Physical model experiment system for structure-karst-unloading coupling degradation high and steep slope collapse

Abstract

The invention discloses a structure-karst-unloading coupling degradation high and steep slope collapse physical model experiment system, and relates to the field of slope collapse risk management, and the system comprises a construction module which is used for constructing a high and steep slope model containing a karst network; the monitoring module is used for synchronously capturing micron-order opening and micro-fracture evolution information of an internal structural surface of the model and stress-seepage field distribution data; according to the method, the model is guaranteed to be highly matched with the actual geological environment through dynamic calibration of the adaptation degree, micron-sized synchronous capture of internal structure evolution, stress distribution and seepage field data of the model is achieved, and the monitoring precision is improved through multi-device collaborative layout and data credibility verification; and cooperatively applying tectonic stress, karst seepage and graded unloading according to a real time sequence, and dynamically adapting to the quasi-steady-state requirements of the stress-seepage field of the model.

Description

Technical Field

[0001] This invention relates to the field of slope collapse risk management technology, specifically to an experimental system for a physical model of collapse on steep, high slopes with coupled structural-karst-unloading deterioration. Background Technology

[0002] High and steep slope collapses are geological disasters in which rock and soil masses become unstable and fall under the action of gravity, influenced by the terrain slope and the properties of the rock and soil, and triggered by external forces such as precipitation and earthquakes. They occur suddenly and have a strong impact, which can easily cause casualties, damage to engineering facilities and traffic interruption, posing a serious threat to the lives and production of residents in mountainous areas.

[0003] The invention patent application with application number 201110200855.5 discloses a physical model experimental system for large-scale landslides under the influence of multiple factors. This application aims to solve the problems that "the sliding bed shape is simple, mostly a straight plate in one section, while the actual sliding bed surface is mostly a concave surface, so there are large errors in the geometric boundary conditions; the operation precision of traditional experimental equipment is not enough, and most of the mechanical transmission parts are manually operated, resulting in large experimental errors, so it is mostly limited to qualitative research and the degree of quantitative research is insufficient; traditional model experiments consider only one factor, making it difficult to carry out research on the multi-factor landslide disaster mechanism".

[0004] However, current collapse physics model experiments have three major flaws: The true structure of karst conduits is missing: the manual pre-embedding using gypsum / silicone cannot reproduce the multi-level branching, heterogeneity, and anisotropy of the natural karst network; Monitoring blind spots: The lack of information on the micron-level opening and micro-fracture of structural surfaces leads to a lag in the identification of pre-collapse signs; Results cannot be extrapolated: the experiment is disconnected from the field and numerical simulation, the critical criteria are not verified in both directions, and the false alarm rate is high.

[0005] To address this, we propose an experimental physical model system for the collapse of steep, deteriorated slopes coupled with tectonic-karst-unloading parameters. Summary of the Invention

[0006] To address the aforementioned shortcomings of existing technologies, this invention provides a physical model experimental system for the collapse of steep, high-slope slopes with coupled structural-karst-unloading effects, which can effectively solve the problems of existing technologies.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions; This invention discloses an experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading processes, comprising: The system comprises the following modules: a construction module for building a model of a steep slope containing a karst network; a monitoring module for simultaneously capturing micron-level opening of structural surfaces, microfracture evolution information, and stress-seepage field distribution data within the model; a loading module for applying synergistic loading of tectonic stress, karst seepage, and staged unloading to the model to simulate the real deterioration environment of a steep slope; an interaction module for establishing a two-way interaction channel between experimental data, field measured data, and numerical simulation data; an identification module for acquiring output data from the monitoring module and data within the interaction channel of the interaction module, and extracting pre-collapse characteristic information based on the acquired data; and a verification and feedback module for performing two-way verification and calibration of the collapse critical criterion based on the linkage data from experiments, field observations, and numerical simulations, and simultaneously feeding back the calibration results to the user. The construction module is interconnected with a monitoring module via a wireless network. The monitoring module is interconnected with a loading module and an interaction module via a wireless network. The monitoring module and the interaction module are interconnected with an identification module via a wireless network. The identification module is interconnected with a verification and feedback module via a wireless network. In the steep slope model, the karst network is constructed by referencing the actual development morphology and distribution pattern of natural karst through in-situ dynamic adjustment of forming parameters. This means that it includes a multi-level branching, heterogeneous and anisotropic karst conduit network.

