Intelligent slope monitoring method, system and storage medium based on digital twin

Abstract

The present invention discloses a method, system and storage medium for intelligent monitoring of slopes based on digital twins, which belongs to the field of digital twin technology, and includes: obtaining point cloud data of the slope surface to create a digital twin model; dividing the digital twin model to generate multiple first grids, and calculating the displacement change value at the first grid within a target time length; defining the first grid whose displacement change value is greater than a first threshold as a first area, and dividing the target time length into multiple first time lengths of unequal time; setting multiple rainfall modes, screening out the optimal mode for each first time length, and calculating the corresponding representative rainfall, constructing a prediction model, and predicting and generating a representative rainfall in a future time period according to the optimal mode and representative rainfall of historical rainfall data in the first time length; predicting and generating a safety factor of the slope in the first area in the future time period based on the representative rainfall and the pore water pressure of the soil, and the accuracy of slope stability monitoring is improved through the present invention.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital twins, and specifically relates to a slope intelligent monitoring method, system and storage medium based on digital twins. Background Art

[0002] In slope engineering, intelligent monitoring using digital twin technology is an emerging method that combines advanced sensor technology, data acquisition, data processing and numerical simulation technology to improve the accuracy and real-time performance of slope stability assessment.

[0003] For example, the Chinese patent application with publication number CN117763969A proposes a slope monitoring method, device, equipment and storage medium based on digital twins, which includes: obtaining monitoring data sets of various types of sensors installed on the slope to be monitored, performing sensitivity analysis on the monitoring data in the monitoring data sets, and screening sensitive data sets that meet the preset sensitivity; inputting the sensitive data sets into the pre-built digital twin model, inverting to obtain the inversion data set of the slope to be monitored; fusing the inversion data set to obtain the comprehensive monitoring information curve of the slope to be monitored; and determining the state of the slope to be monitored based on the inclination of the comprehensive monitoring information curve. For another example, the Chinese patent application with publication number CN117828732A discloses a slope stability determination method, system, medium and terminal based on digital twins, which involves the field of digital twin technology and the field of ionic rare earth mines, and its main purpose is to avoid the problem of changes in the stress conditions of the mine slope when injecting liquid leaching agent into the ore body, thereby affecting the stability of the mine slope. The method includes: receiving a determination instruction for the slope stability parameters of a target mine, and carrying a target time in the determination instruction; determining the groundwater level data of the target mine at the target time from the water potential field time distribution data of the target mine; performing a three-dimensional slope stability calculation based on the groundwater level data based on the strength reduction method, and obtaining a prediction result of the mine slope stability of the target mine at the target time.

[0004] However, the above-mentioned existing technologies for intelligent slope monitoring may not be able to fully integrate the impact of multiple factors on slope stability, such as displacement changes, weather conditions and soil properties, which limits the monitoring system's ability to comprehensively evaluate slope stability. Summary of the invention

[0005] To solve the above problems, the present invention provides a slope intelligent monitoring method, system and storage medium based on digital twins to solve the problems in the prior art.

[0006] In order to achieve the above-mentioned purpose of the invention, the present invention proposes a slope intelligent monitoring method based on digital twin, comprising:

[0007] S1: Acquire first point cloud data of the slope surface at a preset time point in the past, and create a digital twin model based on the first point cloud data;

[0008] S2: Acquire the second point cloud data of the slope surface at the current time point, convert the first point cloud data, the second point cloud data and the digital twin model into the same coordinate system, mesh the digital twin model based on a preset size to generate a plurality of first grids, define the duration from the preset time point to the current time point as a target duration, calculate the displacement change value at each of the first grids within the target duration, and display it in the digital twin model;

[0009] S3: defining the first grid whose displacement change value is greater than the first threshold as a first area, obtaining the actual rainfall in the first area within the target duration, establishing a first coordinate system with time as the horizontal axis and the actual rainfall as the vertical axis, plotting a first variation curve of the actual rainfall over time in the first coordinate system, obtaining the extreme value points of the first variation curve, defining the time point corresponding to each extreme value point as a first time point, and dividing the target duration into a plurality of first durations of unequal time based on the first time point;

[0010] S4: setting a plurality of rainfall modes, each of which corresponds to a plurality of historical rainfall data, selecting a rainfall mode from the rainfall modes as an optimal mode based on the actual rainfall within the first time period, obtaining an optimal time point within each of the first time period based on the optimal mode and the historical rainfall data, using the actual rainfall corresponding to the optimal time point as the representative rainfall within the first time period, building a prediction model based on the historical rainfall data, predicting and generating the optimal mode and the representative rainfall within a future time period based on the optimal mode and the representative rainfall within each of the first time periods, and defining the representative rainfall within the future time period as predicted rainfall;

[0011] S5: Obtain historical rainfall within the target time period, obtain a first pore water pressure distribution of the soil in the first area based on the historical rainfall, predict and generate a first safety factor of the first area in the future time period based on the predicted rainfall and the first pore water pressure distribution, take the average of all the first safety factors of the slope as the final safety factor of the slope, and display it in the digital twin model.

