Digital twinning-based rockfall slope geological early warning device and method

Abstract

The invention provides a rockfall slope geology early warning device and method based on digital twinning, and relates to the technical field of slope geology early warning. According to the invention, slope multi-dimensional physical parameters are obtained through the perception acquisition module, standardized processing and multi-protocol transmission are completed nearby through the edge transmission module, and the cloud load is reduced. The digital twinborn module constructs a mechanical twinborn model, dynamically updates and forms a digital twinborn body in combination with standard data, and accurately copies the slope state; and the early warning decision module realizes graded early warning through a risk assessment system. A full-link closed loop of sensing, transmission, modeling and early warning is integrally formed, the problems that traditional early warning data are fragmented, manual dependence is high, and static modeling is disjointed with reality are solved, conversion from passive response to active pre-judgment is achieved, on-site rapid response and global accurate analysis are balanced, and the accuracy, real-time performance and actual combat performance of early warning of the rockfall slope are improved.

Description

Technical Field

[0001] This invention mainly relates to the field of slope geological early warning technology, specifically to a rockfall slope geological early warning device and method based on digital twin. Background Technology

[0002] Mountainous areas account for more than two-thirds of my country's total land area. The construction of major infrastructure projects such as highways, railways, and water conservancy projects often requires traversing complex geological regions, resulting in numerous slope engineering projects. Affected by natural factors such as earthquakes, heavy rainfall, and weathering of soil and rock, as well as human engineering activities, geological disasters such as rockfalls and landslides occur frequently on slopes.

[0003] Traditional early warning systems often employ point-based monitoring, such as collecting localized data using only a few displacement gauges or rain gauges, or relying on regular manual inspections. The former cannot capture the overall deformation trend of the slope and the dynamics of key structural surfaces, such as joint development and rock loosening, easily creating monitoring blind spots; the latter is limited by terrain and weather, such as being difficult to reach in areas with heavy rain or steep slopes, and the low frequency of manual inspections makes it impossible to perceive subtle changes in the slope in real time, resulting in a high rate of missed detection of early hidden dangers.

[0004] Traditional technologies rely mainly on manual data processing and experience-based analysis, which is not only time-consuming and labor-intensive but also prone to judgment errors due to human error. Although some early warning systems introduce simple mechanical models, these models are mostly static presets and are not dynamically updated in conjunction with real-time monitoring data, thus failing to reflect the changes in the mechanical state of slopes under the influence of external dynamic factors such as rainfall and vibration.

[0005] In summary, existing slope rockfall early warning technologies suffer from drawbacks such as incomplete monitoring, inaccurate modeling, and poor adaptability. There is an urgent need for a technical solution that integrates multi-dimensional perception for intelligent risk assessment to improve the accuracy and practicality of slope rockfall early warning. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a geological early warning device and method for rockfall slopes based on digital twins, which addresses the shortcomings of the existing technology.

[0007] The technical solution of this invention to solve the above-mentioned technical problems is as follows: A geological early warning device for rockfall slopes based on digital twins, comprising: The sensing and acquisition module is deployed in the target early warning slope area to collect initial multi-dimensional physical parameters of the slope's geological environment; The edge transmission module, integrated in the edge control cabinet, is used to establish data transmission between the sensing and acquisition module and the cloud server through multiple transmission protocols. It is also used to standardize the initial multi-dimensional physical parameters to obtain standard multi-dimensional physical parameters and output them to the cloud server. The digital twin module, deployed on the cloud server, is used to construct a slope mechanical twin model and update the state of the slope mechanical twin model through the standard multi-dimensional physical parameters to obtain the slope digital twin model and output the core state parameters. The early warning decision module, deployed on the cloud server, is used to construct a rockfall risk assessment index system, determine the rockfall risk level of the core status parameters based on the rockfall risk assessment index system, and generate early warning information based on the judgment results.

