-soil layer slope stability prediction model to predict the safety factor of the target slope.

[0096] The soil samples of the target slope can be obtained by on-site collection methods, and the slope height, slope angle, groundwater level and other parameters can be determined through on-site geological surveys; the slope parameters of each layer in the slope soil can be determined through geotechnical tests; and real-time or predicted rainfall parameters can be obtained. The slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters are input into the trained rainfall vegetation multi-soil layer slope stability prediction model to predict the safety factor of the target slope. The predicted safety factor can provide real-time or early warning for natural disasters such as slope landslides and mud-rock flows.

[0097] The technical solution of the embodiment of the present invention, by acquiring rainfall parameters and slope parameters of each layer of the slope, the slope includes a vegetation layer and several soil layers, and establishing a vegetation multi-layer soil slope model that takes into account the combined effects of vegetation, rainfall status and multiple soil layers, to analyze the slope stability, the present invention can more realistically reflect the actual state of the slope containing multiple soil layers of vegetation under rainfall conditions, and can obtain a more accurate slope safety factor. By introducing a machine learning model for predicting the stability of a rainfall vegetation multi-soil layer slope, a rainfall vegetation multi-soil layer slope stability prediction model with good generalization ability is obtained through training with a large number of soil slope data samples, and a stability analysis result that is closer to the actual situation is predicted. On the other hand, the present invention can also predict the safety of the slope in real time or in advance and issue an alarm based on the slope parameters of the transplant target slope and the real-time or predicted rainfall parameters.

[0098] In one embodiment, the pre-set vegetation multi-layer soil slope model includes:

[0099] An unsaturated seepage control model considering different rainfall parameters and initial groundwater levels, a moisture content model considering any point in each layer of a slope with vegetation and multiple soil layers, a seepage model, a shear strength calculation model for unsaturated soil in each layer of a slope with vegetation and multiple soil layers, and a safety factor calculation model for a slope with vegetation and multiple soil layers under different rainfall conditions.

[0100] Among them, for soil slopes with different rainfall and initial groundwater levels, the transient saturated zone in the soil slope can be moved downward under the action of seepage, which eventually leads to the change of groundwater level. The unsaturated seepage control model is:

[0101]

[0102] H 总 represents the total water head, which is related to rainfall intensity, rainfall duration, initial groundwater level and slope angle; k xi , k yi are the horizontal permeability and vertical permeability of the i-th layer, and the permeability k of the i-th soil layer. wi Correlation; θ w (h i ) is the moisture content of the i-th layer of the soil slope during the seepage process; φ i is the matrix suction of the i-th layer; x is the horizontal direction, y is the vertical direction; t is the rainfall time, that is, the unsaturated seepage control is related to the rainfall duration.

[0103] In this embodiment, the Figure 2 The vegetation multi-layer soil slope model shown in the figure includes a vegetation layer, 3 soil layers and a groundwater layer. Figure 2 H in the equation is the vertical distance from the upper edge of the slope to the horizontal plane of the lower edge of the slope. is the slope angle of the slope. Figure 2 The vegetation multi-layer soil slope model is only an exemplary description, and can actually include multiple soil layers, and the number of soil layers can be set to be greater than or equal to 1.

[0104] Then the moisture content model of any point in each layer of the slope including vegetation and multi-layer soil is:

[0105]

[0106] θ si ,θ ri is the saturated water content and residual water content in the i-th layer of soil; H is the slope height; H i represents the height of a point in the i-th soil layer; represents the VG model parameters related to the soil characteristics of the i-th layer; ni and mi represent the VG model parameters of the i-th layer; h i Indicates the height difference of each layer; Represents the matrix suction of the vegetation layer; i=1, vegetation layer, h0=0.

[0107] The widely recognized classic VG saturated-unsaturated seepage model is used to describe the relationship between soil slopes in the seepage process:

[0108]

[0109] In the formula, k wi represents the permeability coefficient of the i-th soil layer; represents the matrix suction of the i-th soil layer; k 0i is the saturated permeability coefficient of the i-th soil layer.