[0012] Further, calculating the displacement change value at each of the first grids includes the following steps:

[0013] Define a vertex of the first mesh in the digital twin model as a first node, and define the three-dimensional position coordinates after the first point cloud data and the second point cloud data are converted as first position coordinates and second position coordinates, respectively;

[0014] Obtain a first number of the first point cloud data and a second number of the second point cloud data in the first grid. If both the first number and the second number are less than or equal to 1, a vector formed by taking the first node as a starting point and the first position coordinate of the first point cloud data as an end point is defined as a first vector, and a vector formed by taking the first node as a starting point and the second position coordinate of the second point cloud data as an end point is defined as a second vector;

[0015] If the first number is greater than 1, the barycentric coordinates of all the first point cloud data in the first grid are calculated, and a vector formed by taking the first node as a starting point and the barycentric coordinates of the first point cloud data as an end point is used as the first vector; if the second number is greater than 1, the barycentric coordinates of all the second point cloud data in the first grid are calculated, and a vector formed by taking the first node as a starting point and the barycentric coordinates of the second point cloud data as an end point is used as the second vector;

[0016] Get the unit vector of the first grid in the normal direction , respectively calculate the first component of the first vector and the second vector in the normal direction based on the first formula and the second formula and the second component , the first formula is: , the second formula is: ,in, and are respectively the first vector and the second vector, and the difference between the first component and the second component is defined as the displacement change value of the first grid.

[0017] Further, selecting the optimal mode from the rainfall modes comprises the following steps:

[0018] Calculate the average value of the actual rainfall of the first change curve at multiple first time points. If the first duration is greater than the second threshold, the average value is greater than the third threshold, and the first time point is a maximum point, match the optimal mode to the first mode. If the first duration is greater than the second threshold, the average value is less than the third threshold, and the first time point is a minimum point, match the optimal mode to the second mode. If the first duration is less than the second threshold, the average value is greater than the third threshold, and the first time point is a maximum point, match the optimal mode to the third mode. If the first duration is less than the second threshold, the average value is less than the third threshold, and the first time point is a minimum point, match the optimal mode to the fourth mode.

[0019] Further, obtaining the optimal time point within each of the first durations includes the following steps:

[0020] Obtain the average rainfall of multiple pieces of historical rainfall data corresponding to the optimal mode at the first time point, calculate the difference between the actual rainfall at each of the first time points and the corresponding average rainfall, and define the time point when the difference is less than or equal to a fourth threshold as the optimal time point.

[0021] Furthermore, building a prediction model includes the following steps:

[0022] The time period in which each piece of the historical rainfall data is located is defined as a historical time length, the historical time length is divided into multiple first time lengths, the corresponding optimal mode is matched for each first time length, and the representative rainfall under the optimal mode is calculated, a division point is set in the historical time length, the optimal mode and the representative rainfall of each first time length in the time period before the division point are used as input features, one of the first time lengths is selected after the division point, and the corresponding optimal mode and the representative rainfall are used as output features, this step is repeated to process each piece of the historical rainfall data, a neural network model is constructed, and the neural network model is trained based on the input features and the corresponding output features to obtain the prediction model.

[0023] Further, generating a first safety factor of the first area at the future time period comprises the following steps:

[0024] If the predicted rainfall in the future time period is greater than the fifth threshold value, the first pore water pressure distribution is used as the initial value and the predicted rainfall is used as the boundary condition, and the second pore water pressure distribution of the soil in the future time period is predicted based on the one-dimensional permeability model, and the second pore water pressure of the soil in the first area is obtained based on the second pore water pressure distribution, and the effective stress of the soil in the first area is calculated based on the second pore water pressure, and the effective stress is the difference between the total stress of the soil in the first area and the second pore water pressure, and the shear strength of the soil under the second pore water pressure is obtained, and the ratio of the shear strength to the effective stress is used as the first safety factor of the first area in the future time period.

[0025] The present invention also provides a slope intelligent monitoring system based on digital twins, which is used to implement the above-mentioned method. The system mainly includes:

[0026] A creation module, used to obtain first point cloud data of the slope surface at a preset time point in the past, and create a digital twin model based on the first point cloud data;

[0027] a calculation module, used to obtain second point cloud data of the slope surface at the current time point, convert the first point cloud data, the second point cloud data and the digital twin model into the same coordinate system, mesh the digital twin model based on a preset size to generate a plurality of first grids, define the duration from the preset time point to the current time point as a target duration, calculate the displacement change value at each of the first grids within the target duration, and display it in the digital twin model;

[0028] A division module, used for defining the first grid whose displacement change value is greater than a first threshold as a first area, obtaining the actual rainfall in the first area within the target duration, establishing a first coordinate system with time as the horizontal axis and the actual rainfall as the vertical axis, plotting a first variation curve of the actual rainfall over time in the first coordinate system, obtaining extreme value points of the first variation curve, defining a time point corresponding to each extreme value point as a first time point, and dividing the target duration into a plurality of first durations of unequal time based on the first time point;

