Slope support teaching model, experimental method and early warning control method thereof

Abstract

The invention discloses a slope support teaching model, an experimental method and an early warning control method thereof, belongs to the technical field of teaching models, and aims to solve the problem that the landslide early warning teaching of an existing slope model is mainly based on a traditional method for introducing a static threshold value and cannot meet the requirements of an intelligent monitoring era. In order to solve the problem that teaching of a machine learning early warning method based on'dynamic threshold value + automatic decision 'is lack of a whole-process experimental platform integrating data acquisition, feature extraction, model training and model verification, the invention develops a slope landslide dynamic process capable of simulating a slope landslide dynamic process under a multi-working-condition coupling effect, the slope teaching experiment system can also consider the influence of a support structure, simulates the interaction of support and soil, and is suitable for early warning method verification.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of teaching models, in particular to a slope support teaching model, an experimental method and a warning control method thereof. BACKGROUND

[0002] In the field of civil engineering, landslides of mountain slopes are common geological disasters, and slope protection monitoring and landslide mechanism research are important contents of teaching and scientific research of civil engineering majors. At present, the experimental devices for slope simulation on the market have the following problems: some devices can only simulate a single working condition and cannot realize the coordinated simulation of multiple working conditions; the structure of the slope model of some devices is simple, and it is difficult to accurately restore the stress state and sliding characteristics of the real slope, and there is a lack of simulation of supporting structures, resulting in a large deviation between the experimental results and the actual situation. The traditional method of introducing static threshold is mainly used in the teaching of landslide warning based on slope model, which cannot meet the needs of the intelligent monitoring era. SUMMARY

[0003] The present application aims to at least solve one of the technical problems existing in the prior art. To this end, the present application proposes a slope support teaching model, an experimental method and a warning control method thereof.

[0004] The slope support teaching model according to the first aspect of the present application comprises: a box body provided with a containing cavity at the top; a bottom plate provided in the containing cavity in an inclined manner, the bottom plate being provided with a plurality of through avoiding holes, and the top surface of the bottom plate being used to cover the soil body; a vibration mechanism provided on the bottom plate, the vibration mechanism being used to vibrate the soil body; a lifting mechanism provided below the bottom plate; a plurality of supporting piles, the plurality of supporting piles being provided in the plurality of avoiding holes in a one-to-one correspondence, the lifting mechanism being connected to the plurality of supporting piles, and the lifting mechanism driving the plurality of supporting piles to move in the up-down direction so that the supporting piles are extended above the bottom plate or are recovered below the bottom plate; a water seepage mechanism provided on the bottom plate, the water seepage mechanism being used to seep water into the interior of the soil body; a pressurizing mechanism provided at the top of the containing cavity, the pressurizing mechanism being used to pressurize the soil body; a spraying mechanism provided at the top of the containing cavity, the spraying mechanism being used to spray water to the surface of the soil body.

[0005] The slope support teaching model according to the embodiment of the present application has at least the following beneficial effects: the soil body is covered on the bottom plate in the containing cavity to simulate the soil layer structure of the slope, the soil body is vibrated by the vibration mechanism to simulate the vibration condition of the slope, water is infiltrated into the soil body by the water infiltration mechanism to simulate the water infiltration condition of the slope, the soil body is pressurized by the pressurization mechanism to simulate the pressure condition of the slope, and water is sprayed on the surface of the soil body by the spraying mechanism to simulate the condition of the slope in the rainfall environment; and the plurality of support piles are lifted into the soil body above the bottom plate by the lifting mechanism, so that the situation of setting the support piles on the slope can be simulated; the above simulation operations are independent of each other and can be linked with each other, the simulation operations can be appropriately adjusted according to the experimental requirements, a plurality of working conditions can be cooperatively simulated, the stress state and sliding characteristics of the real slope can be accurately restored, and the simulation of the support structure can be realized, so that the experimental structure is closer to the actual situation.

[0006] According to some embodiments of the present application, the slope support teaching model further comprises: a gyroscope sensor arranged on the support pile, the gyroscope sensor measuring the displacement and inclination angle of the support pile; a tension meter arranged in the interior of the soil body, the tension meter measuring the displacement amount of the soil body; a seepage meter arranged in the interior of the soil body, the seepage meter measuring the water content of the soil body; a plurality of pressure sensors respectively arranged in the interior of the soil body, below the pressurization mechanism, and on the side wall of the support pile; a plurality of temperature sensors respectively arranged in the interior of the soil body and on the periphery of the containing cavity; a plurality of vibration sensors respectively arranged on the bottom plate, in the interior of the soil body, and on the support pile; a plu rality of rain gauges arranged in the containing cavity, the rain gauges being arranged below the spraying mechanism, and the rain gauges being used to measure the surface rainfall of the soil body.

[0007] The slope support teaching model experiment method according to the second aspect of the embodiment of the present application is applied to the slope support teaching model according to the above-mentioned embodiment, and the experiment method comprises: obtaining the soil body according to the experimental requirements, and covering the soil body on the top surface of the bottom plate; controlling the lifting mechanism to drive all the support piles to be recycled below the bottom plate; performing a simulation experiment and obtaining real-time parameters of the soil body, the simulation experiment comprising: controlling the vibration mechanism to vibrate the soil body, controlling the water infiltration mechanism to infiltrate water into the soil body, controlling the pressurization mechanism to pressurize the soil body, and controlling the spraying mechanism to spray water on the surface of the soil body; The lifting mechanism is controlled to drive all the support piles to extend above the base plate, and the simulation experiment is repeated to obtain the real-time parameters.

[0008] The teaching model experimental method for slope support according to embodiments of the present invention has at least the following beneficial effects: covering the bottom plate in the containment cavity with soil to simulate the soil structure of the slope; vibrating the soil through a vibration mechanism to simulate slope vibration; seeping water into the soil through a seepage mechanism to simulate slope seepage; pressurizing the soil through a pressurizing mechanism to simulate slope pressure; and spraying water onto the surface of the soil through a spraying mechanism to simulate slope conditions under rainfall. Furthermore, the lifting mechanism can drive multiple support piles to rise into the soil above the bottom plate, simulating the installation of support piles on the slope. The above simulation operations are independent yet interconnected, and the simulation operations can be appropriately adjusted according to the experimental needs to achieve coordinated simulation of multiple working conditions. It can accurately reproduce the stress state and sliding characteristics of the real slope, and also simulate the support structure, making the experimental structure closer to the actual situation.

[0009] According to some embodiments of the present invention, the experimental method further includes: A grid beam is installed on the surface of the soil, the simulation experiment is repeated, and the real-time parameters are obtained. An anchor rod is installed between the lattice beam and the soil. A tension sensor is installed on the anchor rod to measure the traction force of the anchor rod. The simulation experiment is repeated to obtain the real-time parameters. A protective net is set on the surface of the soil, the simulation experiment is repeated, and the real-time parameters are obtained.

[0010] According to some embodiments of the present invention, the real-time parameters include the displacement and inclination angle of the support pile, the displacement of the soil, water content, pressure, temperature, vibration, and surface precipitation.