[0008] Furthermore, the in-situ dynamic control forming parameters of the karst network include the number of branch levels, the mean equivalent pore size, the coefficient of variation of pore size, the connectivity probability, and the spatial distribution entropy. These forming parameters are dynamically calibrated using a natural karst fit model. ; In the formula: The compatibility between karst networks and natural karst; These are the weighting coefficients; For the number of branch levels; The average number of branching levels of the natural karst in the target area; For aperture adaptation attenuation coefficient; This is the average equivalent aperture. The average equivalent pore size of the natural karst in the target area; The pore size variation coefficient; The pore size variation coefficient of the natural karst in the target area; The probability of connectivity; It represents the spatial distribution entropy; The spatial distribution entropy of natural karst in the target area; Real-time calculation of fit during regulation ,like If the value is less than the preset adaptation threshold, the molding parameters are adjusted sequentially according to their weight priority until... Once the threshold requirement is met, the actual morphology and distribution pattern of the karst network are verified by micro-CT scanning after the formation process.

[0009] Furthermore, the monitoring module integrates distributed fiber optic sensors, miniature pore water pressure gauges, high-precision strain gauges, and high-speed cameras, with each device deployed in a three-dimensional grid with preset location encryption. The data capture process is based on synchronization with a unified time base; Simultaneously capture data and perform trustworthiness verification: ; In the formula: To ensure the overall reliability of the monitoring data; The number of similar devices participating in the verification; Let be the accuracy weight of the i-th device; This represents the monitoring data value of the i-th device; This is the average value of monitoring data for similar equipment. The inherent measurement error of the i-th device; This is the time deviation influence coefficient; The sampling time deviation of the i-th device; Among them, if If the data is less than the preset confidence threshold, the abnormal data will be removed and the corresponding identification process will be refreshed and restarted.

[0010] Furthermore, the collaborative loading in the loading module conforms to: The loading is performed sequentially from tectonic stress preloading to steady-state application of karst seepage to staged unloading, where the coupling loading parameters are: ; In the formula: The coupled loading total stress of the model at time t; This represents the initial preload value of the structural stress. , These are the seepage-stress coupling coefficient and the unloading-stress attenuation coefficient. Let be the karst seepage velocity at time t; The cumulative unloading strain at time t; The additional stress caused by seepage at time t; The grade difference of the graded unloading is based on The growth rate is dynamically adjusted so that the stress-seepage field of the model reaches a quasi-steady state after each unloading stage before the next loading stage is carried out.

[0011] Furthermore, the bidirectional interaction channel of the interaction module adopts a three-layer architecture of data standardization, spatiotemporal alignment, and feature fusion: Data standardization layer: Converts experimental data, field measured data, and numerical simulation data into a preset data format, and uses standardization processing logic based on data type to eliminate differences in units; Spatiotemporal alignment layer: temporal alignment is based on interpolation matching using a unified timestamp, while spatial alignment maps data from different sources to the same three-dimensional coordinate system through coordinate transformation; Feature fusion layer: The weighted average fusion method is used to fuse feature data from the same source. The fusion weights are dynamically allocated based on the credibility of the data source.

[0012] Furthermore, the collapse precursor feature information extracted by the identification module includes the structural surface aperture growth rate, micro-fracture density, stress mutation coefficient, and seepage velocity variation rate. The feature extraction process includes: Data preprocessing: Wavelet threshold denoising algorithm is used to smooth and denoise the monitoring data to remove abnormal peaks caused by environmental interference; Feature calculation: Opening growth rate , This represents the increment of the structural surface opening within a preset time interval. To calculate the preset time interval for the opening growth rate, Initial aperture; microfracture density , V represents the number of micro-fractures, V represents the volume of the monitoring area, and the stress mutation coefficient. , This represents the maximum stress value within a preset time period. The average stress value during the same period; the seepage velocity variation rate. , The maximum seepage velocity within a preset time period. (This refers to the average seepage velocity during the same period). Significance screening: Calculate the coefficient of variation for each feature. , The standard deviation of the eigenvalues. For the mean of the eigenvalues, select Features that exceed a preset significance threshold are considered as valid precursor features of collapse, thus forming a precursor feature set.