[0029] A prediction module is used to set a plurality of rainfall modes, each of which corresponds to a plurality of historical rainfall data, select a rainfall mode from the rainfall modes as an optimal mode based on the actual rainfall in the first time period, obtain an optimal time point in each of the first time period based on the optimal mode and the historical rainfall data, use the actual rainfall corresponding to the optimal time point as a representative rainfall in the first time period, build a prediction model based on the historical rainfall data, predict and generate the optimal mode and the representative rainfall in a future time period according to the optimal mode and the representative rainfall in each of the first time periods, and define the representative rainfall in the future time period as predicted rainfall;

[0030] A monitoring module is used to obtain historical rainfall within the target time period, obtain a first pore water pressure distribution of the soil in the first area based on the historical rainfall, predict and generate a first safety factor of the first area in the future time period based on the predicted rainfall and the first pore water pressure distribution, take the average of all the first safety factors of the slope as the final safety factor of the slope, and display it in the digital twin model.

[0031] The present invention also provides a computer storage medium, which stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the above-mentioned method.

[0032] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0033] By acquiring the first point cloud data at a preset time point in the past, the present invention can create an accurate digital twin model of the slope. This model can reflect the initial state of the slope in detail and provide an accurate basis for subsequent analysis and prediction; the digital twin model is gridded to generate multiple first grids, and the displacement change of the slope in each first grid from the preset time point to the current time point is calculated to more carefully analyze the displacement change of the slope in each local area; the first grid whose displacement change value is greater than the preset threshold is defined as the first area. The present invention focuses on the first area with a large landslide risk or strong instability, and analyzes and predicts the rainfall in the first area.

[0034] The present invention also divides the target time into multiple first time periods of different time periods, matches a corresponding rainfall pattern for each first time period, each rainfall pattern includes multiple historical rainfall data, and constructs a prediction model based on the rainfall data. The prediction model is a neural network model, and the historical time period where each historical rainfall data is located is also divided into multiple first time periods. The optimal pattern is learned to be assigned to each first time period, and the representative rainfall under the optimal pattern is calculated. The neural network model learns the rainfall change trend under each rainfall pattern corresponding to each first time period in the historical time period, and predicts and generates the optimal rainfall pattern and predicted rainfall in the future time period, thereby improving the prediction accuracy of the future rainfall trend; the first pore water pressure distribution of the slope soil in the first area of ​​the future time period is determined according to the predicted rainfall in the future time period and the historical rainfall in the target time period, and the first safety factor of the slope is obtained according to the first pore water pressure distribution. By predicting the slope stability under future rainfall conditions, preventive measures can be taken in advance to reduce disasters such as landslides and mud-rock flows. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flowchart of the steps of the digital twin-based intelligent slope monitoring method of the present invention;

[0036] Figure 2 is the first grid diagram of the present invention;

[0037] Figure 3 is a first change curve diagram of the present invention;

[0038] Figure 4 This is a system structure diagram of the digital twin-based intelligent slope monitoring system of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0040] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0041] like Figure 1 As shown, a slope intelligent monitoring method based on digital twins includes:

[0042] S1: Obtain first point cloud data of the slope surface at a preset time point in the past, and create a digital twin model based on the first point cloud data.

[0043] Specifically, in order to clearly monitor the displacement changes and stability of the slope, it is first necessary to construct a digital twin model of the slope based on the three-dimensional point cloud data of the slope to obtain the entire area of ​​the entire slope, and then analyze the displacement changes of the slope; in this embodiment, the slope is not limited to natural slopes, clay slopes, coastal or riverside slopes, etc.; a three-dimensional laser scanner or a scanning radar of a drone is used to obtain point cloud data of the slope surface at a preset time point in the past (0:00 in the morning), and the point cloud data is preprocessed, including noise removal, filtering and data registration, etc., and the processed point cloud data is converted into a three-dimensional model using three-dimensional modeling software such as AutoCAD, and the key geological features of the slope, such as structural surfaces, cracks, slopes, etc., are extracted from the three-dimensional model. Finally, these multi-source data are integrated to construct a digital twin model.

[0044] S2: Obtain the second point cloud data of the slope surface at the current time point, convert the first point cloud data, the second point cloud data and the digital twin model into the same coordinate system, mesh the digital twin model based on a preset size to generate multiple first grids, define the time between the preset time point and the current time point as the target time length, calculate the displacement change value at each first grid within the target time length, and display it in the digital twin model.

[0045] Specifically, coordinate transformation operations such as translation, rotation and scaling are performed on the first point cloud data and the second point cloud data of the slope surface, and the first point cloud data, the second point cloud data and the digital twin model are aligned with the coordinates to ensure that they are in the same coordinate system. In order to make the calculation of the slope displacement change more accurate, the model is gridded according to a predetermined size, such as 2m*2m. The slope can be divided into smaller parts through grid division. Each grid can represent a small block of the slope, and its displacement can be calculated independently. The displacement change value of each first grid within the target time length is calculated and finally displayed in the digital twin model. The specific calculation method will be introduced later.

