A method and system for predicting the most dangerous sliding surface of geogrid slopes

Abstract

The present invention provides a method and system for predicting the most dangerous sliding surface of a geogrid slope, which relates to the technical field of the most dangerous sliding surface prediction, including obtaining the design data and historical monitoring data of the geogrid slope, and building a physical calculation model of the most dangerous sliding surface based on the design data. The design data is processed using a Markov random field, and the processed data samples are input into the physical calculation model of the most dangerous sliding surface to obtain a data set of the most dangerous sliding surface. The data set and the design data are divided into a training set and a test set, and are input into a preset neural network for training and verification. The neural network is optimized by an improved coronavirus group immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model. The historical monitoring data is input into the optimized model to predict the most dangerous sliding surface parameters of the geogrid slope. The present invention can accurately predict the most dangerous sliding surface of the geogrid slope and improve the accuracy of the slope stability analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of most dangerous sliding surface prediction, and in particular to a method and system for predicting the most dangerous sliding surface of a geogrid slope. Background Art

[0002] Currently, the main methods for determining the most dangerous sliding surface on geogrid slopes and evaluating slope stability include theoretical methods, numerical methods, and field monitoring methods. These methods are significantly influenced by human factors. The number of parameters and the coupling of multiple parameters increase the labor time cost and monitoring difficulty, increasing the probability of structural failure. Therefore, to prevent disasters, a large amount of manual operation is required for post-monitoring, resulting in a significant waste of time and financial costs, and invisibly increasing construction costs.

[0003] Furthermore, all of these methods have limitations. For example, theoretical methods require a priori assumptions about the sliding surface pattern, while numerical simulation methods require not only the physical and mechanical parameters of the rock mass but also deformation parameters, which are rarely considered in engineering surveys. This makes it difficult for either method to accurately identify the most dangerous potential sliding surface on a slope, and thus cannot meet practical engineering needs. Furthermore, both methods are static analyses and cannot reflect the dynamic trends of the slope's stability under changing internal and external factors.

[0004] Therefore, a method and system for predicting the most dangerous sliding surface of a geogrid slope is needed, which can improve optimization performance and operating speed, and is used to improve the accuracy of the most dangerous sliding surface prediction of a geogrid slope and reduce costs. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for predicting the most dangerous sliding surface of a geogrid slope to improve the above-mentioned problem. To achieve the above-mentioned purpose, the technical solution adopted by the present invention is as follows:

[0006] In a first aspect, the present application provides a method for predicting the most dangerous sliding surface of a geogrid slope, comprising:

[0007] Obtain design data of geogrid slopes and historical monitoring data of geogrid slopes;

[0008] Build a physical calculation model of the most dangerous sliding surface of the geogrid slope based on the design data of the geogrid slope;

[0009] The design data of the geogrid slope is processed based on the Markov random field, and all the processed data samples are sent to the physical calculation model of the most dangerous sliding surface for calculation to obtain the data set of the most dangerous sliding surface;

[0010] The dataset of the most dangerous sliding surface and the design data of the geogrid slope are divided into a training set and a test set;

[0011] The training set and the test set are sent to a preset neural network for training and verification, and the preset neural network is optimized by a preset improved coronavirus herd immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model of the geogrid slope;

[0012] The historical monitoring data of the geogrid slope is sent to the optimized most dangerous sliding surface prediction model of the geogrid slope to predict the most dangerous sliding surface parameters of the geogrid slope.

[0013] In a second aspect, the present application also provides a most dangerous sliding surface prediction system for geogrid slopes, comprising:

[0014] An acquisition unit, used for acquiring design data of the geogrid slope and historical monitoring data of the geogrid slope;

[0015] A calculation unit, used to build a physical calculation model of the most dangerous sliding surface of the geogrid slope based on the design data of the geogrid slope;

[0016] A processing unit is used to process the design data of the geogrid slope based on the Markov random field, and send all the processed data samples to the physical calculation model of the most dangerous sliding surface for calculation to obtain the data set of the most dangerous sliding surface;

[0017] A classification unit is used to divide the dataset of the most dangerous sliding surface and the design data of the geogrid slope into a training set and a test set;

[0018] An optimization unit is used to send the training set and the test set to a preset neural network for training and verification, and optimize the preset neural network using a preset improved coronavirus herd immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope;

[0019] The prediction unit is used to send the historical monitoring data of the geogrid slope to the optimized most dangerous sliding surface prediction model of the geogrid slope to predict the most dangerous sliding surface parameters of the geogrid slope.