[0011] According to a third aspect of the present invention, an early warning control method for a slope support teaching model is applied to the slope support teaching model described in the above embodiments, the early warning control method comprising: The slope support teaching model is controlled to simulate experimental conditions and obtain real-time parameters. The real-time parameters are denoised to obtain experimental data, and an adaptive window is established based on the experimental data. A dynamic threshold is calculated based on the adaptive window and the experimental data, and the warning level is divided according to the dynamic threshold. A training database is established based on the experimental conditions, the experimental data, and the warning level. A prediction model is obtained by training a machine learning algorithm based on the training database. The prediction model is used to give an early warning for the slope support teaching model.

[0012] The slope support teaching model early warning control method according to the embodiment of the present application has at least the following beneficial effects: a specific experimental condition is simulated by artificially controlling the slope support teaching model, real-time parameters are obtained from the experimental condition by using various sensors in the teaching model, experimental data are obtained after eliminating abnormal parameters of the real-time parameters, a self-adaptive window is established to monitor specific experimental data, a dynamic threshold is calculated based on the self-adaptive window and data corresponding to the self-adaptive window, the dynamic threshold is used to divide early warning levels, and the above steps are repeatedly performed to obtain a training database, a prediction model is obtained by training a machine learning algorithm using the training database; when a student starts the slope support teaching model to simulate an actual condition, the prediction model gives an early warning for the actual condition, so as to determine the early warning level of the slope under the actual condition. The landslide early warning teaching of the existing slope model mainly uses the traditional method of introducing a static threshold, which cannot meet the needs of the intelligent monitoring era. The machine learning early warning method teaching based on "dynamic threshold + automatic decision" faces the problem of lacking a full-process experimental platform integrating data collection, feature extraction, model training, and model verification. The present application develops a slope teaching experimental system capable of simulating the dynamic process of slope landslide under the coupling action of multiple conditions, considering the influence of supporting structures and simulating the interaction between supporting structures and soil bodies, and suitable for early warning method verification.

[0013] According to some embodiments of the present application, the self-adaptive window is established according to the experimental data, including: A data change rate is obtained according to the experimental data, and a high fluctuation period, a medium fluctuation period, or a stable period is determined according to the data change rate; A window size is determined according to the high fluctuation period, the medium fluctuation period, or the stable period, and the self-adaptive window is set according to the window size; The experimental data of the self-adaptive window are color-labeled according to the high fluctuation period, the medium fluctuation period, or the stable period.

[0014] According to some embodiments of the present application, the dynamic threshold is calculated according to the self-adaptive window and the experimental data, including: Historical data and real-time data are obtained according to the self-adaptive window and the experimental data, and a data change trend is obtained according to the historical data and the real-time data; The dynamic threshold is calculated by using the percentile method in combination with time weighting and trend prediction according to the historical data and the real-time data.

[0015] According to some embodiments of the present application, the early warning control method further includes: Setting a fixed threshold, setting a warning condition according to the experimental data and the fixed threshold; Warning the slope support teaching model according to the warning condition.

[0016] According to some embodiments of the application, the warning of the slope support teaching model by the prediction model comprises: Controlling the slope support teaching model to simulate any working condition to obtain actual parameters; Using the prediction model to obtain the warning level according to the actual parameters. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a structural schematic diagram of the filling soil body of the slope support teaching model of an embodiment of the application; Figure 2 is a structural schematic diagram of the slope support teaching model of an embodiment of the application when the soil body is not filled, and the side plate of the box is hidden; Figure 3 is a sectional view schematic diagram of the slope support teaching model of an embodiment of the application when the soil body is not filled; Figure 4 is a flowchart of the experimental method of the slope support teaching model of an embodiment of the application; Figure 5 is a flowchart of the warning control method of the slope support teaching model of an embodiment of the application.

[0018] Fig. 10 is a structural schematic diagram of the slope support teaching model of an embodiment of the application. DETAILED DESCRIPTION

[0019] The embodiments of the application are described in detail below, and examples of the embodiments are shown in the drawings, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary, only for explaining the application, and cannot be understood as a limitation of the application.

[0020] In the description of the application, it should be understood that the orientation or positional relationship indicated by the terms front, back, up, down, axial, circumferential, etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the application.

[0021] In the description of this invention, "multiple" means two or more; "greater than," "less than," and "exceeding" are understood to exclude the stated number; "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0022] In the description of this invention, it should be noted that terms such as "set up," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0023] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are some embodiments of the present invention, not all embodiments.

[0024] In the field of civil engineering, landslides are common geological hazards that pose a serious threat to engineering construction and safety. Therefore, slope protection monitoring and landslide mechanism research are important contents of teaching and research in civil engineering. In soil mechanics teaching, landslide mechanism analysis mainly focuses on the derivation of limit equilibrium slice formulas and static model demonstrations, making it difficult for students to understand the dynamic evolution process of landslides. Although existing slope simulation systems can simulate the landslide process under single working conditions (groundwater or seismic motion), they do not consider the coupling effect of multiple working conditions and the influence of support structures. In slope support teaching, existing experimental devices mostly focus on support construction technology and lack dynamic demonstration of the support-soil interaction mechanism (such as the real-time change of anchor tension with landslide deformation), which is key for students to understand the working principle and design principle of support structures.

[0025] Existing teaching methods for landslide early warning based on slope models primarily rely on traditional methods that introduce static thresholds, which cannot meet the needs of the era of intelligent monitoring. Teaching machine learning-based early warning methods based on "dynamic thresholds + automatic decision-making" faces the problem of lacking a comprehensive experimental platform integrating data acquisition, feature extraction, model training, and model validation. This invention addresses these teaching pain points by developing a slope teaching experimental system capable of simulating the dynamic process of slope landslides under multiple coupled working conditions, simultaneously considering the influence of support structures and simulating the interaction between support and soil, and suitable for verifying early warning methods.

[0026] According to Figures 1 to 3 As shown, this invention provides a teaching model for slope support.

[0027] The slope support teaching model includes a box 100, a base plate 200, a vibration mechanism 300, a lifting mechanism 400, support piles 410, a seepage mechanism, a pressurization mechanism 500, a sprinkler mechanism 600, a lattice beam 700, an anchor bolt 800, and a protective net.

[0028] The box 100 is provided with a top-open accommodating cavity 110, and a bottom plate 200 is arranged in the accommodating cavity 110, the rear end of the bottom plate 200 is higher than the front end, so that the bottom plate 200 is inclined to the horizontal plane, and the top surface of the bottom plate 200 forms a slope structure, when the soil body 10 covers the top surface of the bottom plate 200, the left and right side walls of the box 100 limit the left and right sides of the soil body 10 on the top surface of the bottom plate 200, so as to avoid the soil body 10 from flowing to the left and right sides, so as to simulate a slope structure extending in the left-right direction.

[0029] In addition, in the embodiment, in order to accurately simulate the actual situation of the slope, the middle part of the bottom plate 200 is provided with a step extending in the left-right direction, and the step is parallel to the horizontal plane, so that the top surface of the bottom plate 200 forms a two-layer slope structure.