[0046] S3: Define the first grid whose displacement change value is greater than the first threshold as the first area, obtain the actual rainfall in the first area within the target time length, establish a first coordinate system with time as the horizontal axis and the actual rainfall as the vertical axis, plot a first change curve of the actual rainfall over time in the first coordinate system, obtain the extreme value points of the first change curve, define the time point corresponding to each extreme value point as the first time point, and divide the target time length into multiple first time lengths of unequal time based on the first time point.

[0047] Specifically, the first area is a grid area where the slope displacement change value is large, the instability is strong, and landslides are prone to occur. The first threshold is set based on the actual situation, and the important factors causing landslides in the first area include the rainfall in the area. Heavy rainfall or continuous rainfall will increase the pore water pressure of the soil and reduce the shear strength of the soil, thereby increasing the risk of landslides. This embodiment mainly analyzes and predicts the stability of the slope based on precipitation. The purpose of setting the first change curve is to intuitively show the changes in the actual rainfall in the first area over time during the duration of rainfall. Figure 3 , obtain extreme points in the first change curve, such as s1, s2, s3, s4, s5, etc., and divide 1 day into multiple first time periods of different time periods, including 0-4h, 4-8h, 12-20h, 20-22h, and 22-24h.

[0048] S4: multiple rainfall modes are set, each rainfall mode corresponds to multiple historical rainfall data, a rainfall mode is selected from the rainfall modes as the optimal mode based on the actual rainfall in the first time period, the optimal time point in each first time period is obtained based on the optimal mode and the historical rainfall data, the actual rainfall corresponding to the optimal time point is used as the representative rainfall in the first time period, a prediction model is constructed based on the historical rainfall data, the optimal mode and representative rainfall in the future time period are predicted and generated according to the optimal mode and representative rainfall in each first time period, and the representative rainfall in the future time period is defined as the predicted rainfall.

[0049] Specifically, according to the rainfall change trend of the slope in the historical time period, a variety of rainfall patterns are obtained, including but not limited to the long-term heavy rainfall pattern, the long-term weak rainfall pattern, the short-term heavy rainfall pattern and the short-term weak rainfall pattern. The rainfall pattern that best matches the rainfall pattern in each first time period is screened out from the multiple rainfall patterns as the optimal pattern. In this way, the change of the rainfall pattern in the target time period can be obtained, and a prediction model is constructed. The optimal pattern of each first time period in the target time period and the average rainfall of the optimal pattern are input into the prediction model to obtain the predicted rainfall in the future time period. This step splits the target time period into multiple first time periods of unequal time, so as to better analyze the rainfall in each stage of the target time period in a more detailed manner. The rainfall change trend of each unequal time period is different, and the corresponding rainfall pattern is matched for each unequal first time period based on the change trend.

[0050] By matching the corresponding rainfall patterns to multiple first time periods within the target time period, the corresponding representative rainfall is calculated. Based on these data, a prediction model is used to predict the rainfall pattern and representative rainfall in future time periods, so that countermeasures can be taken in advance.

[0051] Step S5: Obtain historical rainfall within the target time period, obtain the first pore water pressure distribution of the soil in the first area based on the historical rainfall, predict and generate the first safety factor of the first area in the future time period based on the predicted rainfall and the first pore water pressure distribution, take the average of all first safety factors of the slope as the final safety factor of the slope, and display it in the digital twin model.

[0052] Specifically, it is defined that if there is no rainfall for a long period of time before this rainfall, the soil is in a dry state, otherwise it is in a wet state, and the first absorption capacity of the soil in the first area is obtained. The first absorption capacity is the maximum value of the water absorbed by the soil in a dry state. The standard for defining the dry state is based on a one-dimensional permeability model combined with the first absorption capacity to obtain the initial pore water pressure distribution of the first area in a dry state, and the initial pore water pressure distribution is defined as the first pore water pressure distribution. If the historical rainfall is greater than or equal to a preset threshold, the first absorption capacity of the soil in the first area in the historical time period is obtained, and the initial pore water pressure distribution of the first area is obtained based on the first absorption capacity. Taking the initial pore water pressure distribution as the initial value and the historical rainfall as the boundary condition, the first pore water pressure distribution of the first area in the target time length is calculated based on the one-dimensional permeability model.

[0053] Based on the predicted rainfall and the first pore water pressure distribution, a one-dimensional permeability model is used to predict and generate the first safety factor of the first area in the future time period. Each first area has a first safety factor, and the average value of all first safety factors is used as the final safety factor of the slope. If the final safety factor is less than the preset threshold, it is determined that the slope has a landslide hazard. All first areas are located, and the slope of each area is repaired in advance according to the first safety factor of the area.

[0054] By acquiring the first point cloud data at a preset time point in the past, the present invention can create an accurate digital twin model of the slope. This model can reflect the initial state of the slope in detail and provide an accurate basis for subsequent analysis and prediction; the digital twin model is gridded to generate multiple first grids, and the displacement change of the slope in each first grid from the preset time point to the current time point is calculated to more carefully analyze the displacement change of the slope in each local area; the first grid whose displacement change value is greater than the preset threshold is defined as the first area. The present invention focuses on the first area with a large landslide risk or strong instability, and analyzes and predicts the rainfall in the first area.