[0020] The beneficial effects of the present invention are:

[0021] The improved geogrid slope most dangerous slip surface prediction system of the present invention boasts a high computational speed, reducing time costs and performance losses. The system can consider multiple parameters and possesses strong adaptability, fault tolerance, and self-improvement capabilities, improving monitoring and calculation accuracy. Furthermore, the system can automatically search for solutions that meet any user requirements, reducing monitoring and calculation labor costs.

[0022] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the embodiments of the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 Schematic diagram of the process of the most dangerous sliding surface prediction method of the geogrid slope according to an embodiment of the present invention;

[0025] Figure 2 Schematic diagram of the most dangerous sliding surface prediction system for geogrid slopes according to an embodiment of the present invention.

[0026] In the figure: 701, acquisition unit; 702, calculation unit; 703, processing unit; 704, classification unit; 705, optimization unit; 706, prediction unit. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions 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 part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0028] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are used only to distinguish the description and should not be understood as indicating or implying relative importance.

[0029] Example 1:

[0030] This embodiment provides a method for predicting the most dangerous sliding surface of a geogrid slope.

[0031] See also Figure 1 , the figure shows that the method includes step S1, step S2, step S3, step S4, step S5 and step S6.

[0032] Step S1, obtaining design data of the geogrid slope and historical monitoring data of the geogrid slope;

[0033] It is understandable that the design data of the geogrid slope is obtained, namely the slope height H, the slope soil weight γ, the internal friction angle Cohesion c, slope angle β, and geogrid pullout coefficient μ; historical monitoring data on the geogrid slope includes sliding surface parameters at each pre-set point, the real-time stress state at each measuring point, and the points on the potentially most dangerous sliding surface, along with their inclination and stress state. This step acquires detailed and accurate design data and historical monitoring data, providing the necessary foundation for subsequent preprocessing, model optimization, and real-time monitoring. The comprehensiveness and authenticity of this data directly impacts the accuracy and practicality of the sliding surface prediction model, helping the engineering team develop effective slope management and maintenance strategies.

[0034] Step S2: constructing a physical calculation model of the most dangerous sliding surface of the geogrid slope based on the design data of the geogrid slope;

[0035] It can be understood that the horizontal strip is divided into n segments using the horizontal strip division idea. The sliding surface corresponding to any segment i is a straight line and is α i Therefore, the most dangerous sliding surface of the slope is composed of n straight lines. If the number of segments is large enough, a smooth and continuous most dangerous sliding surface of the slope can be obtained. Step S2 includes steps S21, S22, S23 and S24.

[0036] Step S21: Perform force analysis on any horizontal slider on the slope to calculate the normal force and tangential force on the sliding surface of the horizontal slider;

[0037] Take any horizontal slider i on the slope for force analysis and calculate the normal force F on the sliding surface of the horizontal slider i ni and tangential force Fti ;

[0038] Specifically, the normal force F on the sliding surface of the horizontal slider i is ni The calculation is as follows;

[0039]

[0040] F vi =γh(ni)t i

[0041] h=H / n

[0042] l i =h / sinα i

[0043] Where: G i Represents the gravity of the horizontal slider; F vi and F vi-1 Respectively expressed as the vertical inter-bar force on the upper and lower sides; k is the strength reduction coefficient; α i is the angle between the straight line of the sliding surface corresponding to any segment i and the horizontal plane; γ is the weight of the slope rock and soil; is the internal friction angle; c is the cohesion; h is the thickness of the slider; l i is the sliding surface length of slider i; n is the number of horizontal strip segments of the slope soil; t i is the length of the upper surface of slider i.

[0044] Step S22: Calculate the minimum tension of the slope geogrid based on the normal force and tangential force on the sliding surface of the horizontal slider;

[0045] Specifically, the minimum tension of the slope geogrid is calculated as follows:

[0046]

[0047] Among them, when α1,α2,...,α n The tension has a minimum value when the following differential conditions are met.