[0008] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: A geological early warning method for rockfall slopes based on digital twins, comprising: Collect initial multi-dimensional physical parameters of the slope geological environment; The initial multidimensional physical parameters are standardized to obtain standard multidimensional physical parameters. A slope mechanics twin model is constructed, and the state of the slope mechanics twin model is updated using the standard multidimensional physical parameters to obtain a slope digital twin model and output the core state parameters. A rockfall risk assessment index system is constructed. Based on the rockfall risk assessment index system, the rockfall risk level of the core state parameters is judged, and early warning information is generated based on the judgment results.

[0009] The beneficial effects of this invention are: it transforms the multi-dimensional data of physical slopes into the dynamic state of a virtual digital twin, solving the problems of fragmented early warning data and strong reliance on manual intervention in traditional early warning systems, and realizing the transformation from passive response to proactive early warning; the edge transmission module processes data locally, reducing the data transmission pressure on the cloud; the digital twin and early warning decision-making module are deployed in the cloud, relying on cloud computing power to realize complex mechanical modeling and risk assessment, balancing rapid on-site response and accurate global analysis. Attached Figure Description

[0010] Figure 1 This is a functional module block diagram of the rockfall slope geological early warning device provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the geological early warning device for rockfall slopes provided in an embodiment of the present invention; Figure 3 A flowchart illustrating the geological early warning method for rockfall slopes provided in this embodiment of the invention. Detailed Implementation

[0011] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0012] Example 1: As Figure 1As shown, this embodiment of the invention provides a geological early warning device for rockfall slopes based on digital twins, comprising: The sensing and acquisition module is deployed in the target early warning slope area to collect initial multi-dimensional physical parameters of the slope's geological environment; The edge transmission module, integrated in the edge control cabinet, is used to establish data transmission between the sensing and acquisition module and the cloud server through multiple transmission protocols. It is also used to standardize the initial multi-dimensional physical parameters to obtain standard multi-dimensional physical parameters and output them to the cloud server. The digital twin module, deployed on the cloud server, is used to construct a slope mechanical twin model and update the state of the slope mechanical twin model through the standard multi-dimensional physical parameters to obtain the slope digital twin model and output the core state parameters. The early warning decision module, deployed on the cloud server, is used to construct a rockfall risk assessment index system, determine the rockfall risk level of the core status parameters based on the rockfall risk assessment index system, and generate early warning information based on the judgment results.

[0013] In the above embodiments, the multi-dimensional data of the physical slope is transformed into the dynamic state of the virtual digital twin, solving the problems of fragmented early warning data and strong reliance on manual intervention in traditional early warning systems, and realizing the transformation from passive response to proactive early warning. The edge transmission module processes data locally, reducing the data transmission pressure on the cloud. The digital twin and early warning decision-making module are deployed in the cloud, relying on cloud computing power to realize complex mechanical modeling and risk assessment, balancing rapid on-site response and accurate global analysis.

[0014] like Figure 2 As shown, preferably, the sensing and acquisition module includes a MEMS inertial sensor, a MEMS pressure sensor, a laser rangefinder, an environmental sensor, and a visual camera device; Initial multi-dimensional physical parameters of the slope geological environment were collected, including: The MEMS inertial sensor is provided in multiple locations, which are deployed in potential rockfall areas and key structural surfaces of the slope to collect vibration acceleration data of the slope body along the x, y, and z axes. The MEMS pressure sensor is provided in multiple ways, and the multiple MEMS pressure sensors are buried inside the slope at different depths to collect stress data inside the slope. The laser ranging sensor is provided in multiple units, which are deployed on the top and both sides of the slope to collect laser displacement data on the slope surface. The environmental sensors are provided in multiple locations, which are deployed in an open area around the slope to collect environmental rainfall data. The visual camera device is provided in multiple locations, which are deployed at different angles on the slope to collect visual images of the slope surface. This yields the initial multidimensional physical parameters.