[0055] The present invention also divides the target time into multiple first time periods of different time periods, matches a corresponding rainfall pattern for each first time period, each rainfall pattern includes multiple historical rainfall data, and constructs a prediction model based on the rainfall data. The prediction model is a neural network model, and the historical time period where each historical rainfall data is located is also divided into multiple first time periods. The optimal pattern is learned to be assigned to each first time period, and the representative rainfall under the optimal pattern is calculated. The neural network model learns the rainfall change trend under each rainfall pattern corresponding to each first time period in the historical time period, and predicts and generates the optimal rainfall pattern and predicted rainfall in the future time period, thereby improving the prediction accuracy of the future rainfall trend; the first pore water pressure distribution of the slope soil in the first area of ​​the future time period is determined according to the predicted rainfall in the future time period and the historical rainfall in the target time period, and the first safety factor of the slope is obtained according to the first pore water pressure distribution. By predicting the slope stability under future rainfall conditions, preventive measures can be taken in advance to reduce disasters such as landslides and mud-rock flows.

[0056] As a preferred technical solution of the present invention, calculating the displacement change value at each first grid includes the following steps:

[0057] A vertex of a first mesh in the digital twin model is defined as a first node, and three-dimensional position coordinates after conversion of the first point cloud data and the second point cloud data are defined as first position coordinates and second position coordinates, respectively;

[0058] Obtain a first number of first point cloud data and a second number of second point cloud data in the first grid. If both the first number and the second number are less than or equal to 1, a vector formed by taking the first node as the starting point and the first position coordinate of the first point cloud data as the end point is defined as a first vector, and a vector formed by taking the first node as the starting point and the second position coordinate of the second point cloud data as the end point is defined as a second vector;

[0059] If the first number is greater than 1, the barycentric coordinates of all first point cloud data in the first grid are calculated, and a vector formed by taking the first node as the starting point and the barycentric coordinates of the first point cloud data as the end point is used as the first vector; if the second number is greater than 1, the barycentric coordinates of all second point cloud data in the first grid are calculated, and a vector formed by taking the first node as the starting point and the barycentric coordinates of the second point cloud data as the end point is used as the second vector;

[0060] Get the unit vector of the first mesh in the normal direction , calculate the first component of the first vector and the second vector in the normal direction based on the first formula and the second formula respectively and the second component , the first formula is: , the second formula is: ,in, and are respectively the first vector and the second vector, and the difference between the first component and the second component is defined as the displacement change value of the first grid.

[0061] Specifically, Figure 2 As shown in the figure, it is the first grid diagram, and point K in the first grid is defined as the first node. The first grid has b1, b2, c1, c2 and c3, where b1 and b2 are the first point cloud data, and c1, c2 and c3 are the second point cloud data. Because the number of the first point cloud data and the second point cloud data is greater than 1, the centroid coordinates of point B of b1 and b2 and the centroid coordinates of point C of c1, c2 and c3 are calculated, and point K and point B are connected as the first vector , the coordinates of point K and point C are used as the second vector , calculate the first component of the first vector in the normal direction , and the second component of the second vector in the normal direction , the difference between the first and second components By calculating the components of the first vector and the second vector in the normal direction as the displacement change value of the first grid, the displacement of the first grid in the vertical direction of the slope at different time points is obtained. By only focusing on the normal component, the error caused by terrain undulation, measurement error or data noise can be reduced, thereby improving the accuracy of displacement measurement.

[0062] Screening out the best pattern from rainfall patterns includes the following steps:

[0063] Calculate the average value of the actual rainfall of the first change curve at multiple first time points. If the first duration is greater than the second threshold, the average value is greater than the third threshold, and the first time point is a maximum point, the optimal mode is matched to the first mode. If the first duration is greater than the second threshold, the average value is less than the third threshold, and the first time point is a minimum point, the optimal mode is matched to the second mode. If the first duration is less than the second threshold, the average value is greater than the third threshold, and the first time point is a maximum point, the optimal mode is matched to the third mode. If the first duration is less than the second threshold, the average value is less than the third threshold, and the first time point is a minimum point, the optimal mode is matched to the fourth mode.

[0064] Specifically, Figure 3As shown, the actual rainfall at the first time points 0:30, 1:00, 1:30, 2:00, 2:30, 3:00, 3:30, and 4:00 is obtained at intervals of 30 minutes, which are 22, 24, 26, 28, 30, 34, 36, and 38 respectively. The average actual rainfall of the first duration 0-4h is 29.75, because the first duration 4h is greater than the second threshold, which is assumed to be 2 hours, and the average value 29.75 is greater than the third threshold, which is assumed to be 25, and the first time point 4h is the maximum. value point, it is determined that the optimal prediction mode match of the first time length 0-s1 is the first mode, that is, the long-term heavy rainfall mode. According to this step, the rainfall mode in each first time length is obtained. The rainfall mode of the first time length 4-8h is the second mode, that is, the long-term weak rainfall mode, the rainfall mode of the first time length 8-12h is the first mode, that is, the long-term heavy rainfall mode, the rainfall mode of the first time length 12-20h is the second mode, that is, the long-term weak rainfall mode, and the rainfall mode of the first time length 20-22h is the third mode, that is, the short-term heavy rainfall mode.