[0110] Through the above model, based on the constructed vegetation multi-layer soil slope model, the change of moisture content of each soil layer with rainfall under different rainfall and different soil slope conditions can be simulated. Figure 4 In one embodiment, the soil water saturation changes with rainfall time under uniform rainfall conditions. The rainfall intensity is set to 0.006m / h, the rainfall time starts from 0h and ends at 9h, and the rainfall type is uniform rainfall. The soil moisture content after rainfall is as follows Figure 5 (a) - Figure 5 As shown in (d), Figure 5 It consists of two layers: the top vegetation layer and the soil layer below the vegetation layer. Figure 5 (a) - Figure 5 The horizontal axis of (d) is length (unit: meter), and the vertical axis is height (unit: meter). Figure 5 The color from blue to dark red indicates the water saturation rate of the soil layer. When it reaches dark red, the water saturation rate is 1 or close to 1, and the corresponding water content reaches or is close to saturation; dark blue indicates that the water content is 0.2-0.3. Figure 5 (a) shows the water saturation of each layer of the slope at 0h; Figure 5 (b) shows the water saturation of each layer of the slope at 2h; Figure 5 (c) shows the water saturation of each layer of the slope at 6 hours; Figure 5 (d) in the figure shows the water saturation of each layer of the slope at 9h. Figure 5 (a) - Figure 5 As can be seen from (d) in the figure, as the rainfall progresses, rainwater will gradually penetrate from the vegetation layer to the soil layer below.

[0111] By using the modified Mohr-Coulomb failure criterion and combining the corresponding soil-water characteristic curve to obtain the corresponding matrix suction, the shear strength model of each layer of unsaturated soil considering vegetation and multi-layer slopes can be obtained:

[0112]

[0113] is the shear strength of unsaturated soil; c' is the effective cohesion; is the effective internal friction angle; represents the matrix suction of the i-th layer; is the net normal stress, which varies with the moisture content of the soil layer; is the suction parameter, which changes with the moisture content of the soil layer; the effective cohesion of each layer is related to the initial cohesion of each layer, the moisture content of any point in each layer and the soil quality; the effective internal friction of each layer is related to the initial internal friction of each layer, the moisture content of any point in each layer and the soil quality.

[0114] The safety factor calculation model of vegetated multi-soil slope under different rainfall conditions is:

[0115]

[0116] C is the cohesion of the slope after rainfall; is the internal friction angle of the slope after rainfall; , are the cohesion and internal friction angle at critical failure of the slope, respectively.

[0117] In one embodiment, the method for obtaining the cohesion and internal friction angle at the critical failure of the slope can be: using the finite element strength reduction method to obtain the cohesion at the critical failure of the slope and internal friction angle .

[0118] It should be noted that the finite element strength reduction coefficient method simulates the process of the slope from stability to progressive failure by gradually reducing the shear strength parameters of the soil, such as the internal friction angle and cohesion, so as to determine the safety factor of the slope. This method is based on finite element analysis. By establishing a finite element model of the slope, different properties are given to each layer of soil material on the slope. After the ground stress is balanced, the stress field of the entire rainfall slope is calculated. Subsequently, by gradually increasing the safety factor of the slope (i.e., the reduction factor F of the shear strength of the soil), the slope safety factor is gradually reduced. s ), assign the reduced strength parameters to the calculation model, and recalculate to simulate the response of the slope under the increasing external force. When the soil strength itself is not enough to resist the shear stress, soil instability will occur naturally, so that the stability and failure process of the slope can be truly reflected. This application is based on the constructed vegetation multi-layer soil slope model to simulate and establish a finite element model of the vegetation multi-layer soil slope model. The finite element model after modeling is as follows Figure 3 shown.

[0119] When applying the finite element strength reduction factor method, multiple parameters need to be considered, such as the friction angle, cohesion, elastic modulus, Poisson's ratio and density of each soil layer; these parameters have an important influence on the stability and failure process of the soil. By substituting the soil slope parameters of each layer and performing calculations, the safety and stability of the slope can be evaluated more accurately.

[0120] In one embodiment, the present invention further includes constructing a rainfall-vegetation multi-soil layer slope stability prediction model based on a random forest model.

[0121] In one embodiment, step S104 uses the data set to train a rainfall-vegetation multi-soil layer slope stability prediction model, including:

[0122] S1041, dividing the data set into a training set and a test set;

[0123] S1042, for each training sample, using the red-billed blue magpie optimization algorithm to optimize the parameters of the random forest model to obtain the optimal parameter configuration of the random forest model for each training sample;

[0124] S1043. Use a test set to test the random forest model to evaluate the performance of the random forest model.