[0015] The above embodiments cover five dimensions: vibration (MEMS inertial sensor), stress (MEMS pressure sensor), displacement (laser rangefinder sensor), environment (rainfall), and vision (camera equipment). The MEMS pressure sensor is buried at a certain depth, and the laser sensor is deployed at the top of the slope. The targeted deployment of the sensors solves the problems of narrow monitoring dimensions and large blind spots of traditional single sensors. The MEMS sensor collects triaxial vibration acceleration, the laser rangefinder sensor captures surface displacement, and the image supplementation by the vision equipment ensures that the data accuracy matches the requirements of digital twin modeling, avoiding model distortion caused by insufficient data accuracy.

[0016] Preferably, the edge transmission module includes a data standardization processing unit, which is used to perform data standardization processing on the initial multi-dimensional physical parameters, including: The vibration acceleration data is denoised using the Kalman filter algorithm to obtain standard vibration acceleration data. The internal stress data, laser displacement data, and environmental rainfall data of the slope were normalized using the min-max normalization algorithm to obtain standard internal stress data, standard laser displacement data, and standard environmental rainfall data. The slope surface visual image is subjected to grayscale processing and dehazing processing to obtain an optimized slope surface visual image. Based on the ROI region of interest extraction algorithm, the slope body is extracted from the optimized slope surface visual image to obtain the slope body region image. Standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall, and slope area images are correlated, and the correlated data is encapsulated to obtain standard multi-dimensional physical parameters.

[0017] It should be understood that selecting appropriate algorithms based on the characteristics of different types of data aims to improve data quality, standardize data formats, and provide reliable input for the accurate updating of digital twin models. Vibration acceleration data is easily affected by environmental noise (such as wind and vegetation swaying). The Kalman filter algorithm has the advantage of real-time dynamic noise reduction. It can filter out invalid noise while preserving the real vibration signal (such as acceleration changes caused by rock loosening and structural surface slippage), thus avoiding model updates caused by noisy data. The dimensions of internal slope stress, laser displacement, and rainfall are significantly different (e.g., stress is in Pa, displacement is in mm, and rainfall is in mm). Directly inputting these into the model would lead to an imbalance in mechanical calculations. By using the min-max normalization algorithm, these dimensions are mapped to the same numerical range (e.g., [0,1]), which preserves the relative size relationship of the data (reflecting the changing trend of the real physical quantities) and achieves dimension unification of multi-source data, facilitating subsequent fusion processing and model parameter iteration. Visual images are easily affected by weather (fog, rain) and lighting, resulting in blurred image quality. They also contain irrelevant areas around the slope (such as vegetation and sky). Grayscale processing can simplify the image data dimensions and reduce computational complexity. Dehazing can restore clear slope details. Then, the ROI region of interest extraction algorithm focuses on the core area of ​​the slope and removes irrelevant background to ensure that the image data only contains effective information related to the risk of rockfall (such as rock block morphology and slope cracks). Multi-source data association and encapsulation is the key to the "spatiotemporal consistency" of digital twins. It uses a unified timestamp as a benchmark to associate vibration, stress, displacement, rainfall and image data to ensure that multi-dimensional data at the same time correspond to the same state of the physical slope. This avoids contradictions in model update logic caused by spatiotemporal misalignment of data. The encapsulated data can be directly called by the digital twin module without additional format conversion, thus improving the efficiency of model update.

[0018] In the above embodiments, differentiated processing is adopted for different data characteristics to solve the problems of high noise, inconsistent format, and difficulty in extracting effective information in traditional data, and the data processing efficiency is significantly improved compared with manual processing. The standardized vibration, stress, displacement, rainfall, and image data are associated and encapsulated to ensure that the data is aligned in the time and space dimensions. This provides clean and related input data for the real-time updates of the digital twin model and avoids model bias caused by isolated multi-source data.