[0030] The box 100 is constructed by using a plate material and a frame to build a slope-shaped body similar to a slope, the frame of the box 100 is selected from high-strength aluminum alloy materials, so as to ensure the stability and bearing capacity of the structure, and the side plate of the box 100 is selected from transparent PVC plates which are wear-resistant, corrosion-resistant and convenient for observation, so as to facilitate real-time observation of the change of the soil body 10 inside the slope.

[0031] The top surface of the bottom plate 200 is specially treated, the wear-resistant coating is sprayed on the surface of the bottom plate 200, and the surface roughness is controlled, so that the friction of the surface of the bottom plate is close to the friction of the soil body 10 in the actual engineering, and the authenticity of the simulation is ensured. Specifically, the surface of the bottom plate 200 is pretreated to remove impurities and oxide layers, then the wear-resistant coating is sprayed, and finally the surface of the coating is polished, so that the surface friction is controlled in a small range of error with the friction of common soil.

[0032] A layer of soil body 10 is covered on the top surface of the bottom plate 200, and the soil of the soil body 10 is determined according to the actual soil to be simulated (such as clay, sandy soil, gravel soil, etc.). The thickness of the soil body 10 is determined according to the test requirements, mainly according to the experimental parameters of the maximum static friction and the minimum dynamic friction of the soil body 10, and the thickness of the soil body 10 can be adjusted through the pre-experiment, so as to ensure that the sliding law of the soil body 10 under different thicknesses can be accurately reflected.

[0033] Supporting structure: including supporting pile 410, lattice beam 700, anchor rod 800, protective net, and one or several supporting structures can be selected according to actual experimental needs.

[0034] Support pile simulation: The lifting mechanism 400 is arranged below the step of the base plate 200, the lifting mechanism 400 is connected with a plurality of support piles 410, the support piles 410 are arranged vertically, the plurality of support piles 410 are distributed in the left-right direction, the step of the base plate 200 is provided with a plurality of avoidance holes 210 distributed in the left-right direction, the plurality of support piles 410 are correspondingly and slidingly arranged in the avoidance holes 210, the lifting mechanism 400 drives all the support piles 410 to move in the up-down direction, the lifting mechanism 400 drives all the support piles 410 to move upward and extend above the base plate 200 and extend into the soil body 10, and the lifting mechanism 400 drives all the support piles 410 to move downward and be recovered to below the base plate 200.

[0035] The support piles 410 are made of metal materials with appropriate strength, and the length is determined according to the thickness of the soil, so as to ensure that the support piles 410 can penetrate the soil body 10.

[0036] The lifting mechanism 400 is selected from a driving motor capable of accurately controlling displacement, the lifting mechanism 400 is connected with the support piles 410 through a screw rod transmission mechanism, and the lifting speed and displacement of the support piles 410 can be accurately controlled. By controlling the jacking and lowering of the support piles 410, the insertion and extraction of the support piles 410 in the soil body 10 are realized, so as to simulate the landslide comparison experiment of the slope with and without the support piles 410. When it is necessary to simulate the working condition with the support piles 410, the support piles 410 are jacked to penetrate the soil body 10 to a set height; when it is necessary to simulate the working condition without the support piles 410, the support piles 410 are lowered to below the base plate 200 and are separated from the soil body 10.

[0037] The support piles 410 for monitoring are made of a flexible plastic material, and a through hole for installing a sensor is arranged in the center. Compared with the traditional rigid material, the flexible plastic can produce more obvious deformation when the soil body 10 slides to generate lateral pressure, so as to facilitate the magnification of the deformation of the support piles 410 in the teaching model, and the students can more intuitively observe the deformation process of the support piles 410, and the difficulty of teaching understanding is reduced.

[0038] Gyroscope sensor installation: A plurality of gyroscope sensors are installed in the central through hole of the monitoring support pile 410 in the vertical direction, the installation interval of the sensors is determined according to the length of the support pile 410 and the monitoring accuracy requirement, so as to ensure that the displacement and inclination changes of the support pile 410 at different height positions can be fully captured. The gyroscope sensors are selected from types with high measurement accuracy and fast response speed, and are tightly connected with the central hole of the support pile 410 through special fixing parts, so as to avoid the influence of sensor loosening on data accuracy.

[0039] Mechanical parameter calculation principle: when the supporting pile 410 is deformed by the lateral pressure applied by the sliding soil, the gyro sensor at different positions in the vertical direction will collect the displacement value and inclination value of the position where it is located in real time. After uploading these data to the data processing module, the bending deformation, bending moment, shear force and other mechanical parameters of the supporting pile 410 can be directly calculated through the geometric relationship and the basic formula of material mechanics, without complex data conversion, simplifying the data processing process in teaching experiment, and facilitating students to quickly understand the stress and deformation law of the supporting pile 410.

[0040] Lattice beam simulation: lattice beams 700 are made of suitable materials, and the arrangement interval of the lattice beams 700 is determined according to the size of the slope model. The lattice beams 700 are connected to the frame of the slope slope body through a convenient connection mode, which can simulate the supporting effect of lattice beams 700 with different parameters on the slope, and analyze the improvement effect of the lattice beams 700 on the slope stability in combination with other monitoring data.

[0041] Anchor rod simulation: anchor rods 800 are made of high-strength materials, and the length is determined according to the thickness of the soil. One end of the anchor rod 800 is fixed to the frame at the top of the slope through an anchor, and the other end is inserted into the soil. The insertion depth can be controlled by adjusting the anchor, which simulates the fixing mode and supporting effect of the anchor rod 800 in the actual anchoring engineering.

[0042] A tension sensor is installed between the anchor rod 800 and the top nut. The tension sensor is an anchor cable meter. The anchor cable meter is selected to match the specifications of the anchor rod 800, and is connected with the anchor rod 800 and the top nut through a threaded connection to ensure accurate monitoring of the tension change of the anchor rod 800 during the stress process. The anchor cable meter is equipped with a high-precision force sensor, which can collect dynamic change data of the tension of the anchor rod 800 in real time. The data is transmitted to the data acquisition system through a signal line, and in combination with the displacement monitoring of the slope, the interaction between the anchor rod 800 and the slope can be observed in real time.

[0043] Protection net simulation: different materials and mesh specifications of the protection net are used, and the appropriate type is selected according to the experimental requirements. The protection net is connected to the frame of the slope slope body through a convenient connection mode, and is covered on the surface of the soil, which can simulate the restraining effect of the protection net on the slope soil, preventing the soil from collapsing, scattering and other situations during sliding, rainfall or vibration.

[0044] Landslide simulation system: including seismic motion simulation system, rainfall simulation system, soil seepage simulation system, slope top load simulation system.