[0065] Obtaining the optimal time point in each first duration includes the following steps:

[0066] The average rainfall at the first time point of multiple historical rainfall data corresponding to the optimal mode is obtained, the difference between the actual rainfall at each first time point and the corresponding average rainfall is calculated, and the time point when the difference is less than or equal to the fourth threshold is defined as the optimal time point.

[0067] Specifically, the optimal mode corresponding to the first duration 0-4h is obtained: in the long-term heavy rainfall mode, the first time points in this rainfall mode include 0:30, 1:00, 1:30, 2:00, 2:30, 3:00, 3:30, and 4:00, and the average rainfall at these first time points are obtained, which are 25, 26, 27, 28, 29, 30, 31, and 33 respectively. The error between the average rainfall and the actual rainfall at each first time point is calculated, and the difference at the first time point 2:00 is 0 by calculation, which is less than or equal to the fourth threshold value 0, and the moment 2:00 is defined as the optimal time point. By confirming the optimal time point, a rainfall with the highest accuracy and the smallest error can be obtained as a representative rainfall. Taking the representative rainfall as a benchmark, the prediction accuracy of future rainfall events can be improved.

[0068] Building a predictive model involves the following steps:

[0069] The time period of each historical rainfall data is defined as the historical time length, and the historical time length is divided into multiple first time lengths. The corresponding optimal mode is matched in each first time length, and the representative rainfall under the optimal mode is calculated. A division point is set in the historical time length, and the optimal mode and representative rainfall of each first time length in the time period before the division point are used as input features. After the division point, a first time length is selected, and the corresponding optimal mode and representative rainfall are used as output features. Repeat this step to process each historical rainfall data, build a neural network model, and train the neural network model based on the input features and the corresponding output features to obtain a prediction model.

[0070] Specifically, based on the method of dividing the first time period within the target time period, and the method of matching the optimal pattern for each first time period and calculating the representative rainfall, each historical time period is divided into multiple first time periods (the historical time period also includes the target time period, which can be rainfall data for multiple days), a point is selected in the time series of the historical time period, and the historical rainfall data is divided into two parts: before the division point and after the division point. This division point is used to distinguish the model training data from the model verification data. The optimal pattern and representative rainfall of all first time periods before the division point are used as input features (i.e., feature sets), and the rainfall pattern and representative rainfall of the first "first time period" after the division point are used as output features (i.e., target values). The purpose of this is to train the model so that it can predict the optimal pattern and representative rainfall in the future time period based on the historical rainfall data. A neural network model is constructed using the above input features and output features. Through multiple iterations, the model parameters are adjusted. After training, the neural network model can capture the time series characteristics and patterns of the rainfall data, thereby predicting the optimal pattern and representative rainfall in the future time period. The pore water pressure distribution of the soil in the future time period is obtained in advance based on the representative precipitation in the future time period, and the stability of the slope is further analyzed.

[0071] Generating a first safety factor for a first region at a future time period comprises the following steps:

[0072] If the predicted rainfall in the future time period is greater than the fifth threshold, the first pore water pressure distribution is used as the initial value and the predicted rainfall is used as the boundary condition. The second pore water pressure distribution of the soil in the future time period is predicted based on the one-dimensional permeability model. The second pore water pressure of the soil in the first area is obtained based on the second pore water pressure distribution. The effective stress of the soil in the first area is calculated based on the second pore water pressure, and the effective stress is the difference between the total stress of the soil in the first area and the second pore water pressure. The shear strength of the soil under the second pore water pressure is obtained, and the ratio of the shear strength to the effective stress is used as the first safety factor of the first area in the future time period.

[0073] Specifically, the predicted rainfall in the future time period is obtained and compared with the preset fifth threshold. If the predicted rainfall exceeds this threshold, it means that a large amount of water may infiltrate the soil, affecting the stability of the slope. The one-dimensional permeability model is used to predict the soil pore water pressure distribution in the future time period with the current pore water pressure distribution as the initial value and the predicted rainfall as the boundary condition. Based on the predicted pore water pressure distribution, the effective stress of the soil in the first area is calculated. The effective stress refers to the difference between the total soil stress and the pore water pressure, which is an important parameter for evaluating soil stability. When the second pore water pressure is known, the shear strength of the soil is obtained. Shear strength is the ability of soil to resist shear deformation, which is crucial for evaluating slope stability. The ratio of the soil shear strength to the effective stress is used as the first safety factor of the first area in the predicted future time period. The safety factor is a direct indicator for evaluating slope stability. The higher the coefficient, the more stable the slope. By predicting the safety factor in the future time period, the potential risk of slope instability can be discovered in advance, providing a time window for taking preventive measures.