[0048]

[0049] Where: F P is the minimum tension of the slope geogrid; F Pi G is the minimum tension of the geogrid on the sliding surface of the horizontal slider i of the slope; i Represents the gravity of the horizontal slider; F ni is the normal force on the sliding surface of the horizontal slider i; is the internal friction angle; c is the cohesion; h is the thickness of the slider; l i is the sliding surface length of slider i; n is the number of horizontal strip segments of the slope soil; k is the strength reduction coefficient; αi is the angle between the straight line of the slip surface corresponding to any segment i and the horizontal plane;

[0050] Step S23, calculating the horizontal length of the sliding surface point at the position of the horizontal slider from the slope surface according to the minimum tension of the slope geogrid;

[0051] Specifically, the horizontal distance L between the sliding point at the position of the horizontal slider i and the slope surface is Ci The calculation is as follows;

[0052] L Ci =L+L ai

[0053]

[0054] Where: L Ci The horizontal length L from the sliding surface point at the position of the horizontal slider i to the slope surface is the length of the reinforcement material penetrating into the stable rock (soil); L ai is the length of the anchorage area; σ vi is the soil pressure acting on the horizontal slider i; μ is the pull-out coefficient of the geogrid; F P is the minimum tension of the slope geogrid; n is the number of horizontal strip segments of the slope soil; β is the slope angle; and h is the thickness of the slider.

[0055] Step S24: determining the position of the sliding surface point based on the horizontal length between the sliding surface point at the position of the horizontal slider and the slope surface, and connecting the sliding surface points at the positions of the sliders to obtain the most dangerous sliding surface.

[0056] It can be understood that after the position of the sliding surface point is determined in this step, the sliding surface points at the positions of the sliders can be connected to obtain the most dangerous sliding surface.

[0057] Step S3: Processing the design data of the geogrid slope based on the Markov random field, and sending all the processed data samples to the physical calculation model of the most dangerous sliding surface for calculation to obtain the data set of the most dangerous sliding surface;

[0058] It can be understood that multiple data samples are generated using random field images, and the most dangerous sliding surface physical model of the slope is introduced to calculate the most dangerous sliding surface corresponding to each data sample. The data set is composed of all data samples. In this step, step S3 includes step S31, step S32 and step S33.

[0059] Step S31: exporting the design data of the geogrid slope into 500 random field RGB images through a preset software, deleting redundant blank areas of the random field RGB images, and converting the random field RGB images into grayscale images to obtain at least two random field images;

[0060] It can be understood that the design data of the geogrid slope is obtained through FLAC 3D Export 500 random field RGB images; delete the extra blank areas of the random field images, convert the RGB images to grayscale images, convert the ui nt8 type (0-255) data to double precision type (0-1), and adjust the image size to fit the input layer size.

[0061] Step S32: randomly combining the design data parameters of the geogrid slope corresponding to the random field images to obtain a plurality of data samples;

[0062] Step S33: Send all data samples to the physical calculation model of the most dangerous slip surface for calculation to obtain the most dangerous slip surface corresponding to each data sample.

[0063] It can be understood that the random field images corresponding to the design data parameters of multiple geogrid slopes are randomly combined to generate multiple data samples, wherein one data sample contains a random field image corresponding to each design data parameter; and the most dangerous sliding surface physical model of the slope is used to determine the most dangerous sliding surface corresponding to each data sample.

[0064] Step S4, dividing the data set of the most dangerous sliding surface and the design data of the geogrid slope into a training set and a test set;

[0065] It is understandable that dividing the dataset into training set and test set in a ratio of 7:3 can better achieve the training effect and achieve the purpose of balanced training and testing.

[0066] Step S5: sending the training set and the test set to a preset neural network for training and verification, and optimizing the preset neural network using a preset improved coronavirus herd immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope;

[0067] It is understandable that this step optimizes the neural network by improving the coronavirus herd immunity optimization algorithm. It is an improvement based on the coronavirus herd immunity algorithm. First, in the initialization population stage, the Picewise chaos mapping method is introduced to make the individual genes more evenly distributed within the constraints of the basic reproduction rate and the age of the maximum infected case. Then, differential mutation, crossover, and selection strategies are introduced to optimize the individual genes. The optimal individual genes found within the constraints are the hidden layer parameters and learning rate parameters of the corresponding neural network. The hidden layer parameters and learning rate parameters of the optimal neural network are output as the most accurate position of the most dangerous sliding surface of the slope. A more accurate prediction of the most dangerous sliding surface of the slope is achieved. The optimized model can better utilize complex and diverse slope data information, improve the robustness and accuracy of the prediction model, and provide important technical support for engineering decision-making. Step S5 includes steps S51, S52, S53, and S54.