[0045] The seismic vibration simulation system (optimized vibration conduction structure) is characterized in that the vibration mechanism 300 is arranged on the bottom plate 200, the vibration mechanism 300 comprises a vibration plate 310, an elastic member 320 and a vibrator 330, the top end of the vibration plate 310 is hinged to the bottom plate 200, the vibration plate 310 rotates around an axis extending in the left-right direction, the elastic member 320 is a spring, the two ends of the elastic member 320 are connected to the vibration plate 310 and the bottom plate 200 respectively, so that the vibration plate 310 and the bottom plate 200 can move elastically relative to each other, and the vibrator 330 is arranged on the bottom surface of the vibration plate 310, and the vibrator 330 drives the vibration plate 310 to vibrate after being started, so as to simulate the influence of an earthquake on a slope.

[0046] The vibrator 330 is selected to be a vibration motor with adjustable frequency and adjustable amplitude, the motor parameters can be adjusted through relevant control components, vibration output with different frequencies and amplitudes is realized, and different intensity working conditions of an earthquake are simulated. The vibration conduction structure of the elastic member 320 and the vibration plate 310 is adopted, a plurality of elastic members 320 are uniformly arranged between the vibration plate and the bottom plate, and the elastic member 320 is selected to be a spring with good damping performance and bearing capacity.

[0047] In some embodiments, the vibration plate 310 is arranged in a local distribution manner between the bottom plate 200 and the soil body 10, a plurality of combined structures of the vibration plate 310 and the elastic member 320 are arranged in key areas (such as middle and bottom areas of the slope which are prone to sliding) of the slope model according to the size of the slope model and experimental requirements, each combined structure of the vibration plate 310 and the elastic member 320 can be independently controlled, differential vibration simulation of different areas of the slope is realized, and the situation that different parts of the slope are subjected to uneven stress in an actual earthquake is more close to reality. The elastic member 320 plays a damping role on the one hand, reduces the interference of the vibrator 330 in operation on other parts of the device, and realizes accurate conduction of vibration energy on the other hand, so as to ensure that vibration energy is efficiently transmitted to the soil body 10 through the vibration plate 310, and the accuracy of the earthquake simulation is improved.

[0048] The soil body seepage simulation system is characterized in that a water seepage mechanism (not shown in the figure) is arranged on the bottom plate 200, a plurality of water seepage pipes are arranged between the bottom plate 200 and the soil body 10, and the water seepage mechanism is used for simulating the state of sliding caused by seepage of the soil body 10. The water seepage pipes on the bottom plate 200 are made of corrosion-resistant materials, the plurality of water seepage pipes are uniformly arranged along the length direction of the bottom plate 200, the water inlet end of the water seepage pipe is connected to an external water supply system through a pipeline, the water outlet end is arranged on the top surface of the bottom plate 200, and control components and flow monitoring components are arranged on the water outlet end, so as to control the opening and closing of the water seepage pipe and monitor the seepage flow.

[0049] The water infiltration pipe in the soil body 10 is provided with a plurality of through water infiltration holes and is buried at different depths of the soil body 10, so that water can uniformly infiltrate into the soil body 10, the seepage of the soil body 10 is simulated, and the sliding condition of the soil body 10 under different seepage flow and seepage speed can be simulated by adjusting the pressure of the water supply system and the flow of the water infiltration pipe.

[0050] The rainfall simulation system includes a spraying pipeline 610, a plurality of spray heads 620 and a plurality of rain gauges 630.

[0051] The spraying pipeline 610 and the plurality of spray heads 620 are installed on the top frame of the box body 100 and are used for simulating the influence of natural rainfall on the slope. The spraying pipeline 610 is made of corrosion-resistant material and includes a main pipeline and branch pipelines. The main pipeline is arranged along the length direction of the top frame of the slope, and the branch pipelines extend from the main pipeline to above the surface of the slope. The plurality of spray heads 620 are distributed on the spraying pipeline 610 at intervals and face the soil body 10, so that the rainfall can cover the entire surface of the slope. The spray angle of the spray head 620 can be adjusted, and different intensities of rainfall (such as light rain, moderate rain and heavy rain) can be simulated by adjusting the water spraying amount.

[0052] The water supply end of the spraying pipeline 610 is connected with an external constant-pressure water supply system. Control components and flow monitoring components are installed on the main pipeline. The opening and closing of the spraying pipeline 610 are controlled by the control components, and different rainfall intensities are simulated by adjusting the pressure of the constant-pressure water supply system and the opening degree of the control components.

[0053] The rain gauges 630 are placed at different positions (such as the top, middle and bottom of the slope) of the surface of the slope and are used for monitoring the rainfall intensity and rainfall amount in real time. The rain gauges are selected from types with measurement accuracy meeting the experimental requirements. The collected rainfall data are uploaded to a teaching platform in real time through a data signal line and a data acquisition system, so that the influence of rainfall on the stability of the slope can be analyzed.

[0054] The slope top load simulation system includes a pressurizing mechanism 500 arranged at the top of the containing cavity 110. The pressurizing mechanism 500 is arranged above the top of the slope. The pressurizing mechanism 500 includes a linear actuator and a pressurizing plate. The linear actuator is connected with the pressurizing plate and drives the pressurizing plate to move in the up-down direction. The pressurizing plate is located above the top of the slope formed by the soil body 10. The pressurizing plate is driven by the linear actuator to move downward to pressurize the top of the slope, so as to simulate the state of sliding caused by overload of the bearing capacity of the surface of the slope.

[0055] The linear actuator is driven by a servo motor capable of accurately controlling pressure and is connected with the pressurizing plate through a screw transmission mechanism. The pressurizing plate is made of a plate material with appropriate strength and can be adjusted in the pressurizing area according to the experimental requirements. The servo motor can accurately control the pressure of the pressurizing plate on the surface of the soil body 10, and the pressure control precision meets the experimental requirements.

[0056] In some embodiments, the pressurizing mechanism 500 includes a pressurizing plate and a weight, which can be a known weight of a counterweight, and the pressurizing plate is placed on the top of the slope, and then the weight is placed on the top of the pressurizing plate, and the pressurizing effect can also be achieved, which is used to simulate the state of sliding caused by overload of the surface bearing capacity of the slope.

[0057] The counterweight is placed above the pressurizing plate, and the surface pressure of the soil body 10 is adjusted by adding counterweights of different weights, which can be used to verify the pressure accuracy of the above-mentioned linear drive, and also can be used as a backup pressure increasing mode when the linear drive fails, to ensure the smooth progress of the experiment.

[0058] Data acquisition system: The slope support teaching model is equipped with a perfect data acquisition system, which includes the collection of various variables through pressure sensors, anchor cable meters, temperature sensors, humidity sensors, gyroscope sensors, pull line meters, vibration sensors, seepage meters and rain gauges, and the data is uploaded to the teaching platform to provide data support for teaching analysis.

[0059] The pressure sensor selects a type with appropriate measurement accuracy, which is installed inside the soil body (different depths and positions), below the pressurizing plate and on the side wall of the support pile 410 (different depths), respectively, for collecting the internal pressure of the soil body 10, the pressure on the surface of the soil body 10 and the contact force between the support pile 410 and the soil body 10.