[0013] Furthermore, the collapse threshold criterion in the verification and feedback module is: ; In the formula: The collapse risk index at time t; The number of valid precursor features; is the weight coefficient of the j-th precursor feature; Let be the real-time value of the j-th precursor feature at time t; The critical threshold for the j-th precursor feature; The verification and feedback module uses a preset early warning threshold and calculates it in real time. A comparison is made with these data to issue an early warning.

[0014] Furthermore, in the verification and feedback module, the bidirectional verification and calibration are iteratively optimized as follows: Substitute experimental data into the criterion calculation If the misjudgment rate exceeds the preset misjudgment threshold when compared with the actual collapse condition on site, then... Adjust the weighting coefficients, where The weight is adjusted until the misjudgment rate drops below the threshold. The calibrated criterion parameters are then synchronously stored in the system's preset built-in database.

[0015] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: This invention can accurately replicate the multi-level branching, heterogeneity, and anisotropy of natural karst. Through dynamic calibration, it ensures a high degree of fit between the model and the actual geological environment, achieving micron-level synchronous capture of internal structural evolution, stress distribution, and seepage field data. Furthermore, it enhances monitoring accuracy through multi-device collaborative deployment and data reliability verification. It applies tectonic stress, karst seepage, and staged unloading in a real-time sequence, dynamically adapting to the quasi-steady-state requirements of the model's stress-seepage field. It establishes a data interaction channel between experiments, field measurements, and numerical simulations, effectively extracting key precursor features such as the structural surface opening rate and micro-fracture density. Finally, through bidirectional verification and calibration optimization of critical criteria, it effectively improves the accuracy and reliability of collapse risk assessment, providing support for early warning and prevention decisions regarding high and steep slope collapse disasters. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the structure of an experimental system for a physical model of a deteriorated steep slope collapse caused by a combination of structural-karst-unloading coupling. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] The present invention will be further described below with reference to embodiments.