[0019] Preferably, constructing a slope mechanics twin model includes: Import the slope point cloud data captured by drone, and fit the slope surface features based on the slope point cloud data and the RANSAC algorithm to obtain a three-dimensional geometric twin model that matches the actual slope, and determine the spatial coordinate mapping relationship of each region of the actual slope. Based on the spatial morphology of the three-dimensional geometric twin model, the three-dimensional geometric twin model is divided into finite element meshes, and lithological basic mechanical parameters are assigned to the mesh units after division. Then, based on the structural surface distribution characteristics of the target early warning slope area, structural surface mechanical parameters are assigned to the mesh units and unit interfaces in the area where the structural surface is located, thus obtaining a preliminary slope mechanical twin model. Boundary conditions and load adaptation settings are applied to the preliminary slope mechanical twin model to obtain the slope mechanical twin model.

[0020] It should be understood that a three-dimensional geometric twin model is first constructed based on the point cloud data of the slope taken by drone. The core reason is that the point cloud data can accurately restore the actual spatial shape of the slope (such as slope, slope undulation, and outline size), while the RANSAC algorithm has the advantage of strong noise resistance and can effectively filter out abnormal points in the point cloud data (such as vegetation occlusion and invalid points caused by measurement errors), ensuring the matching degree between the geometric model and the actual slope. Assigning lithological basic mechanical parameters (such as elastic modulus, Poisson's ratio, and density) to grid cells is because lithology is the core determinant of slope mechanical properties (for example, the elastic modulus of granite is much higher than that of silty clay, and the difference in resistance to deformation is significant). By assigning parameter values ​​to grid cells, the virtual model can have "intrinsic mechanical properties" consistent with the physical slope. Assigning structural surface mechanical parameters (such as shear strength and interface stiffness) to the area where the structural surface is located is because the structural surface (joints, faults) is the weak link of the slope, and its mechanical properties are much lower than those of the rock block itself. It is the main inducing factor of rockfall and landslide. Targeted assignment can accurately simulate the influence of the structural surface on the slope stability. Boundary condition settings (such as fixed constraints at the bottom of the slope and horizontal constraints on the sides of the slope) must conform to the actual stress boundaries of the slope (such as the bottom of the slope being fixedly connected to the foundation and unable to generate displacement). Load adaptation settings (such as superimposing the self-weight of the slope) must restore the natural stress state of the slope to avoid distortion of mechanical calculation results due to the disconnect between the boundary or load and reality, thus laying a reliable foundation for subsequent state updates based on standard multi-dimensional physical parameters.

[0021] In the above embodiments, a three-dimensional geometric twin model is first constructed based on UAV point cloud and RANSAC algorithm, and then the lithological foundation mechanical parameters and structural surface mechanical parameters are assigned to it. This solves the limitations of traditional pure geometric models that have no mechanical properties and cannot simulate stress deformation, and meets the core requirements of digital twin geometric-physical mapping. By setting specific boundary conditions and loads, the stress environment of the mechanical model is made consistent with that of the physical slope, avoiding calculation distortion caused by improper boundary / load settings, and laying a reliable foundation for subsequent state updates and stability analysis.