[0060] The anchor cable meter collects the change data of the anchor rod traction force, which is directly transmitted to the data acquisition card through the signal line without additional conversion, and the dynamic change curve of the traction force can be displayed in real time on the teaching platform, which is convenient for students to observe the stress condition of the anchoring system.

[0061] The temperature sensor selects a type with a measurement range and accuracy that meets the experimental requirements, which is uniformly arranged inside the soil body and around the device environment, for collecting the temperature of the soil body and the environment temperature, and analyzing the influence of temperature change on the physical and mechanical properties of the soil body and the stability of the slope.

[0062] The humidity sensor selects a type with high measurement accuracy, and multiple humidity sensors are installed at different depths inside the soil body 10 for collecting the humidity change of the soil body 10 and monitoring the distribution of humidity during the seepage and rainfall process of the soil body 10.

[0063] The displacement and inclination data of the anti-slide pile collected by the gyroscope sensor are transmitted to the data processing module, and the mechanical parameters of the anti-slide pile are automatically calculated by the preset basic formula, and the calculation results are uploaded to the teaching platform in real time, which simplifies the data processing process.

[0064] The fixed end of the tension meter is installed on the fixed frame of the slope model, and the tension end is connected with the representative monitoring point (such as the top of the slope and the potential sliding area) on the surface of the soil body. When the soil body slides, the tension line will stretch or contract with the displacement of the soil body. The tension meter converts the displacement into an electrical signal through the internal coding mechanism and directly outputs the overall displacement value of the soil body without collecting indirect data such as soil pressure, avoiding complex conversion, and uploading data directly to the teaching platform, which is convenient for students to intuitively understand the sliding trend of the soil body.

[0065] The vibration sensor is selected from a type with appropriate sensitivity and measurement range. Multiple vibration sensors are installed on the vibration plate 310, the bottom plate 200, the soil body 10 inside, and the supporting pile, for collecting the vibration signals generated by the vibrator 330 and the responses of the soil body 10 and the supporting pile 410 during the vibration process, analyzing the influence of the earthquake on the slope stability and the amplification effect of the supporting pile on the vibration.

[0066] The seepage meter is used to measure the water content of the soil body. The detection end of the seepage meter is buried in the monitoring points at different depths of the soil body. The seepage meter directly outputs the water content data of the soil body by sensing the permeation characteristics of the water in the soil body without conversion through indirect parameters such as soil moisture content. The data is uploaded to the teaching platform in real time, which can intuitively reflect the correlation between the water content of the soil body and the slope sliding, simplify the monitoring process, and improve the efficiency of teaching experiments.

[0067] The rainfall intensity and rainfall data collected by the rain gauge 630 are transmitted to the data acquisition card through the data signal line and processed and uploaded together with the data of other sensors.

[0068] The data acquisition system includes a data acquisition card and a data processing module. The analog signals collected by the sensors are converted into digital signals by the data acquisition card and then transmitted to the data processing module for filtering, amplification, analysis, and other processing. Then the processed data is uploaded to the teaching software through the wireless communication module. The teaching software can display, store, query, and analyze the data in real time, which is convenient for teachers and students to conduct teaching research.

[0069] The slope support teaching model can simulate the coupling of multiple working conditions, including earthquake action (optimized vibration conduction structure), soil seepage, surface bearing capacity overload, and rainfall action. It can also simulate the slope state with or without supporting structures (supporting piles, lattice beams, anchoring, and protective nets). It can comprehensively and realistically restore the stress, deformation, and landslide process of the mountain slope, providing an intuitive and lively experimental scene for civil engineering professional teaching, helping students better understand the knowledge related to slope protection and landslide mechanism.

[0070] The optimized vibration simulation system adopts a local distribution structure of the shock absorber 330, the elastic member 320 and the vibration plate 310 in cooperation, which can not only reduce vibration interference, but also realize accurate conduction of vibration energy, and can also realize differential vibration simulation on different regions of the slope, and improve the precision of the simulation of the seismic working condition.

[0071] The slope support teaching model is equipped with a perfect data acquisition system, which can accurately acquire various parameters of the slope under different working conditions, such as pressure, temperature, humidity, displacement, vibration, rainfall, support pile deformation, anchor rod tension and the like, and upload the data to the teaching platform in real time, so as to facilitate teachers and students to compare and analyze the data, summarize the law, provide reliable experimental data support for scientific research, and promote the innovation and development of slope protection technology.

[0072] The components are designed to have good adjustability and replaceability, such as adjustable frequency and amplitude of the vibrator 330, adjustable rainfall intensity, replaceable types and thickness of the soil body 10, adjustable types and parameters of the support structure, etc., which can meet the needs of different teaching experiments and scientific research, and have a wide range of applications.

[0073] The structure of the present application is reasonable, and high-strength, wear-resistant and corrosion-resistant materials are used to ensure the stability and service life of the device, and the device is easy to install, disassemble and maintain, and the operation is simple and easy to understand, so it is convenient to promote the use in university laboratories and scientific research institutions.

[0074] Referring to Figure 4 The present application provides a kind of slope support teaching model experimental method, applied to the slope support teaching model described in the above embodiment, experimental method includes the following steps.

[0075] Step S100, according to the experimental requirements to obtain soil body 10, cover soil body 10 on the top surface of the bottom plate 200.

[0076] According to the experimental requirements to select suitable soil, evenly lay the soil on the slope bottom plate 200, control the thickness of the soil body 10 to reach the set value, check the connection of each system, ensure the normal operation of the equipment, calibrate various sensors, and ensure the accuracy of data acquisition.

[0077] Step S200, control the lifting mechanism 400 to drive all support piles 410 to be recycled to the bottom of the bottom plate 200.

[0078] Control the lifting mechanism 400 to drive all support piles 410 to be lowered to the bottom of the bottom plate 200, and separate from the soil body 10, to simulate the situation of the slope without support pile working condition.

[0079] Step S300, the real-time parameters of the soil body 10 are obtained by performing a simulation experiment, and the simulation experiment includes: controlling the vibration mechanism 300 to vibrate the soil body 10, controlling the water infiltration mechanism to infiltrate water into the soil body 10, controlling the pressurizing mechanism 500 to pressurize the soil body 10, and controlling the spraying mechanism 600 to spray water on the surface of the soil body 10; and the real-time parameters include temperature, humidity, pressure, vibration amount, displacement amount and surface rainfall amount.

[0080] According to the experimental scheme, the vibration simulation (the vibration mechanism 300 is set to a specific frequency and amplitude), the soil body 10 seepage simulation (the water infiltration mechanism is set to a specific seepage flow), the surface bearing capacity simulation system (the pressurizing mechanism 500 is added to apply a specific pressure), and the slope surface rainfall system (the spraying mechanism 600 is set to a specific rainfall intensity) are started respectively, and the data acquisition system is started at the same time, the parameters of the soil body under different working conditions are collected, and are uploaded to the teaching platform.

[0081] Step S400, the lifting mechanism 400 drives all the supporting piles 410 to extend above the bottom plate 200, and the simulation experiment is repeatedly performed to obtain the real-time parameters.