[0068] Step S51: sending the training set to a preset neural network for training to obtain a trained neural network;

[0069] It can be understood that this step optimizes the neural network parameters by training the training set, giving it high predictive and generalization capabilities. The training results reflect the model's performance and effectiveness in predicting the most dangerous slip surface on the slope, providing a foundation for subsequent testing and verification.

[0070] Step S52: optimizing the hidden layer parameters and learning rate parameters of the trained neural network based on the improved coronavirus herd immunity optimization algorithm to obtain optimized hidden layer parameters and learning rate parameters;

[0071] It can be understood that this step optimizes the hidden layer parameters and learning rate of the neural network by improving the coronavirus herd immunity optimization algorithm, which significantly improves the prediction accuracy and generalization ability of the model. The optimized model can more accurately predict the most dangerous sliding surface conditions of geogrid slopes, providing more reliable support for practical engineering applications. In this step, step S52 includes steps S521, S522, S523, and S524.

[0072] Step S521: Initialize the hidden layer parameters and learning rate parameters in the preset neural network, and set the parameters of the preset improved coronavirus herd immunity optimization algorithm, including the algorithm population size, the initial number of infected cases, the problem dimension, the maximum number of iterations, the basic reproduction rate, and the maximum age of infected cases;

[0073] Step S522: randomly generate individual genes with a population size of a preset number, evenly distribute the individual genes based on the Picewise chaotic mapping method, and change the initial positions of the individual genes to obtain optimized individual gene positions;

[0074] Specifically, the Picewise chaotic map is described as follows:

[0075]

[0076] Where, parameter p = 0.4, x(t) is the initial position of the individual gene, and X(t+1) represents the position of the individual gene after update;

[0077] The changed population distribution expression is described as:

[0078]

[0079] Where, X k represents the population distribution after the improvement of Picewise chaotic mapping, Indicates the improved individual gene position.

[0080] Step S523: bringing the optimized individual gene positions into the improved coronavirus group immunity optimization algorithm, and performing differential mutation, crossover, and selection on the hidden layer parameters and learning rate parameters of the neural network based on the improved coronavirus group immunity optimization algorithm with the optimized individual gene positions, to obtain the screened excellent individual genes;

[0081] It's understandable that during the execution of this improved coronavirus herd immunity optimization algorithm, the position and fitness of the current individual gene are calculated during each iteration. Although the algorithm itself updates the individual genes, it is inevitable that there will be poor individuals. If we can optimize these individual genes after iteration, the search for optimal values ​​can theoretically be accelerated. Inspired by genetic algorithms, we propose using a differential mutation crossover selection strategy to identify superior individuals before the next individual position and fitness calculation. First, individual genes are mutated using a differential strategy, followed by adaptive crossover, and finally, the greedy principle is used to select the superior individual genes.

[0082] By improving the coronavirus herd immunity optimization algorithm, we optimized the neural network parameters, enabling them to achieve better fits on the training data. By gradually optimizing individual genes, the optimization process becomes more focused and efficient, resulting in optimal individual gene information. The neural network parameters corresponding to this optimal position information significantly improve the model's predictive and generalization capabilities, providing reliable optimization parameters for subsequent steps.

[0083] Step S524: Calculate the fitness value of the selected excellent individual genes, and update the individual genes based on the calculated fitness value until the number of iterations reaches the maximum number of iterations, thereby obtaining the optimal hidden layer parameters and learning rate parameters.

[0084] It can be understood that this step includes step S5241, step S5242, step S5243 and step S5244.

[0085] Step S5241: Use differential mutation, crossover, and selection strategies to screen individual genes to obtain excellent individual genes after screening.