[0125] In this embodiment, for each training sample, the parameters of the random forest model can be optimized using the red-billed blue magpie optimization algorithm, wherein the red-billed blue magpie optimization algorithm is an algorithm based on swarm intelligence, which simulates the hunting process of the red-billed blue magpie, including searching, attacking prey and storing food.

[0126] The optimization algorithm of the red-billed blue magpie specifically includes:

[0127] Define the search space of hyperparameters; the hyperparameters include the number of trees, the maximum depth of each tree, the minimum number of node split samples for each tree, the minimum number of leaf node samples for each tree, and the maximum number of features. A reasonable search range can be defined for each hyperparameter, for example, the number of decision trees may be [10, 1000]. Other hyperparameters, the maximum depth of each tree, the minimum number of node split samples for each tree, the minimum number of leaf node samples for each tree, and the maximum number of features, can be set as required.

[0128] Generate an initial population, each individual in the initial population represents a candidate solution for a set of hyperparameters;

[0129] Find the goal of each individual. The goal formula is:

[0130]

[0131] is the mean of the individual group targets; Q is the group size; is the mth individual in this group; m is a constant greater than or equal to 1 and less than or equal to Q; N is a random number; The target of the current individual calculated for this iteration; The target of the current individual calculated in the previous iteration;

[0132] Get the optimal solution found in the population of this iteration. The formula for the optimal solution is:

[0133]

[0134] is the optimal solution currently found; M is the adjustment coefficient; is a normally distributed random number;

[0135]

[0136] The adjustment factor M is:

[0137]

[0138] t is the current iteration number; T is the maximum iteration number;

[0139] The fitness of individuals in each iteration is evaluated by minimizing the loss function. The fitness evaluation formula is:

[0140]

[0141] In the formula, is the safety factor of each sample slope; The machine learning modeling predicted value for each sample; n p Indicates the number of samples;

[0142] When the number of iterations reaches the maximum number of iterations or the individual fitness is less than or equal to the fitness threshold, the iteration is stopped and the optimal hyperparameters are output. The optimal hyperparameters define the optimal parameter configuration of the random forest model.

[0143] In a specific embodiment, the blue magpie optimized random forest model of the present invention and the existing random search optimized random forest model and grid search optimized random forest model are tested respectively, and the test results are compared. Figure 5 The slope model shown includes vegetation and a soil layer. The slope height of the slope model is 5m, the slope angle is 34°, and the initial groundwater level is 2 meters below the ground. The parameters of each layer of the slope model are shown in Table 1 below:

[0144] Table 1 Parameters of each layer of slope model

[0145]

[0146] In addition, multiple groups of rainfall parameters can be set to test the performance of the solution of the present invention and the existing solution under various rainfall parameters. The set rainfall parameters are shown in Table 2 below:

[0147] Table 2 Rainfall parameter group settings

[0148]

[0149] During the rainfall process, the safety factor was recorded every ten minutes, and a total of 465 data sets were collected.

[0150] 75% of the data set was divided into a training set and 25% into a test set. The same computer was used to test the random forest model optimized by random search, the random forest model optimized by grid search, and the random forest model optimized by red-billed blue magpie. The test results are shown in Table 3 below:

[0151] Table 3 Test data of three schemes

[0152]

[0153] Table 3 includes high determination coefficient R² and low error indicators RMSE and MAE (root mean square error RMSE, mean absolute error MAE). It can be seen from Table 3 that the root mean square error and mean absolute error of the test results of the red-billed blue magpie optimization random forest model of the scheme of the present invention are smaller, and the high determination coefficient is closer to 1, that is, the present invention shows better stability and reliability than the two existing schemes, and has a more prominent effect on rainfall warning. In addition, the total time required for the scheme of the present invention to process a batch of data is relatively short, so it can quickly respond to changes in slope stability and improve the efficiency of emergency services and disaster management.

[0154] In addition, you can refer to the comparison chart of the predicted value (orange) and the true value (blue) of the red-billed blue magpie optimized random forest model in the present invention, as shown in the figure. Figure 6 As shown, Figure 6 The horizontal axis is the data point of the test set, the vertical axis is the safety factor, the orange curve represents the predicted value, and the blue curve represents the actual value. It can be seen that the predicted value of the safety factor of the scheme of the present invention can be close to the actual value, and its prediction result has outstanding stability and reliability.