[0020] Example: This embodiment presents a physical model experimental system for the collapse of a steep, high-slope slope, coupled with a structure-karst-unloading coupling mechanism. Figure 1 As shown, it includes: A building block for constructing models of steep slopes containing karst networks; In-situ dynamic control of karst network formation parameters includes branching level, mean equivalent pore size, pore size variation coefficient, connectivity probability, and spatial distribution entropy. These formation parameters are dynamically calibrated using a natural karst fit model. ; In the formula: The compatibility between karst networks and natural karst; These are the weighting coefficients; For the number of branch levels; The average number of branching levels of the natural karst in the target area; For aperture adaptation attenuation coefficient; This is the average equivalent aperture. The average equivalent pore size of the natural karst in the target area; The pore size variation coefficient; The pore size variation coefficient of the natural karst in the target area; The probability of connectivity; It represents the spatial distribution entropy; The spatial distribution entropy of natural karst in the target area; The above formula integrates five key characteristics of natural karst: branching level, mean equivalent pore size, pore size variation coefficient, connectivity probability, and spatial distribution entropy. The weight coefficients reflect the sensitivity of each characteristic to the impact of collapse, and the sum of the weights meets the normalization requirements. An exponential term is applied to attenuate the deviation between the equivalent pore size and natural karst. The fit is calculated in real time and the forming parameters are adjusted according to the weight priority. At the same time, the pore size fit attenuation coefficient is used to adapt to the concentration of the pore size distribution of natural karst. Finally, combined with micro CT scanning verification, it is ensured that the constructed karst network is highly consistent with the natural karst in terms of morphology and distribution pattern. Real-time calculation of fit during regulation ,like If the value is less than the preset adaptation threshold, the molding parameters are adjusted sequentially according to their weight priority until... Once the threshold requirement is met, the actual morphology and distribution pattern of the karst network are verified by micro-CT scanning after the formation process. in, The values ​​of all values ​​are in the range (0,1), and The sum is 1. Each weight coefficient is dynamically adjusted based on the sensitivity analysis of the impact of various characteristics of natural karst on collapse. That is, the higher the sensitivity, the larger the corresponding weight coefficient value. The preset value range is 0.01 to 0.1. When the concentration of the natural karst pore size distribution in the target area is higher... The larger the value, the lower the value. The smaller the value; The monitoring module is used to simultaneously capture micron-level openings, micro-fracture evolution information, and stress-seepage field distribution data of the internal structural surfaces of the model. The monitoring module integrates distributed fiber optic sensors, miniature pore water pressure gauges, high-precision strain gauges, and high-speed cameras. Each device is deployed using a three-dimensional grid with preset location densification. The density of the three-dimensional grid is determined based on the model scale and monitoring accuracy requirements. The sensor deployment density at key structural surfaces (including karst pipe walls and potential sliding surfaces) is 2 to 3 times the grid density. The data acquisition process is based on a unified time reference for synchronization, and uses a preset GPS timing module to provide an absolute timestamp. The sampling time deviation of each monitoring device is controlled within a preset time threshold. Simultaneously capture data and perform trustworthiness verification: ; In the formula: To ensure the overall reliability of the monitoring data; The number of similar devices participating in the verification; Let be the accuracy weight of the i-th device; This represents the monitoring data value of the i-th device; This is the average value of monitoring data for similar equipment. The inherent measurement error of the i-th device; This is the time deviation influence coefficient; The sampling time deviation of the i-th device; The above formula is based on monitoring data from multiple monitoring devices of the same type. It introduces influencing factors such as device accuracy weight, data dispersion, and sampling time deviation. After normalizing the deviation of each device's data from the mean, it combines the time deviation attenuation effect to weighted summation, and then takes the reciprocal to obtain the comprehensive reliability. It also ensures data synchronization through unified GPS time synchronization, controls the sampling time deviation within the threshold, and removes abnormal data and re-identifies it when the reliability is lower than the preset value, thus fully ensuring the reliability and effectiveness of the monitoring data. Among them, if If the data is less than the preset confidence threshold, the abnormal data will be removed and the corresponding identification process will be refreshed and restarted. The loading module is used to apply synergistic loading of tectonic stress, karst seepage and staged unloading to the model to simulate the real deterioration environment of steep slopes; Co-loading in the loading module follows the following rules: The loading is performed sequentially from tectonic stress preloading to steady-state application of karst seepage to staged unloading, where the coupling loading parameters are: ; In the formula: The coupled loading total stress of the model at time t; This represents the initial preload value of the structural stress. , These are the seepage-stress coupling coefficient and the unloading-stress attenuation coefficient. Let be the karst seepage velocity at time t; The cumulative unloading strain at time t; The additional stress caused by seepage at time t; The above formula follows the real time sequence of tectonic stress preloading, steady-state application of karst seepage, and staged unloading. Based on the initial tectonic stress converted from geological stress survey data of the target area, the seepage-stress coupling coefficient k1 reflects the strengthening effect of seepage on tectonic stress, and the unloading-stress attenuation coefficient k2 reflects the attenuation effect of unloading on total stress. The dynamic changes of real-time karst seepage velocity and cumulative unloading strain are incorporated. The additional stress of seepage is calculated by the unit weight of water and pore water pressure head. The staged unloading difference is dynamically adjusted according to the cumulative unloading strain growth rate to ensure that the stress-seepage field of the model reaches quasi-steady state after each loading stage, thereby accurately simulating the real deterioration environment under the coupling effect of tectonic-karst-unloading. The grade difference of unloading is based on The growth rate is dynamically adjusted so that the stress-seepage field of the model reaches a quasi-steady state after each unloading stage before the next loading stage is carried out. in, Calculated based on similarity ratio using geological stress survey data of the target area; The preset value range is [0.05, 0.3]. The higher the permeability coefficient of the model material, the better the connectivity of the karst network, and the more significant the strengthening effect of seepage on tectonic stress, the larger the value will be, and vice versa. The preset value range is [0.1, 0.4]. The larger the value is when the model material has more obvious elastic-plastic deformation characteristics and weaker brittleness, and the more prominent the attenuation effect of unloading on the total coupled stress, the larger the value is when the