[0086] The three main stages are described as follows:

[0087] Step 1: Mutation

[0088] According to each mutation vector X pi (t), i=1,2,3,…,N, generate mutation difference vector

[0089]

[0090] Where, X P1 ,X P2 ,X P3 are three random positions in individual genes; iter max Indicates the maximum number of iterations; t indicates the current number of iterations; exp represents the natural exponential function, that is, the exponential function with the real number e (e≈2.71828) as the base; τ is an intermediate variable with no practical meaning;

[0091] The random indexes P1, P2, and P3 are randomly selected integers from [1, NP] and P1≠P2≠P3≠i;

[0092] F is a real number scaling factor, and its value is between [0,2]. It controls the P2 (t)-X P3 (t) The amplitude of the differential variation; F0 represents the variation rate;

[0093] Step 2: Cross

[0094] After the previous stage is completed, the perturbation parameter vector is cross-operated to make the perturbation parameters diverse, thereby generating a trial vector. The specific description is:

[0095]

[0096] Where CR is the random crossover parameter, and its value range is [0.8-1];

[0097] rand is a random number between [0,1]; Ui (t+1) represents the test vector; X i (t) represents the target vector; X i+1 (t+1) represents the target vector for the next iteration;

[0098] Step 3: Select

[0099] After the initial individual gene undergoes the mutation and crossover phases, the test vector U is transformed into i (t+1) and the target vector X i (t) to determine whether it can become the next generation individual gene iteration individual. i The fitness value of (t+1) is less than the target vector X i (t), then X i+1 (t+1) is set to U i (t+1): Otherwise, keep the old value X i (t).

[0100] Its expression is described as:

[0101]

[0102] Where U i (t+1) represents the experimental vector; X i (t) represents the target vector; X i+1 (t+1) represents the target vector for the next iteration;

[0103] Step S5242: Calculate the fitness values ​​of the selected excellent individual genes based on the preset fitness function, sort the fitness values ​​of the calculated individual genes, and select the individual gene with the optimal value f g and the worst value f w ;

[0104] Among them, the preset fitness function is as follows:

[0105]

[0106] in, represents the gene at position i of individual j in the population at the tth iteration; y i Represents the output of the neural network model, Q i The position information of the individual gene is determined by the fitness of the individual gene.

[0107] Step S5243: The preset function calculates the position information of the individual gene, optimizes it using a dynamic step factor, obtains the position information of the individual gene after iteration, and updates the position of the individual gene;

[0108] It can be understood that this step uses dynamic step factors β and K. In the early stages of the algorithm, a small β is conducive to local optimization, and an increasing K is conducive to optimizing individual genes within the constraints. In the later stages of the algorithm, a large β is conducive to global optimization and enhances the ability to escape local optimality. A sharp decrease in K in the later stages is beneficial to increasing the convergence speed of the algorithm. The position information of individual genes is calculated through a function to obtain the position of individual genes after iteration. The formula for updating the position of individual genes is as follows:

[0109]

[0110] Where, represents the gene at position i of the infected individual j in the population at the tth iteration; Indicates the gene at position i of susceptible individual j in the population at the tth iteration, Indicates the gene at position i of immune individual j in the population at the tth iteration, BR r represents the basic reproduction rate, r represents a random number between 0 and 1, is the new individual gene value, is the individual gene value before updating. β and K represent the dynamic step size factor

[0111] The update formulas of β and K dynamic step factors are described as follows:

[0112]

[0113] Where, f g represents the optimal value of individual genes and f w Indicates the worst value of individual genes; iter max represents the maximum number of iterations; t represents the current number of iterations; exp represents the natural exponential function, that is, the exponential function with the real number e (e≈2.71828) as the base; rand is a random number between [0,1];

[0114] Step S5244: Calculate the immunity rate of each generated individual gene, and update the individual gene at a certain number of iterations specified by the parameter maximum infection case age until the optimal individual gene is reached.

[0115] The calculation formula for the immunity rate of individual genes is as follows:

[0116]

[0117] Among them, αk is the coefficient; k(x,x k ) is the kernel function; b is a constant; f(x) represents the immunity rate of individual genes;

[0118] Step S53: assigning the optimized hidden layer parameters and learning rate parameters to the trained neural network to obtain an assigned neural network;

[0119] It can be understood that this step updates the neural network by using the parameters optimized by the improved coronavirus herd immunity optimization algorithm, so that the model parameters can better adapt to the characteristics of the data and the prediction task. The optimized hidden layer parameters and learning rate parameters can improve the prediction accuracy and generalization ability of the model, thereby more accurately predicting the most dangerous sliding surface of the geogrid slope.