[0155] In order to better simulate the real environment, the present invention takes into account more influencing factors, optimizes resource allocation, and does not need to be modeled in detail, which can reduce computing resources and thus reduce costs. In addition, after the present invention substitutes rainfall parameters and slope parameters into the optimization model, it can use real-time or predicted rainfall parameters for prediction, and no longer needs to perform a large amount of historical data calculation, and can quickly obtain prediction results, reducing the additional costs that may be incurred by waiting for the calculation results.

[0156] Figure 7 This is a schematic diagram of the structure of a device for predicting the stability of a rainfall-vegetated multi-soil-layer slope provided by Embodiment 3 of the present invention. Figure 7As shown, the device comprises:

[0157] A parameter acquisition unit 701 is used to acquire rainfall parameters and slope parameters of each layer of the slope, wherein the slope includes a vegetation layer and several soil layers;

[0158] A calculation unit 702, used for substituting the rainfall parameter and the slope parameter of each layer of the slope into the preset vegetation multi-layer soil slope model to calculate the safety factor of the slope;

[0159] The data construction unit 703 is used to take the rainfall parameter, the slope parameter of each layer of the slope and the safety factor of the slope as a group of samples to obtain a data set including multiple groups of samples; the data set includes multiple groups of samples with different rainfall parameters and different slope parameters;

[0160] A training unit 704 is used to train a rainfall-vegetation-multi-soil-layer slope stability prediction model using the data set to obtain a trained rainfall-vegetation-multi-soil-layer slope stability prediction model;

[0161] The prediction unit 705 is used to obtain the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters, and input the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters into the trained rainfall-vegetation multi-layer slope stability prediction model to predict the safety factor of the target slope.

[0162] In one embodiment, the training unit 704 is further configured to divide the data set into a training set and a test set;

[0163] For each training sample, the parameters of the random forest model are optimized using a red-billed blue magpie optimization algorithm to obtain the optimal parameter configuration of the random forest model for each training sample;

[0164] The random forest model is tested using a test set to evaluate the performance of the random forest model.

[0165] The message data processing resource scheduling device provided in the embodiment of the present invention can execute the message data processing resource scheduling device method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0166] The technical solution of the embodiment of the present invention, by acquiring rainfall parameters and slope parameters of each layer of the slope, the slope includes a vegetation layer and several soil layers, and establishing a vegetation multi-layer soil slope model that takes into account the combined effects of vegetation, rainfall status and multiple soil layers, to analyze the slope stability, the present invention can more realistically reflect the actual state of the slope containing multiple soil layers of vegetation under rainfall conditions, and can obtain a more accurate slope safety factor. By introducing a machine learning model for predicting the stability of a rainfall vegetation multi-soil layer slope, a rainfall vegetation multi-soil layer slope stability prediction model with good generalization ability is obtained through training with a large number of soil slope data samples, and a stability analysis result that is closer to the actual situation is predicted. On the other hand, the present invention can also predict the safety of the slope in real time or in advance and issue an alarm based on the slope parameters of the transplant target slope and the real-time or predicted rainfall parameters.