[0082] The lifting mechanism 400 drives the supporting piles 410 to penetrate the soil body 10 to a set height, and the supporting effect of the supporting piles 410 on the slope is simulated. The experimental steps of the simulation experiment are repeated, the corresponding experimental data are collected, and the data under the condition without supporting piles are compared and analyzed.

[0083] Step S500, the lattice beam 700 is arranged on the surface of the soil body 10, and the simulation experiment is repeatedly performed to obtain the real-time parameters.

[0084] Step S600, the anchor rod 800 is arranged between the lattice beam 700 and the soil body 10, and the simulation experiment is repeatedly performed to obtain the real-time parameters.

[0085] Step S700, the protective net is arranged on the surface of the soil body 10, and the simulation experiment is repeatedly performed to obtain the real-time parameters.

[0086] The supplementary supporting structures such as the lattice beam 700, the anchor rod 800 and the protective net are installed respectively, the experimental steps of the simulation experiment are repeated, the working conditions of the slope under the action of different supplementary supporting structures are simulated, the experimental data are collected, and the influence of different supplementary supporting structures on the stability of the slope is analyzed.

[0087] Referring to Figure 5 The present application provides a kind of early warning control method of slope support teaching model, it is applied to the slope support teaching model described in above embodiment, and early warning control method includes the following steps.

[0088] Step S800, the real-time parameters are obtained by controlling the simulation experiment working condition of the slope support teaching model.

[0089] Different experimental conditions of the slope support teaching model are simulated, and real-time parameters are obtained in each simulation process, so as to obtain real-time parameters of the slope support teaching model in response to different experimental conditions.

[0090] In step S810, the real-time parameters are denoised to obtain experimental data, and an adaptive window is established according to the experimental data.

[0091] The obviously abnormal data in the real-time parameters are eliminated to obtain experimental data conforming to the actual situation, and then an adaptive window mechanism based on the data fluctuation degree is adopted to automatically adjust the size of the sliding window according to the data change rate of the experimental data, so as to determine an accurate adaptive window to observe the situation of the experimental data.

[0092] In step S820, a dynamic threshold is calculated according to the adaptive window and the experimental data, and the dynamic threshold is used to divide the early warning levels.

[0093] The dynamic threshold corresponding to each monitoring parameter is calculated from the adaptive window and the corresponding experimental data, and then the dynamic threshold is used to divide the early warning levels, for example, the early warning levels are divided into four levels: blue-normal (all characteristic indexes are within the normal range), yellow-attention (a single index slightly exceeds the threshold), orange-alert (multiple indexes exceed the threshold or a single index significantly exceeds the threshold), and red-danger (multiple indexes seriously exceed the threshold, and the landslide risk is extremely high).

[0094] In step S830, a training database is established according to the experimental conditions, the experimental data, and the early warning levels.

[0095] The experimental data obtained from each experimental condition and the early warning levels determined from the experiment are integrated to establish a training database, so that the database contains experimental data of multiple experimental conditions and corresponding early warning levels.

[0096] In step S840, a prediction model is obtained by training a machine learning algorithm based on the training database.

[0097] In step S850, the prediction model is used to perform early warning on the slope support teaching model.

[0098] The prediction model is trained based on the training database, and when a student uses the slope support teaching model to perform a simulation experiment, the prediction model can perform real-time early warning based on the parameters measured by each sensor in the slope support teaching model, so as to obtain an accurate early warning level.

[0099] In some embodiments, step S810 includes the following steps.

[0100] In step S811, a data change rate is obtained according to the experimental data, and a high fluctuation period, a medium fluctuation period, or a stable period is determined according to the data change rate.

[0101] Step S812, determining the window size according to the high fluctuation period, the medium fluctuation period or the stable period, and setting the adaptive window according to the window size.

[0102] Step S813, color marking the experimental data of the adaptive window according to the high fluctuation period, the medium fluctuation period or the stable period.

[0103] The system adopts an adaptive window mechanism based on the degree of data fluctuation, and automatically adjusts the size of the sliding window according to the data change rate. In the high fluctuation period (data change rate > 10% / min), a 1-minute window is used, in the medium fluctuation period (data change rate 1%~10% / min), a 3-minute window is used, and in the stable period (data change rate <1% / min), a 10-minute window is used, to ensure that the characteristics of the slope change can be accurately captured in different states. The student enters the adaptive window module, and the system marks the characteristic data of the high fluctuation period, the medium fluctuation period and the stable period with different colors according to the characteristic change rate calculation result, which is convenient for students to observe.

[0104] In some embodiments, step S820 includes the following steps.

[0105] Step S821, obtaining historical data and real-time data according to the adaptive window and the experimental data, and obtaining data change trend according to the historical data and the real-time data.

[0106] Step S822, calculating the dynamic threshold value by using the percentile method combined with time weighting and trend prediction according to the historical data and the real-time data.

[0107] Based on the historical data and the real-time change trend in the sliding window, the dynamic threshold value of each feature is calculated. The percentile method is used, combined with time weighting and trend prediction, to calculate the single feature threshold value and the multi-feature combined threshold value, to realize more accurate early warning judgment. The student enters the dynamic threshold value calculation module, and the system will calculate the dynamic threshold value of different features and display it in the form of a dashed line on the feature extraction result graph.

[0108] In some embodiments, step S850 includes the following steps.

[0109] Step S851, controlling the slope support teaching model to simulate any working condition to obtain actual parameters.

[0110] Step S852, obtaining the early warning level by using the prediction model according to the actual parameters.

[0111] In a specific experimental scenario, the student starts the model verification module. The prediction model of the system will analyze the feature data in real time when the experiment is performed, and will issue an alarm when the early warning condition is met.

[0112] In some embodiments, the early warning control method further includes the following steps.

[0113] Step S900, set a fixed threshold, and set a warning condition according to the experimental data and the fixed threshold.

[0114] Step S910, according to the warning condition, the slope support teaching model is warned.

[0115] The system allows users to select static threshold method and machine learning method for landslide warning at the same time, which is convenient for students to compare the warning effect of two different methods.

[0116] Teaching software: mainly used for four modules of experimental condition setting, real-time display of experimental results, experimental data analysis and landslide warning. The modules cooperate with each other to realize the intelligent control and analysis of the whole process of slope simulation experiment.

[0117] 1. Experimental condition setting module Multi-parameter cooperative control system: adopts hierarchical control architecture, users input experimental parameters through graphical interface, software system converts parameter instructions into digital signals, and transmits them to corresponding controllers through communication interface to realize accurate control of each subsystem of mechanical structure.

[0118] Landslide working condition parameter control: seismic frequency and amplitude can be selected from 5 levels, frequency range 0.1-5Hz, amplitude range 0-10mm. Seepage flow, 5 levels can be selected, flow range 0-500ml / min, flow is 0, seepage is not considered. Slope top load pressure, 5 adjustable, pressure range 0-50kPa, pressure is 0, slope top pressure is not considered. Rainfall intensity is divided into 5 adjustable levels (light rain 2mm / h, medium rain 5mm / h, heavy rain 10mm / h, heavy rain 20mm / h, heavy rain 30mm / h, intensity is 0, rainfall is not considered), through accurate control of water supply system pressure and electromagnetic valve opening, accurate simulation of different rainfall intensity is realized.