model material is more brittle, the weaker the elastic-plastic deformation, and the more gradual the attenuation effect of unloading on the total coupled stress. Determined based on the natural karst seepage rate and model similarity ratio; It is calculated from the unloading displacement and the characteristic length of the model; ,in Indicates the density of water. The pore water pressure head in the model at time t; The interaction module is used to establish a two-way interaction channel between experimental data and field measured data and numerical simulation data; The two-way interaction channel of the interaction module adopts a three-layer architecture of data standardization, spatiotemporal alignment, and feature fusion. Data standardization layer: Converts experimental data, field measured data, and numerical simulation data into a preset data format, and uses standardization processing logic based on data type to eliminate dimensional differences. Data types include mechanical data, seepage data, and geometric data. Spatiotemporal alignment layer: Temporal alignment is based on interpolation matching using a unified timestamp, while spatial alignment maps data from different sources to the same three-dimensional coordinate system through coordinate transformation, with the model coordinate system as the reference. Feature fusion layer: A weighted average fusion method is used to fuse feature data from the same source. The fusion weights are dynamically allocated based on the credibility of the data source. The communication stability of the interactive channel is monitored in real time through the data transmission success rate. If the transmission fails for a preset number of consecutive times, it will automatically switch to the preset backup communication link and record the transmission abnormality log. The identification module is used to acquire the output data of the monitoring module and the data in the interaction channel of the interaction module, and to extract the pre-collapse feature information based on the acquired data. The collapse precursor features extracted by the identification module include the structural surface aperture growth rate, micro-fracture density, stress mutation coefficient, and seepage velocity variation rate. The feature extraction process includes: Data preprocessing: Wavelet threshold denoising algorithm is used to smooth and denoise the monitoring data to remove abnormal peaks caused by environmental interference; Feature calculation: Opening growth rate , This represents the increment of the structural surface opening within a preset time interval. To calculate the preset time interval for the opening growth rate, The initial aperture was determined by averaging three consecutive monitoring data points during the initial experimental phase; microfracture density was also considered. , V represents the number of micro-fractures, V represents the volume of the monitoring area, and the stress mutation coefficient. , This represents the maximum stress value within a preset time period. The average stress value during the same period; the seepage velocity variation rate. , The maximum seepage velocity within a preset time period. (This refers to the average seepage velocity during the same period). The above formula targets key indicators of impending collapse: The opening growth rate is calculated by the ratio of the structural surface opening increment within a preset time interval to the initial opening and the time interval. The initial opening is the average of three consecutive monitoring data in the initial stage of the experiment. Microfracture density is the ratio of the number of microfractures in the monitoring area to the volume of the area. The stress mutation coefficient and the seepage velocity variation rate are calculated by the ratio of the difference between the maximum and average stress and the maximum and average seepage velocity within a preset time period to the corresponding average values. Finally, features with a significance higher than a preset threshold are selected by using the coefficient of variation to construct an effective precursor feature set and accurately capture key precursor information that is sensitive to collapse. Significance screening: Calculate the coefficient of variation for each feature. , The standard deviation of the eigenvalues. For the mean of the eigenvalues, select Features that exceed a preset significance threshold are considered as valid precursor features of collapse, thus forming a precursor feature set. The verification and feedback module is used to perform bidirectional verification and calibration of the collapse critical criterion based on the linkage data of experiments, field and numerical simulation, and simultaneously feed the calibration results back to the user terminal. The collapse threshold criterion in the verification and feedback module is: ; In the formula: The collapse risk index at time t; The number of valid precursor features; is the weight coefficient of the j-th precursor feature; Let be the real-time value of the j-th precursor feature at time t; The critical threshold for the j-th precursor feature; The above formula integrates multiple effective precursor features. The real-time collapse risk comprehensive index is obtained by weighted summation of the ratios of the real-time values ​​of each precursor feature to the corresponding critical threshold. The index is then combined with a preset warning threshold for early warning. At the same time, the index calculated from experimental data is compared with the actual collapse state on site, and the weight coefficients are iteratively adjusted according to the weight correction amount until the misjudgment rate of the criterion is reduced to below the preset threshold. The calibrated parameters are stored in the built-in database, thus realizing the two-way verification and calibration of the collapse critical criterion to improve the accuracy and on-site applicability of the criterion. The verification and feedback module determines early warning thresholds based on preset criteria and calculates them in real time. A comparison with this data will trigger an early warning. in, ∈ (0,1), Furthermore, the stronger the correlation between the precursor features and the occurrence of the collapse, the higher the sensitivity to the slope deterioration process. That is, the more significant the change of the feature before the collapse and the more sufficient the lead time, the larger its weight coefficient value, and vice versa. In the verification and feedback module, the two-way verification and calibration are iteratively optimized as follows: Substitute experimental data into the criterion calculation If the misjudgment rate exceeds the preset misjudgment threshold when compared with the actual collapse condition on site, then... Adjust the weighting coefficients, where The weight adjustment is performed until the misjudgment rate drops below the threshold, and the calibrated criterion parameters are synchronously stored in the system's preset built-in database. in, ∈[-0.2,0.2], and follows the rule that the larger the deviation between the criterion prediction value of the j-th precursor feature and the actual collapse state on site, and the higher the sensitivity of the feature to the collapse, the larger the value; and the smaller the deviation and the lower the sensitivity of the influence, the smaller the value. Among them, the karst network in the high and steep slope model is constructed by referring to the actual development form and distribution pattern of natural karst and by dynamically adjusting the forming parameters in situ, that is, a karst conduit network with multi-level branches, heterogeneity and anisotropy. The building module is connected to the monitoring module via a wireless network. The monitoring module is connected to the loading module and the interaction module via a wireless network. The monitoring module and the interaction module are connected to the identification module via a wireless network. The identification module is connected to the verification and feedback module via a wireless network.