[0120] Step S54: sending the training set and the test set to the assigned neural network for training and verification, to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope.

[0121] This step optimizes the neural network model by assigning the optimized parameters to the neural network, enabling more accurate training and prediction. The optimized neural network will achieve higher accuracy and generalization capabilities when predicting the most dangerous sliding surface of geogrid slopes, providing more reliable support for practical engineering applications. In this step, step S54 includes steps S541, S542, and S543.

[0122] Step S541: Send the hidden layer parameters and learning rate parameters to the neural network to construct a forward calculation formula, and obtain the forward calculation formula of the neural network after assignment;

[0123] It can be understood that the forward calculation formula of the neural network in this step is as follows:

[0124]

[0125] ht=ot·tanh(ct)

[0126] Among them, inpft represents the forget gate of the neural network, Both represent the recursive weight of the neural network, σ represents the sigmoid function, and h t-1 represents the short-term memory of the neural network at time t-1, h t represents the short-term memory of the neural network at time t, x t Indicates the input operation parameter sequence at time t, are all bias items corresponding to the neural network, I t is the input gate, c t is the state of the memory cell at time t, is the intermediate value of the updated memory cell state, Tanh is the hyperbolic tangent activation function; c t- 1 is the state of the memory cell at time t-1, o t is the output gate.

[0127] Step S542: sending the training set to the assigned neural network for further training, wherein the output of the assigned neural network at a preset time is calculated by a forward calculation formula, and the loss value at the preset time is calculated based on the output;

[0128] It can be understood that this step calculates the loss value at the preset moment by using a preset loss value calculation formula, wherein the loss value calculation formula is shown as follows:

[0129] Δ t =(T yt -y pt ) 2

[0130] Among them, Δ t Represents the loss value at time t, T yt Represents the observed value of the training sample at time t, y pt Represents the output of the neural network model at time t.

[0131] Step S543: Update the loss value at the preset moment, and predict the data at each moment to obtain the output result corresponding to the historical monitoring data in the neural network.

[0132] This step gradually optimizes the model parameters by updating the loss values ​​and predicting the data at each moment, enabling the model to more accurately predict the most dangerous sliding surface of the geogrid slope. Predictions based on historical monitoring data can verify the model's performance and provide valuable prediction results for practical engineering applications.

[0133] Step S6: sending the historical monitoring data of the geogrid slope to the optimized most dangerous sliding surface prediction model of the geogrid slope to predict the most dangerous sliding surface parameters of the geogrid slope.

[0134] It is understandable that the most dangerous sliding surface prediction system of geogrid slope can consider multiple parameters, has strong adaptability, fault tolerance and self-improvement capabilities, and improves the accuracy of monitoring and calculation.

[0135] Example 2:

[0136] like Figure 2 As shown, this embodiment provides a most dangerous sliding surface prediction system for geogrid slopes, see Figure 2 The system includes an acquisition unit 701 , a calculation unit 702 , a processing unit 703 , a classification unit 704 , an optimization unit 705 and a prediction unit 706 .

[0137] An acquisition unit 701 is used to acquire design data of the geogrid slope and historical monitoring data of the geogrid slope;

[0138] A calculation unit 702 is configured to construct a physical calculation model of the most dangerous sliding surface of the geogrid slope based on the design data of the geogrid slope;

[0139] The processing unit 703 is used to process the design data of the geogrid slope based on the Markov random field, and send all the processed data samples to the physical calculation model of the most dangerous sliding surface for calculation to obtain the data set of the most dangerous sliding surface;

[0140] a classification unit 704 for dividing the data set of the most dangerous sliding surface and the design data of the geogrid slope into a training set and a test set;

[0141] The optimization unit 705 is used to send the training set and the test set to a preset neural network for training and verification, and optimize the preset neural network using a preset improved coronavirus herd immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope;

[0142] The prediction unit 706 is configured to send the historical monitoring data of the geogrid slope to the optimized most dangerous sliding surface prediction model of the geogrid slope to predict the most dangerous sliding surface parameters of the geogrid slope.