[0119] Support structure parameter control: support pile height is continuously adjustable (0-300mm, height is 0, support pile is not considered), support pile spacing is divided into 3 levels (dense 100mm, medium spacing 150mm, sparse 200mm), anchor rod anchoring length is adjustable (50-200mm), the geometric parameters of each support structure are accurately controlled by servo motor.

[0120] Multi-condition coupling control algorithm: time sequence control algorithm is adopted, the starting time, duration and intensity change curve of different working conditions can be set, the accurate simulation of complex working conditions such as earthquake-rainfall coupling, seepage-load coupling is realized, and the slope instability process under the action of multiple factors in actual engineering is more truly restored.

[0121] 2. Real-time display module of experimental results Slope digital twin model construction: A three-dimensional visualization technology is used to construct the slope digital twin model, which accurately reflects the geometric structure of the actual device, the layout position of the sensors, and the configuration of the supporting structure. The model supports interactive operations such as 360-degree rotation, scaling, and sectioning, allowing users to intuitively observe the internal structure of the slope and the distribution of sensors.

[0122] Real-time multi-sensor data display system: The system supports the simultaneous display of real-time data from eight types of sensors, including gyroscopic sensors (support pile displacement and inclination), anchor cable meters (anchor rod tension), line meters (soil displacement), seepage meters (soil moisture content), pressure sensors (soil pressure), temperature sensors (environmental and soil temperature), vibration sensors (vibration response), and rain gauges (rainfall intensity). Users can select the sensor data they want to display, and the data is displayed in real time as a time-value curve graph.

[0123] Data visualization optimization algorithm: An adaptive data compression algorithm is used to optimize the real-time display of large amounts of data while ensuring data accuracy. For high-frequency sampling data, the system automatically performs sliding average processing to eliminate noise interference; for multiple groups of similar sensor data, the system supports simultaneous display in the same chart and uses different colors and line types to distinguish them, making it easier for students to compare and observe.

[0124] Visual-data linkage analysis function: The system links real-time sensor data with physical changes in the slope and displays them together. When an abnormal fluctuation occurs in a sensor data, the corresponding location in the digital twin model is automatically highlighted, guiding students to link visual observation results with sensor monitoring data and deepen their understanding of the dynamic process of slope landslide.

[0125] 3. Experimental data analysis module Historical data management system: Time series databases are used to store experimental data, supporting multi-dimensional retrieval by time, sensor type, and experimental conditions. The data storage format is standardized, containing complete information such as timestamp, sensor number, value, and working condition parameters, ensuring data integrity and traceability.

[0126] Multi-condition comparison analysis algorithm: The system has built-in multiple data analysis algorithms, supporting the comparison and analysis of various sensor data under different working conditions. For example, students can select experimental data under different seismic levels, and the system automatically extracts data at key time nodes, generates comparison charts, and calculates statistical indicators (mean, peak, and change rate), quantitatively analyzing the impact of seismic motion on slope stability.

[0127] Supporting structure effect evaluation system: By comparing and analyzing the experimental data before and after setting the supporting structure, the system can automatically calculate the supporting effect evaluation indexes, including soil displacement reduction rate, stress redistribution coefficient, supporting structure bearing ratio, etc., and generate a supporting effect evaluation report, helping students understand the mechanism and application conditions of different supporting structures.

[0128] Parameter sensitivity analysis function: The system can perform sensitivity analysis on different supporting structure parameters (such as supporting pile spacing, anchor rod anchoring length), and calculate the influence weight of each parameter on slope stability by comparing the experimental results under different parameter combinations, providing quantitative basis for supporting scheme optimization.

[0129] 4. Landslide warning module (1) Traditional static threshold warning system: Users can pre-set multiple warning trigger conditions, including selecting sensor types (any combination of 8 sensors), sensor numbers (supporting multiple sensor joint warning), trigger types (absolute value threshold, rate of change threshold, cumulative value threshold), and cycle trigger notification interval (adjustable from 1 minute to 60 minutes). The system supports composite condition warning, such as "soil displacement rate ≥ 1 mm / min and anchor rod tension increment ≥ 50 N" simultaneously.

[0130] (2) Intelligent warning system based on machine learning.

[0131] The system provides machine learning landslide warning method full-process step-by-step learning function for landslide warning feature extraction, model training, and model verification.

[0132] ① Data preprocessing module Wavelet denoising algorithm: An improved wavelet denoising method is used to process abnormal values and noise in sensor data. The algorithm first decomposes the signal into different frequency components through wavelet decomposition, then removes noise components using an adaptive threshold method, and finally obtains the purified data through wavelet reconstruction, improving data quality and warning accuracy. Students select specific historical experimental data for wavelet denoising, and the system automatically generates a comparison chart of feature extraction results before and after denoising, making it easy for students to observe the data denoising processing effect of different wavelet bases.

[0133] Feature extraction algorithm: An 8-dimensional feature extraction program is used to extract deformation features (displacement rate), mechanical features (soil pressure rate of change, anchor rod tension rate of change, pile strain rate), hydrological features (rainfall infiltration coefficient), and time series features (3-hour displacement cumulative amount, 24-hour stress variation coefficient, rainfall lag effect quantitative index) from multi-sensor data. After feature extraction of the denoised data, the system supports displaying multiple sets of feature data in the same chart, and uses different colors and line types to distinguish them, making it easy for students to compare and observe the warning effect of different features.

[0134] ②Dynamic label marking module Adaptive window mechanism: The system adopts an adaptive window mechanism based on the degree of data fluctuation, automatically adjusting the size of the sliding window according to the data change rate. In the high fluctuation period (data change rate > 10% / min), a 1-minute window is used, in the medium fluctuation period (data change rate 1%~10% / min), a 3-minute window is used, and in the stable period (data change rate < 1% / min), a 10-minute window is used, ensuring that the characteristics of the slope change can be accurately captured in different states. Students enter the adaptive window module, and the system will mark the feature data of high fluctuation period, medium fluctuation period and stable period with different colors according to the characteristic change rate calculation results, which is convenient for students to observe.

[0135] Dynamic threshold calculation: Based on the historical data and real-time change trend in the sliding window, the dynamic threshold of each feature is calculated. The percentile method is used to calculate the single feature threshold and multi-feature combined threshold, combining time weighting and trend prediction, to achieve more accurate early warning judgment. Students enter the dynamic threshold calculation module, and the system will calculate the dynamic threshold of different features and display them in the form of dashed lines on the feature extraction result graph.

[0136] Warning level marking: According to the dynamic threshold calculation results, the system automatically marks four levels of warning levels: blue - normal (all feature indicators are within the normal range), yellow - attention (single indicator slightly exceeds threshold), orange - alert (multiple indicators exceed threshold or single indicator significantly exceeds threshold), red - danger (multiple indicators seriously exceed threshold, landslide risk is extremely high). After students enter the warning level marking module, according to the selected percentile threshold, the system marks the feature abnormal data with yellow, orange and red according to the warning level, and displays the reason for level division, so that students can observe the influence of percentile threshold adjustment on the false alarm rate.