[0021] In this embodiment, the construction module builds a high and steep slope model containing a karst network. The monitoring module runs in sequence to simultaneously capture micron-level opening of the internal structural surfaces, micro-fracture evolution information, and stress-seepage field distribution data of the model. The loading module further applies synergistic loading of tectonic stress, karst seepage, and staged unloading to the model to simulate the real deterioration environment of the high and steep slope. The interaction module simultaneously establishes a two-way interaction channel between experimental data, field measured data, and numerical simulation data. Then, the identification module obtains the output data of the monitoring module and the data in the interaction channel of the interaction module. Based on the acquired data, it extracts the pre-collapse feature information. Finally, the verification and feedback module performs two-way verification and calibration of the collapse critical criterion based on the linkage data of experiments, field, and numerical simulation, and simultaneously feeds back the calibration results to the user terminal.

[0022] In the above embodiments, the system can accurately recreate the coupled deterioration environment of high and steep slope structure-karst-unloading, accurately capture key internal data, improve monitoring reliability through multi-source data collaborative verification, efficiently identify collapse precursor characteristics, and optimize early warning criteria through bidirectional verification and calibration, thereby greatly improving the accuracy and timeliness of collapse early warning and providing scientific and reliable technical support for slope disaster prevention and control.

[0023] Referring to the system in the above embodiments, an application example of the system is shown: Taking a steep slope in a karst development area in South China as the research object, this experimental system was used to carry out experimental research on landslide early warning. The specific process is as follows: During model construction, the model parameters such as the number of branches and the average equivalent pore size were dynamically adjusted based on the actual development characteristics of natural karst in the target area. The fit was calculated in real time, and the fit was finally achieved to 0.90 (meeting the preset threshold of 0.85). Micro-CT scan confirmed that the multi-level branching, heterogeneity and anisotropy characteristics of the model karst network were consistent with those of natural karst.

[0024] The monitoring module is equipped with distributed fiber optic sensors, high-precision strain gauges, and other devices arranged in a three-dimensional grid. The sensor density is increased to twice the grid density in key areas such as karst pipe walls and potential sliding surfaces. GPS timing is used to ensure that the sampling time deviation meets the requirements. After synchronously capturing relevant data, the overall reliability of the monitoring data is calculated to be 0.94. After removing a small number of abnormal data, the data is used for subsequent analysis.

[0025] The loading phase is carried out in a coordinated manner according to a preset time sequence: first, the tectonic stress preloading is completed, and after the karst seepage reaches a steady state, the staged unloading is carried out simultaneously. Based on the model material characteristics and regional geological data, parameters such as the seepage-stress coupling coefficient and the unloading-stress attenuation coefficient are determined, and the unloading level difference is dynamically adjusted to ensure that the stress-seepage field reaches a quasi-steady state after each unloading stage.