[0167] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0168] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the stability of a rainfall-vegetated multi-soil slope, characterized in that: include: Obtaining rainfall parameters and slope parameters of each layer of the slope, wherein the slope includes a vegetation layer and several soil layers; Substituting rainfall parameters and slope parameters of each layer of the slope into a preset vegetation multi-layer soil slope model to calculate the safety factor of the slope; The pre-set vegetation multi-layer soil slope model includes: Unsaturated seepage control model considering different rainfall parameters and initial groundwater levels, moisture content model considering any point in each layer of a slope with vegetation and multiple soil layers, seepage model, shear strength calculation model of unsaturated soil in each layer of a slope with vegetation and multiple soil layers, and safety factor calculation model of a slope with vegetation and multiple soil layers under different rainfall conditions; The moisture content model for any point in each layer of a multi-layer slope including vegetation is: θ si ,θ ri is the saturated moisture content and residual moisture content in the i-th layer of soil; H is the slope height; Hi represents the height of a point in the i-th layer of soil; represents the VG model parameters related to the soil characteristics of the i-th layer; ni and mi represent the VG model parameters of the i-th layer; h i Indicates the height difference of each layer; Indicates the matrix suction of the vegetation layer; i=1, vegetation layer, h0=0; The classic VG saturated-unsaturated seepage model is: In the formula, k wi represents the permeability coefficient of the i-th soil layer; represents the matrix suction of the i-th soil layer; k 0i is the saturated permeability coefficient of the i-th soil layer; The unsaturated seepage control model of soil slope under seepage action under different rainfall parameters and initial groundwater levels is: H 总 represents the total water head, which is related to rainfall intensity, rainfall duration, initial groundwater level and slope angle; k xi , k yi are the horizontal permeability and vertical permeability of the i-th layer, and the permeability k of the i-th soil layer. wi Correlation; θ w (h i ) is the moisture content of the i-th layer of the soil slope during the seepage process; φ i is the matrix suction of the i-th layer; x is the horizontal direction, y is the vertical direction; t is the rainfall time, that is, the unsaturated seepage control is related to the rainfall duration; The shear strength model of each layer of unsaturated soil including vegetation and multi-layer slope is: is the shear strength of unsaturated soil; c' is the effective cohesion; is the effective internal friction angle; represents the matrix suction of the i-th layer; is the net normal stress, which varies with the moisture content of the soil layer; is the suction parameter, which changes with the moisture content of the soil layer; the effective cohesion of each layer is related to the initial cohesion of each layer, the moisture content of any point in each layer, and the soil quality; the effective internal friction of each layer is related to the initial internal friction of each layer, the moisture content of any point in each layer, and the soil quality; The safety factor calculation model of vegetated multi-soil slope under different rainfall conditions is: C is the cohesion of the slope after rainfall; is the internal friction angle of the slope after rainfall; , are the cohesion and internal friction angle at critical failure of the slope respectively; The rainfall parameter, the slope parameter of each layer of the slope and the safety factor of the slope are taken as a group of samples to obtain a data set containing multiple groups of samples; the data set contains multiple groups of samples with different rainfall parameters and different slope parameters; The data set is used to train a rainfall-vegetated multi-soil-layer slope stability prediction model to obtain a trained rainfall-vegetated multi-soil-layer slope stability prediction model; Obtain the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters, and input the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters into the trained rainfall-vegetation multi-soil layer slope stability prediction model to predict the safety factor of the target slope.

2. The method for predicting the stability of multi-layer slopes with rainfall and vegetation according to claim 1 is characterized in that: The slope parameters of each layer of the slope at least include slope height, slope angle, initial groundwater level, saturated permeability coefficient of each layer, residual moisture content of each layer of soil, saturated moisture content of each layer, initial cohesion of each layer, initial internal friction angle of each layer, and VG model parameters; The rainfall parameters include at least rainfall intensity, rainfall duration, and rainfall type.

3. The method for predicting stability of multi-layer slopes with rainfall and vegetation according to claim 1 is characterized in that: Also includes: Using finite element strength reduction method to obtain the cohesion of slope at critical failure and internal friction angle .

4. The method for predicting stability of multi-layer slopes with rainfall and vegetation according to claim 1 is characterized in that: Also includes: A rainfall-vegetation multi-layer slope stability prediction model based on the random forest model was constructed.

5. The method for predicting stability of multi-layer slopes with rainfall and vegetation according to claim 4 is characterized in that: The method of using the data set to train a rainfall-vegetation multi-soil layer slope stability prediction model comprises: Divide the dataset into training and testing sets; For each training sample, the parameters of the random forest model are optimized using a red-billed blue magpie optimization algorithm to obtain the optimal parameter configuration of the random forest model for each training sample; The random forest model is tested using a test set to evaluate the performance of the random forest model.

6. The method for predicting stability of multi-layer slopes with rainfall and vegetation according to claim 5 is characterized in that: The method of optimizing the parameters of the random forest model using the red-billed blue magpie optimization algorithm includes: Defining a search space for hyperparameters; the hyperparameters include the number of trees, the maximum depth of each tree, the minimum number of node split samples for each tree, the minimum number of leaf node samples for each tree, and the maximum number of features; Generate an initial population, each individual in the initial population represents a candidate solution for a set of hyperparameters; Find the goal of each individual. The goal formula is: is the mean of the individual group targets; Q is the group size; is the mth individual in this group; m is a constant greater than or equal to 1 and less than or equal to Q; N is a random number; The target of the current individual calculated for this iteration; The target of the current individual calculated in the previous iteration; Get the optimal solution found in the population of this iteration. The formula for the optimal solution is: is the optimal solution currently found; M is the adjustment coefficient; is a normally distributed random number; The adjustment factor M is: t is the current iteration number; T is the maximum iteration number; The fitness of individuals in each iteration is evaluated by minimizing the loss function. The fitness evaluation formula is: In the formula, is the safety factor of each sample slope; The machine learning modeling predicted value for each sample; n p Indicates the number of samples; When the number of iterations reaches the maximum number of iterations or the individual fitness is less than or equal to the fitness threshold, the iteration is stopped and the optimal hyperparameters are output. The optimal hyperparameters define the optimal parameter configuration of the random forest model.