[0143] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0144] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

[0145] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for predicting the most dangerous sliding surface of a geogrid slope, characterized in that: include: Obtain design data of geogrid slopes and historical monitoring data of geogrid slopes; Build a physical calculation model of the most dangerous sliding surface of the geogrid slope based on the design data of the geogrid slope; The design data of the geogrid slope is processed based on the Markov random field, and all the processed data samples are sent to the physical calculation model of the most dangerous sliding surface for calculation to obtain the data set of the most dangerous sliding surface; The dataset of the most dangerous sliding surface and the design data of the geogrid slope are divided into a training set and a test set; The training set and the test set are sent to a preset neural network for training and verification, and the preset neural network is optimized by a preset improved coronavirus herd immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model of the geogrid slope; Sending the historical monitoring data of the geogrid slope to the optimized most dangerous sliding surface prediction model of the geogrid slope to predict the most dangerous sliding surface parameters of the geogrid slope; The training set and the test set are sent to a preset neural network for training and verification, and the preset neural network is optimized by a preset improved coronavirus herd immunity optimization algorithm, including: Sending the training set to a preset neural network for training to obtain a trained neural network; Optimizing the hidden layer parameters and learning rate parameters of the trained neural network based on an improved coronavirus herd immunity optimization algorithm to obtain optimized hidden layer parameters and learning rate parameters; Assigning the optimized hidden layer parameters and learning rate parameters to the trained neural network to obtain an assigned neural network; Sending the training set and the test set to the assigned neural network for training and verification to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope; The hidden layer parameters and learning rate parameters of the trained neural network are optimized based on the improved coronavirus herd immunity optimization algorithm to obtain the optimized hidden layer parameters and learning rate parameters, including: Initialize the hidden layer parameters and learning rate parameters in the preset neural network, and set the parameters of the preset improved coronavirus herd immunity optimization algorithm, including the algorithm population size, the initial number of infected cases, the problem dimension, the maximum number of iterations, the basic reproduction rate, and the maximum age of infected cases; Randomly generate individual genes with a preset population size, evenly distribute the individual genes based on the Picewise chaos mapping method, and change the initial positions of the individual genes to obtain the optimized individual gene positions; The optimized individual gene positions are brought into the improved coronavirus herd immunity optimization algorithm, and based on the improved coronavirus herd immunity optimization algorithm with the optimized individual gene positions, the hidden layer parameters and learning rate parameters of the neural network are differentially mutated, crossover and selected to obtain the screened excellent individual genes; The fitness values ​​of the excellent individual genes after screening are calculated, and the individual genes are updated based on the calculated fitness values ​​until the number of iterations reaches the maximum number of iterations, thereby obtaining the optimal hidden layer parameters and learning rate parameters.

2. The method for predicting the most dangerous sliding surface of a geogrid slope according to claim 1, characterized in that ,Based on the design data of geogrid slope, a physical calculation model of the most dangerous sliding surface of geogrid slope is built, including: Perform force analysis on any horizontal slider on the slope and calculate the normal force and tangential force on the sliding surface of the horizontal slider; The minimum tension of the slope geogrid is calculated based on the normal force and tangential force on the sliding surface of the horizontal slider; The horizontal length of the sliding surface point at the position of the horizontal slider from the slope surface is calculated based on the minimum tension of the slope geogrid; The position of the sliding surface point is determined based on the horizontal length of the sliding surface point at the horizontal slider position from the slope surface, and the sliding surface points at the positions of each slider are connected to obtain the most dangerous sliding surface.

3. The method for predicting the most dangerous sliding surface of a geogrid slope according to claim 1, characterized in that ,Based on the Markov random field, the design data of the geogrid slope is processed, and all the processed data samples are sent to the physical calculation model of the most dangerous sliding surface for calculation, including: The design data of the geogrid slope is exported into 500 random field RGB images through a preset software, and redundant blank areas of the random field RGB images are deleted and the random field RGB images are converted into grayscale images to obtain at least two random field images; The design data parameters of the geogrid slope are randomly combined to correspond to random field images to obtain multiple data samples; All data samples are sent to the physical calculation model of the most dangerous slip surface for calculation to obtain the most dangerous slip surface corresponding to each data sample.