[0074] like Figure 4 As shown, the present invention also provides a slope intelligent monitoring system based on digital twins, which is used to implement the above method. The system mainly includes:

[0075] A creation module is used to obtain first point cloud data of the slope surface at a preset time point in the past, and to create a digital twin model based on the first point cloud data.

[0076] A calculation module is used to obtain the second point cloud data of the slope surface at the current time point, convert the first point cloud data, the second point cloud data and the digital twin model into the same coordinate system, mesh the digital twin model based on a preset size to generate multiple first grids, define the time between the preset time point and the current time point as the target time length, calculate the displacement change value at each first grid within the target time length, and display it in the digital twin model.

[0077] A division module is used to define a first grid whose displacement change value is greater than a first threshold as a first area, obtain the actual rainfall in the first area within a target time length, establish a first coordinate system with time as the horizontal axis and the actual rainfall as the vertical axis, plot a first change curve of the actual rainfall over time in the first coordinate system, obtain the extreme value points of the first change curve, define the time point corresponding to each extreme value point as a first time point, and divide the target time length into multiple first time lengths of unequal time based on the first time point.

[0078] The prediction module is used to set a variety of rainfall modes, each rainfall mode corresponds to a plurality of historical rainfall data, a rainfall mode is selected from the rainfall modes as the optimal mode based on the actual rainfall in the first time period, the optimal time point in each first time period is obtained based on the optimal mode and the historical rainfall data, the actual rainfall corresponding to the optimal time point is used as the representative rainfall in the first time period, a prediction model is constructed based on the historical rainfall data, and the optimal mode and representative rainfall in the future time period are predicted and generated according to the optimal mode and representative rainfall in each first time period, and the representative rainfall in the future time period is defined as the predicted rainfall.

[0079] The monitoring module is used to obtain the historical rainfall within the target time period, obtain the first pore water pressure distribution of the soil in the first area based on the historical rainfall, and predict and generate the first safety factor of the first area in the future time period based on the predicted rainfall and the first pore water pressure distribution. The average of all the first safety factors of the slope is used as the final safety factor of the slope and displayed in the digital twin model.

[0080] The present invention also provides a computer storage medium, which stores program instructions. When the program instructions are executed, the device where the computer storage medium is located is controlled to execute the above method.

[0081] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0082] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0083] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The above embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0085] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A slope intelligent monitoring method based on digital twins, characterized in that: The method comprises the following steps: S1: Acquire first point cloud data of the slope surface at a preset time point in the past, and create a digital twin model based on the first point cloud data; S2: Acquire the second point cloud data of the slope surface at the current time point, convert the first point cloud data, the second point cloud data and the digital twin model into the same coordinate system, mesh the digital twin model based on a preset size to generate a plurality of first grids, define the duration from the preset time point to the current time point as a target duration, calculate the displacement change value at each of the first grids within the target duration, and display it in the digital twin model; S3: defining the first grid whose displacement change value is greater than the first threshold as a first area, obtaining the actual rainfall in the first area within the target duration, establishing a first coordinate system with time as the horizontal axis and the actual rainfall as the vertical axis, plotting a first variation curve of the actual rainfall over time in the first coordinate system, obtaining the extreme value points of the first variation curve, defining the time point corresponding to each extreme value point as a first time point, and dividing the target duration into a plurality of first durations of unequal time based on the first time point; S4: setting a plurality of rainfall modes, each of which corresponds to a plurality of historical rainfall data, selecting a rainfall mode from the rainfall modes as an optimal mode based on the actual rainfall within the first time period, obtaining an optimal time point within each of the first time period based on the optimal mode and the historical rainfall data, using the actual rainfall corresponding to the optimal time point as the representative rainfall within the first time period, building a prediction model based on the historical rainfall data, predicting and generating the optimal mode and the representative rainfall within a future time period based on the optimal mode and the representative rainfall within each of the first time periods, and defining the representative rainfall within the future time period as predicted rainfall; S5: Obtain historical rainfall within the target time period, obtain a first pore water pressure distribution of the soil in the first area based on the historical rainfall, predict and generate a first safety factor of the first area in the future time period based on the predicted rainfall and the first pore water pressure distribution, take the average of all the first safety factors of the slope as the final safety factor of the slope, and display it in the digital twin model.

2. The method according to claim 1, characterized in that: Calculating the displacement change value at each of the first grids comprises the following steps: Define a vertex of the first mesh in the digital twin model as a first node, and define the three-dimensional position coordinates after the first point cloud data and the second point cloud data are converted as first position coordinates and second position coordinates, respectively; Obtain a first number of the first point cloud data and a second number of the second point cloud data in the first grid. If both the first number and the second number are less than or equal to 1, a vector formed by taking the first node as a starting point and the first position coordinate of the first point cloud data as an end point is defined as a first vector, and a vector formed by taking the first node as a starting point and the second position coordinate of the second point cloud data as an end point is defined as a second vector; If the first number is greater than 1, the barycentric coordinates of all the first point cloud data in the first grid are calculated, and a vector formed by taking the first node as a starting point and the barycentric coordinates of the first point cloud data as an end point is used as the first vector; if the second number is greater than 1, the barycentric coordinates of all the second point cloud data in the first grid are calculated, and a vector formed by taking the first node as a starting point and the barycentric coordinates of the second point cloud data as an end point is used as the second vector; Get the unit vector of the first grid in the normal direction , respectively calculate the first component of the first vector and the second vector in the normal direction based on the first formula and the second formula and the second component , the first formula is: , the second formula is: ,in, and are respectively the first vector and the second vector, and the difference between the first component and the second component is defined as the displacement change value of the first grid.