7. A device for predicting the stability of a multi-layer slope with rainfall and vegetation, characterized in that: include: A parameter acquisition unit, used to acquire rainfall parameters and slope parameters of each layer of the slope, wherein the slope includes a vegetation layer and several soil layers; A calculation unit, used for substituting the rainfall parameter and the slope parameter of each layer of the slope into the preset vegetation multi-layer soil slope model to calculate the safety factor of the slope; The pre-set vegetation multi-layer soil slope model includes: Unsaturated seepage control model considering different rainfall parameters and initial groundwater levels, moisture content model considering any point in each layer of a slope with vegetation and multiple soil layers, seepage model, shear strength calculation model of unsaturated soil in each layer of a slope with vegetation and multiple soil layers, and safety factor calculation model of a slope with vegetation and multiple soil layers under different rainfall conditions; The moisture content model for any point in each layer of a multi-layer slope including vegetation is: θ si ,θ ri is the saturated moisture content and residual moisture content in the i-th layer of soil; H is the slope height; Hi represents the height of a point in the i-th layer of soil; represents the VG model parameters related to the soil characteristics of the i-th layer; ni and mi represent the VG model parameters of the i-th layer; h i Indicates the height difference of each layer; Indicates the matrix suction of the vegetation layer; i=1, vegetation layer, h0=0; The classic VG saturated-unsaturated seepage model is: In the formula, k wi represents the permeability coefficient of the i-th soil layer; represents the matrix suction of the i-th soil layer; k 0i is the saturated permeability coefficient of the i-th soil layer; The unsaturated seepage control model of soil slope under seepage action under different rainfall parameters and initial groundwater levels is: H 总 represents the total water head, which is related to rainfall intensity, rainfall duration, initial groundwater level and slope angle; k xi , k yi are the horizontal permeability and vertical permeability of the i-th layer, and the permeability k of the i-th soil layer. wi Correlation; θ w (h i ) is the moisture content of the i-th layer of the soil slope during the seepage process; φ i is the matrix suction of the i-th layer; x is the horizontal direction, y is the vertical direction; t is the rainfall time, that is, the unsaturated seepage control is related to the rainfall duration; The shear strength model of each layer of unsaturated soil including vegetation and multi-layer slope is: is the shear strength of unsaturated soil; c' is the effective cohesion; is the effective internal friction angle; represents the matrix suction of the i-th layer; is the net normal stress, which varies with the moisture content of the soil layer; is the suction parameter, which changes with the moisture content of the soil layer; the effective cohesion of each layer is related to the initial cohesion of each layer, the moisture content of any point in each layer, and the soil quality; the effective internal friction of each layer is related to the initial internal friction of each layer, the moisture content of any point in each layer, and the soil quality; The safety factor calculation model of vegetated multi-soil slope under different rainfall conditions is: C is the cohesion of the slope after rainfall; is the internal friction angle of the slope after rainfall; , are the cohesion and internal friction angle at critical failure of the slope respectively; A data construction unit is used to take the rainfall parameter, the slope parameter of each layer of the slope and the safety factor of the slope as a group of samples to obtain a data set containing multiple groups of samples; the data set contains multiple groups of samples with different rainfall parameters and different slope parameters; A training unit, used for training a rainfall-vegetated multi-soil-layer slope stability prediction model using the data set to obtain a trained rainfall-vegetated multi-soil-layer slope stability prediction model; A prediction unit is used to obtain the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters, and input the slope parameters of each layer of the target slope and the real-time or predicted rainfall parameters into the trained rainfall-vegetation multi-soil layer slope stability prediction model to predict the safety factor of the target slope.

8. The device for predicting stability of multi-layer slopes with rainfall and vegetation according to claim 7 is characterized in that: The training unit is also used to divide the data set into a training set and a test set; For each training sample, the parameters of the random forest model are optimized using the red-billed blue magpie optimization algorithm to obtain the optimal parameter configuration of the random forest model for each training sample; The random forest model is tested using a test set to evaluate the performance of the random forest model.

Images