[0137] ③Model training module Built-in training database: The system has a built-in training database containing 1000 historical experimental data, covering various combination scenarios of five working conditions (earthquake / rainfall / load / seepage / support). The data comes from simulation analysis and model experiment results, the input layer is an 8-dimensional feature vector, and the output layer is a four-level warning label, ensuring the sufficiency and representativeness of model training.

[0138] Machine learning algorithm training: Random Forest + Long Short Term Memory Network (LSTM) dual model architecture is used to train the training data. Random Forest algorithm is used to process nonlinear feature interaction (such as the coupling effect of soil pressure change rate and rainfall infiltration coefficient), and the feature importance is evaluated by out-of-bag error (OOB). Key indicators are automatically selected. LSTM network is used to capture time-dependent relationships and predict trends for dynamic features such as 3-hour displacement accumulation and 24-hour stress variation coefficient. The gating mechanism (forget gate, input gate, output gate) remembers the long-term landslide evolution law. Students enter the model training module, and the system displays real-time monitoring pictures of model training, including the change curve of OOB error of random forest with the number of trees, the feature importance ranking bar chart, and the loss function descent process (MSE curve) of LSTM training set / verification set. To save students' waiting time, the system provides a pre-trained model for direct calling.

[0139] ④Model verification module In a specific experimental scenario, students start the model verification module. The system will analyze the feature data in real time during the experiment, and issue an alarm when the warning condition is met. The system allows users to select both static threshold method and machine learning method for landslide warning, making it easy for students to compare the warning effects of the two different methods.

[0140] The embodiments of the application are described in detail above with reference to the drawings, but the application is not limited to the above embodiments. Within the knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the application.

Claims

1. A teaching model for slope protection, characterized in that, include: The box has a receiving cavity at the top; A base plate is inclinedly disposed in the receiving cavity, the base plate is provided with a plurality of through-holes, and the top surface of the base plate is used to cover the soil. A vibration mechanism is provided on the base plate, and the vibration mechanism is used to vibrate the soil. A lifting mechanism is located below the base plate; Multiple support piles are provided, each of which is correspondingly installed in a plurality of clearance holes. The lifting mechanism is connected to the multiple support piles and drives the multiple support piles to move in the up and down direction so that the support piles extend above the bottom plate or retract below the bottom plate. A water infiltration mechanism is provided on the base plate, and the water infiltration mechanism is used to infiltrate water into the interior of the soil. A pressurizing mechanism is located at the top of the receiving cavity, and the pressurizing mechanism is used to pressurize the soil. A spraying mechanism is located at the top of the receiving cavity, and the spraying mechanism is used to spray water onto the surface of the soil.

2. The slope support teaching model according to claim 1, characterized in that, The slope protection teaching model also includes: A gyroscope sensor is installed on the support pile, and the gyroscope sensor measures the displacement and tilt angle of the support pile; A string meter is installed inside the soil mass, and the string meter measures the displacement of the soil mass. A permeameter is installed inside the soil mass to measure the water content of the soil mass; Multiple pressure sensors are respectively installed inside the soil, below the pressurization mechanism, and on the side wall of the support pile; Multiple temperature sensors are respectively located inside the soil and around the periphery of the receiving cavity; Multiple vibration sensors are respectively installed in the base plate, inside the soil, and in the support piles; A rain gauge is installed in the receiving cavity and is located below the sprinkler mechanism. The rain gauge is used to measure the surface precipitation of the soil.

3. A teaching model experimental method for slope support, characterized in that, The experimental method, applied to the slope support teaching model as described in any one of claims 1 to 2, includes: The soil is obtained according to the experimental requirements, and the soil is used to cover the top surface of the base plate; Control the lifting mechanism to retract all the support piles to below the base plate; The simulation experiment includes: controlling the vibration mechanism to vibrate the soil, controlling the seepage mechanism to seep water into the soil, controlling the pressurization mechanism to pressurize the soil, and controlling the spraying mechanism to spray water onto the surface of the soil. The lifting mechanism is controlled to drive all the support piles to extend above the base plate, and the simulation experiment is repeated to obtain the real-time parameters.

4. The teaching model experimental method for slope support according to claim 3, characterized in that, The experimental method also includes: A grid beam is installed on the surface of the soil, the simulation experiment is repeated, and the real-time parameters are obtained. An anchor rod is installed between the lattice beam and the soil. A tension sensor is installed on the anchor rod to measure the traction force of the anchor rod. The simulation experiment is repeated to obtain the real-time parameters. A protective net is set on the surface of the soil, the simulation experiment is repeated, and the real-time parameters are obtained.

5. The teaching model experimental method for slope support according to claim 3, characterized in that, The real-time parameters include the displacement and inclination angle of the support piles, the displacement of the soil, water content, pressure, temperature, vibration, and surface precipitation.

6. A method for early warning and control of a teaching model for slope support, characterized in that, The early warning and control method, applied to the slope support teaching model as described in any one of claims 1 to 2, includes: The slope support teaching model is controlled to simulate experimental conditions and obtain real-time parameters. The real-time parameters are denoised to obtain experimental data, and an adaptive window is established based on the experimental data. A dynamic threshold is calculated based on the adaptive window and the experimental data, and the warning level is divided according to the dynamic threshold. A training database is established based on the experimental conditions, the experimental data, and the warning level. A prediction model is obtained by training a machine learning algorithm based on the training database. The prediction model is used to provide early warning for the slope support teaching model.

7. The early warning and control method for the slope support teaching model according to claim 6, characterized in that, The step of establishing an adaptive window based on the experimental data includes: The data change rate is obtained based on the experimental data, and the high fluctuation period, medium fluctuation period or stable period is determined based on the data change rate. The window size is determined based on the high volatility period, the medium volatility period, or the stable period, and the adaptive window is set based on the window size; The experimental data of the adaptive window are color-coded according to the high fluctuation period, the medium fluctuation period, or the stable period.

8. The early warning and control method for the slope support teaching model according to claim 6, characterized in that, The calculation of the dynamic threshold based on the adaptive window and the experimental data includes: Historical data and real-time data are obtained based on the adaptive window and the experimental data, and the data change trend is obtained based on the historical data and the real-time data; The dynamic threshold is calculated using the percentile method, combined with time weighting and trend prediction, based on the historical data and the real-time data.

9. The early warning and control method for the slope support teaching model according to claim 6, characterized in that, The early warning and control method further includes: Set a fixed threshold, and set early warning conditions based on the experimental data and the fixed threshold; The slope support teaching model is given an early warning based on the aforementioned warning conditions.

10. The early warning and control method for the slope support teaching model according to claim 6, characterized in that, The method of using the prediction model to provide early warning for the slope support teaching model includes: The slope support teaching model is controlled to simulate arbitrary working conditions and obtain actual parameters; The warning level is obtained using the prediction model based on the actual parameters.

Images