3. The method according to claim 1, characterized in that Screening out the optimal mode from the rainfall modes comprises the following steps: Calculate the average value of the actual rainfall of the first change curve at multiple first time points. If the first duration is greater than the second threshold, the average value is greater than the third threshold, and the first time point is a maximum point, match the optimal mode to the first mode. If the first duration is greater than the second threshold, the average value is less than the third threshold, and the first time point is a minimum point, match the optimal mode to the second mode. If the first duration is less than the second threshold, the average value is greater than the third threshold, and the first time point is a maximum point, match the optimal mode to the third mode. If the first duration is less than the second threshold, the average value is less than the third threshold, and the first time point is a minimum point, match the optimal mode to the fourth mode.

4. The method according to claim 3, characterized in that Obtaining the optimal time point within each of the first durations comprises the following steps: Obtain the average rainfall of multiple pieces of historical rainfall data corresponding to the optimal mode at the first time point, calculate the difference between the actual rainfall at each of the first time points and the corresponding average rainfall, and define the time point when the difference is less than or equal to a fourth threshold as the optimal time point.

5. The method according to claim 3, characterized in that: Building a predictive model involves the following steps: The time period in which each piece of the historical rainfall data is located is defined as a historical time length, the historical time length is divided into multiple first time lengths, the corresponding optimal mode is matched for each first time length, and the representative rainfall under the optimal mode is calculated, a division point is set in the historical time length, the optimal mode and the representative rainfall of each first time length in the time period before the division point are used as input features, one of the first time lengths is selected after the division point, and the corresponding optimal mode and the representative rainfall are used as output features, this step is repeated to process each piece of the historical rainfall data, a neural network model is constructed, and the neural network model is trained based on the input features and the corresponding output features to obtain the prediction model.

6. The method according to claim 1, characterized in that Generating a first safety factor for the first area in the future time period comprises the following steps: If the predicted rainfall in the future time period is greater than the fifth threshold value, the first pore water pressure distribution is used as the initial value and the predicted rainfall is used as the boundary condition, and the second pore water pressure distribution of the soil in the future time period is predicted based on the one-dimensional permeability model, and the second pore water pressure of the soil in the first area is obtained based on the second pore water pressure distribution, and the effective stress of the soil in the first area is calculated based on the second pore water pressure, and the effective stress is the difference between the total stress of the soil in the first area and the second pore water pressure, and the shear strength of the soil under the second pore water pressure is obtained, and the ratio of the shear strength to the effective stress is used as the first safety factor of the first area in the future time period.

7. A digital twin-based intelligent slope monitoring system, used to implement the method according to any one of claims 1 to 6, characterized in that: The system includes the following modules: A creation module, used to obtain first point cloud data of the slope surface at a preset time point in the past, and create a digital twin model based on the first point cloud data; a calculation module, used to obtain second point cloud data of the slope surface at the current time point, convert the first point cloud data, the second point cloud data and the digital twin model into the same coordinate system, mesh the digital twin model based on a preset size to generate a plurality of first grids, define the duration from the preset time point to the current time point as a target duration, calculate the displacement change value at each of the first grids within the target duration, and display it in the digital twin model; A division module, used for defining the first grid whose displacement change value is greater than a first threshold as a first area, obtaining the actual rainfall in the first area within the target duration, establishing a first coordinate system with time as the horizontal axis and the actual rainfall as the vertical axis, plotting a first variation curve of the actual rainfall over time in the first coordinate system, obtaining extreme value points of the first variation curve, defining a time point corresponding to each extreme value point as a first time point, and dividing the target duration into a plurality of first durations of unequal time based on the first time point; A prediction module is used to set a plurality of rainfall modes, each of which corresponds to a plurality of historical rainfall data, select a rainfall mode from the rainfall modes as an optimal mode based on the actual rainfall in the first time period, obtain an optimal time point in each of the first time period based on the optimal mode and the historical rainfall data, use the actual rainfall corresponding to the optimal time point as a representative rainfall in the first time period, build a prediction model based on the historical rainfall data, predict and generate the optimal mode and the representative rainfall in a future time period according to the optimal mode and the representative rainfall in each of the first time periods, and define the representative rainfall in the future time period as predicted rainfall; A monitoring module is used to obtain historical rainfall within the target time period, obtain a first pore water pressure distribution of the soil in the first area based on the historical rainfall, predict and generate a first safety factor of the first area in the future time period based on the predicted rainfall and the first pore water pressure distribution, take the average of all the first safety factors of the slope as the final safety factor of the slope, and display it in the digital twin model.

8. A computer storage medium, characterized in that The computer storage medium stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

Images