4. A most dangerous sliding surface prediction system for geogrid slopes, characterized by: include: An acquisition unit, used for acquiring design data of the geogrid slope and historical monitoring data of the geogrid slope; A calculation unit, used to build a physical calculation model of the most dangerous sliding surface of the geogrid slope based on the design data of the geogrid slope; A processing unit is used to process the design data of the geogrid slope based on the Markov random field, and send all the processed data samples to the physical calculation model of the most dangerous sliding surface for calculation to obtain the data set of the most dangerous sliding surface; A classification unit is used to divide the dataset of the most dangerous sliding surface and the design data of the geogrid slope into a training set and a test set; An optimization unit is used to send the training set and the test set to a preset neural network for training and verification, and optimize the preset neural network using a preset improved coronavirus herd immunity optimization algorithm to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope; A prediction unit, configured to send the historical monitoring data of the geogrid slope to the optimized most dangerous sliding surface prediction model of the geogrid slope, and predict the most dangerous sliding surface parameters of the geogrid slope; Wherein, the optimization unit includes: A first optimization subunit is configured to send the training set to a preset neural network for training to obtain a trained neural network; A second optimization subunit is used to optimize the hidden layer parameters and learning rate parameters of the trained neural network based on an improved coronavirus herd immunity optimization algorithm to obtain optimized hidden layer parameters and learning rate parameters; A third optimization subunit is used to assign the optimized hidden layer parameters and learning rate parameters to the trained neural network to obtain an assigned neural network; A fourth optimization subunit is used to send the training set and the test set to the assigned neural network for training and verification, so as to obtain an optimized most dangerous sliding surface prediction model for the geogrid slope; Wherein, the second optimization subunit includes: The fifth optimization subunit is used to initialize the hidden layer parameters and learning rate parameters in the preset neural network, and set the parameters of the preset improved coronavirus herd immunity optimization algorithm, including the algorithm population size, the initial number of infected cases, the problem dimension, the maximum number of iterations, the basic reproduction rate, and the maximum age of infected cases; The sixth optimization subunit is used to randomly generate individual genes with a population size of a preset number, uniformly distribute the individual genes based on the Picewise chaos mapping method, and change the initial positions of the individual genes to obtain optimized individual gene positions; The seventh optimization subunit is used to bring the optimized individual gene positions into the improved coronavirus group immunity optimization algorithm, and perform differential mutation, crossover and selection on the hidden layer parameters and learning rate parameters of the neural network based on the improved coronavirus group immunity optimization algorithm with the optimized individual gene positions brought in to obtain the screened excellent individual genes; The eighth optimization subunit is used to calculate the fitness value of the excellent individual genes after screening, and update the individual genes based on the calculated fitness value until the number of iterations reaches the maximum number of iterations, thereby obtaining the optimal hidden layer parameters and learning rate parameters.

5. The most dangerous sliding surface prediction system of geogrid slope according to claim 4, characterized in that: The computing unit includes: The first calculation subunit is used to perform force analysis on any horizontal slider of the slope and calculate the normal force and tangential force on the sliding surface of the horizontal slider; A second calculation subunit is used to calculate the minimum tension of the slope geogrid according to the normal force and the tangential force on the sliding surface of the horizontal slider; A third calculation subunit is configured to calculate the horizontal length of the sliding surface point at the position of the horizontal slider from the slope surface according to the minimum tension of the slope geogrid; The fourth calculation subunit is used to determine the position of the sliding surface point based on the horizontal length between the sliding surface point at the position of the horizontal slider and the slope surface, and connect the sliding surface points at the positions of the sliders to obtain the most dangerous sliding surface.

6. The most dangerous sliding surface prediction system of geogrid slope according to claim 4, characterized in that: The processing unit includes: A first processing subunit is configured to export the design data of the geogrid slope into 500 random field RGB images through a preset software, delete redundant blank areas of the random field RGB images, and convert the random field RGB images into grayscale images to obtain at least two random field images; The second processing subunit is used to randomly combine the design data parameters of the geogrid slope corresponding to the random field images to obtain multiple data samples; The third processing subunit is used to send all data samples to the physical calculation model of the most dangerous slip surface for calculation, so as to obtain the most dangerous slip surface corresponding to each data sample.

Images