[0026] The interaction module processes data through a three-layer architecture: first, it standardizes experimental data, field measured data, and numerical simulation data to eliminate differences in dimensions; then, it maps various types of data to the same coordinate system through spatiotemporal alignment; finally, it completes feature fusion by assigning weights according to data credibility. In the experiment, it switched to the backup link due to one consecutive transmission failure, ensuring the stability of data interaction.

[0027] The identification module first denoises the monitoring data, then calculates four features, including the structural surface opening growth rate and micro-fracture density. Through the coefficient of variation screening, two effective precursor features are determined to form a feature set.

[0028] During the verification and feedback phase, weights were assigned based on the correlation between each effective feature and the collapse (the sum of the weights was 1), and a real-time comprehensive collapse risk index was calculated. After comparison with the actual on-site conditions, the initial judgment error rate was 7%. After adjusting the weight correction amount (the correction amount was in the range of -0.2 to 0.2), the error rate was reduced to 4.2% after iterative optimization. The calibration parameters were stored in the system, and the early warning results were fed back to the user terminal, providing technical support for slope collapse prevention in this area.

[0029] In summary, the system in the above embodiments can accurately replicate the multi-level branching, heterogeneity, and anisotropy characteristics of natural karst. Dynamic calibration ensures a high degree of fit between the model and the actual geological environment, achieving micron-level synchronous capture of internal structural evolution, stress distribution, and seepage field data. Multi-device collaborative deployment and data reliability verification enhance monitoring accuracy. Tectonic stress, karst seepage, and staged unloading are applied collaboratively according to the actual time sequence, dynamically adapting to the quasi-steady-state requirements of the model's stress-seepage field. It establishes a data interaction channel between experiments, field measurements, and numerical simulations, effectively extracting key precursor features such as the structural surface opening rate and micro-fracture density. Finally, bidirectional verification and calibration of critical criteria optimize early warning parameters, effectively improving the accuracy and reliability of collapse risk assessment and providing support for early warning and prevention decisions regarding high and steep slope collapse disasters.

[0030] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading processes, characterized in that... include: A building module for constructing models of steep slopes containing karst networks; The monitoring module is used to simultaneously capture micron-level openings, micro-fracture evolution information, and stress-seepage field distribution data of the internal structural surfaces of the model. The loading module is used to apply synergistic loading of tectonic stress, karst seepage and staged unloading to the model to simulate the real deterioration environment of steep slopes; The interaction module is used to establish a two-way interaction channel between experimental data and field measured data and numerical simulation data; The identification module is used to acquire the output data of the monitoring module and the data in the interaction channel of the interaction module, and to extract the pre-collapse feature information based on the acquired data. The verification and feedback module is used to perform bidirectional verification and calibration of the collapse critical criterion based on the linkage data of experiments, field and numerical simulation, and simultaneously feed the calibration results back to the user terminal. In the steep slope model, the karst network is constructed by referencing the actual development morphology and distribution pattern of natural karst through in-situ dynamic adjustment of forming parameters. This means that it includes a multi-level branching, heterogeneous and anisotropic karst conduit network.

2. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The in-situ dynamic control parameters for the formation of the karst network include the number of branch levels, the mean equivalent pore size, the coefficient of variation of pore size, the connectivity probability, and the spatial distribution entropy. These parameters are dynamically calibrated using a natural karst fit model. ； In the formula: The compatibility between karst networks and natural karst; These are the weighting coefficients; For the number of branch levels; The average number of branching levels of the natural karst in the target area; For aperture adaptation attenuation coefficient; This is the average equivalent aperture. The average equivalent pore size of the natural karst in the target area; The pore size variation coefficient; The pore size variation coefficient of the natural karst in the target area; The probability of connectivity; It represents the spatial distribution entropy; The spatial distribution entropy of natural karst in the target area; Real-time calculation of fit during regulation ,like If the value is less than the preset adaptation threshold, the molding parameters are adjusted sequentially according to their weight priority until... Once the threshold requirement is met, the actual morphology and distribution pattern of the karst network are verified by micro-CT scanning after the formation process.

3. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The monitoring module integrates distributed fiber optic sensors, miniature pore water pressure gauges, high-precision strain gauges, and high-speed cameras. Each device is deployed in a three-dimensional grid with preset location encryption. The data capture process is based on synchronization with a unified time base; Simultaneously capture data and perform trustworthiness verification: ； In the formula: To ensure the overall reliability of the monitoring data; The number of similar devices participating in the verification; Let be the accuracy weight of the i-th device; This represents the monitoring data value of the i-th device; This is the average value of monitoring data for similar equipment. The inherent measurement error of the i-th device; This is the time deviation influence coefficient; The sampling time deviation of the i-th device; Among them, if If the data is less than the preset confidence threshold, the abnormal data will be removed and the corresponding identification process will be refreshed and restarted.

4. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The collaborative loading mechanism in the loading module follows the following rules: The loading is performed sequentially from tectonic stress preloading to steady-state application of karst seepage to staged unloading, where the coupling loading parameters are: ； In the formula: The coupled loading total stress of the model at time t; This represents the initial preload value of the structural stress. , These are the seepage-stress coupling coefficient and the unloading-stress attenuation coefficient. Let be the karst seepage velocity at time t; The cumulative unloading strain at time t; The additional stress caused by seepage at time t; The grade difference of the graded unloading is based on The growth rate is dynamically adjusted so that the stress-seepage field of the model reaches a quasi-steady state after each unloading stage before the next loading stage is carried out.

5. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The two-way interaction channel of the interaction module adopts a three-layer architecture of data standardization, spatiotemporal alignment, and feature fusion: Data standardization layer: Converts experimental data, field measured data, and numerical simulation data into a preset data format, and uses standardization processing logic based on data type to eliminate differences in units; Spatiotemporal alignment layer: Temporal alignment is based on interpolation matching using a unified timestamp, while spatial alignment maps data from different sources to the same three-dimensional coordinate system through coordinate transformation; Feature fusion layer: The weighted average fusion method is used to fuse feature data from the same source. The fusion weights are dynamically allocated based on the credibility of the data source.

6. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The pre-collapse feature information extracted by the identification module includes the structural surface aperture growth rate, micro-fracture density, stress mutation coefficient, and seepage velocity variation rate. The feature extraction process includes: Data preprocessing: Wavelet threshold denoising algorithm is used to smooth and denoise the monitoring data to remove abnormal peaks caused by environmental interference; Feature calculation: Opening growth rate , This represents the increment of the structural surface opening within a preset time interval. To calculate the preset time interval for the opening growth rate, Initial aperture; microfracture density , V represents the number of micro-fractures, V represents the volume of the monitoring area, and the stress mutation coefficient. , This represents the maximum stress value within a preset time period. The average stress value during the same period; the seepage velocity variation rate. , The maximum seepage velocity within a preset time period. (This refers to the average seepage velocity during the same period). Significance screening: Calculate the coefficient of variation for each feature. , The standard deviation of the eigenvalues. For the mean of the eigenvalues, select Features that exceed a preset significance threshold are considered as valid precursor features of collapse, thus forming a precursor feature set.

7. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The collapse threshold criterion in the verification and feedback module is: ； In the formula: The collapse risk index at time t; The number of valid precursor features; is the weight coefficient of the j-th precursor feature; Let be the real-time value of the j-th precursor feature at time t; The critical threshold for the j-th precursor feature; The verification and feedback module uses a preset early warning threshold and calculates it in real time. A comparison is made with these data to issue an early warning.

8. The experimental system for a physical model of collapse on a steep, high-slope slope with coupled structural-karst-unloading coupling as described in claim 7, characterized in that, In the verification and feedback module, the bidirectional verification and calibration are iteratively optimized as follows: Substitute experimental data into the criterion calculation If the misjudgment rate exceeds the preset misjudgment threshold when compared with the actual collapse condition on site, then... Adjust the weighting coefficients, where The weight is adjusted until the misjudgment rate drops below the threshold. The calibrated criterion parameters are then synchronously stored in the system's preset built-in database.

9. The experimental system for a physical model of landslides on steep, deteriorated slopes coupled with tectonic-karst-unloading coupling, as described in claim 1, is characterized in that... The construction module is interconnected with a monitoring module via a wireless network. The monitoring module is interconnected with a loading module and an interaction module via a wireless network. The monitoring module and the interaction module are interconnected with an identification module via a wireless network. The identification module is interconnected with a verification and feedback module via a wireless network.

Images