[0022] Preferably, the standard multi-dimensional physical parameters include standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall, and slope area images; The slope mechanical twin model is updated using the standard multidimensional physical parameters to obtain a digital twin model of the slope and output state parameters, including: Using the timestamp of GPS and Beidou dual-mode clock synchronization as a reference, the standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall and slope area images are synchronized in time. The stiffness matrix [K] and mass matrix [M] of the slope mechanics twin model are constructed based on the lithological foundation mechanical parameters of the grid cell; Based on the mechanical equilibrium equations, the synchronized standard vibration acceleration data is substituted into the acceleration vector {u} of the mechanical equilibrium equations as the motion excitation. The synchronized standard environmental rainfall is converted into the additional self-weight load of the slope and superimposed on the total load vector {F(t)}. The stress and displacement results of the slope mechanical twin model are obtained by solving the equations using the Newmark-β method. The mechanical equilibrium equations are expressed as follows: , in, For displacement vectors, For velocity vector, For acceleration vectors, This is the total load vector. Here is the damping matrix. Using the synchronized standard slope internal stress data as mechanical feedback, the deviation between the model-calculated stress and the actual stress is compared, and the stiffness matrix is ​​iteratively corrected. K The local parameters of the slope mechanics twin model are determined by repeating the solution process until the deviation is less than the preset threshold, thus obtaining the iterative solution results of the twin model. Based on the iterative solution results of the slope mechanical twin model, the corresponding mesh elements of the three-dimensional geometric twin model are driven to perform morphological updates, thereby obtaining a digital twin of the slope. The rock block movement features are obtained from the slope area image, and the spatial position and movement state of the rock blocks in the slope digital twin model are updated based on the rock block movement features. The core state parameters used to characterize slope stability are extracted from the updated digital twin model of the slope and output to the early warning decision module. The core state parameters include the peak value of maximum vibration acceleration, cumulative displacement, stress concentration factor, rock block movement speed and overall slope stability coefficient.

[0023] It should be understood that in the above embodiments, the use of GPS and Beidou dual-mode clocks to synchronize timestamps is to ensure that multi-source data (vibration, stress, displacement, rainfall, images) are fully aligned in the time dimension—multi-dimensional data with the same timestamp correspond to the same state of the physical slope, avoiding logical contradictions in data fusion and model updates due to time deviations. Constructing a stiffness matrix based on lithological mechanical parameters [ K ] and quality matrix [ M ], because the stiffness matrix [ K The mass matrix directly reflects the slope's resistance to deformation (positively correlated with the elastic modulus of the lithology). MThe inertial characteristics of the slope (positively correlated with the density of the lithology) are reflected. These two factors are the core basis for solving the mechanical equilibrium equations, and their accuracy directly determines the reliability of the model calculation results. Standard vibration acceleration data is used as the acceleration vector input, and rainfall is converted into additional load because vibration is a common cause of slope instability (such as vibration caused by earthquakes or blasting), and rainfall increases the self-weight of the slope and reduces the strength of the soil and rock mass (especially silty soil and shale slopes). Both are key dynamic factors affecting slope stability and need to be used as core excitation terms for model updates. The stiffness matrix [K] is corrected by using the internal stress data of the standard slope as mechanical feedback. This is because stress is a direct reflection of the mechanical state of the slope. The deviation between the stress calculated by the model and the actual stress is essentially the difference between the model parameters (such as the preset value of the rock elastic modulus) and the actual slope. Through iterative correction, the virtual model can continuously approach the real mechanical state of the physical slope, solving the problem that the static model cannot adapt to dynamic changes. The update is supplemented by rock block movement features extracted from visual images because the mechanical model focuses on simulating the overall deformation of the slope, while the visual images can accurately capture the loosening and displacement of local rock blocks (such as the positional shift and rotation of individual rock blocks). The two work together to achieve a comprehensive state replication from "whole" to "local", ensuring the integrity and accuracy of the digital twin of the slope. The five core state parameters extracted are key indicators for rockfall risk assessment: the peak value of maximum vibration acceleration reflects the intensity of disturbance to the slope, the cumulative displacement reflects the degree of slope deformation, the stress concentration coefficient locates weak areas, the rock block movement speed is directly related to the probability of rockfall, and the overall slope stability coefficient is the core basis for risk judgment. This parameter set can fully support the hierarchical assessment of the early warning decision module and avoid misjudgment of risk due to missing parameters.

[0024] In the above embodiments, standard multi-dimensional physical parameters are used as input: vibration acceleration is used as the excitation of the mechanical equation, rainfall is converted into additional load, and the stiffness matrix is ​​corrected by combining stress data feedback. The Newmark-β method is used for iterative solution, so that the deviation between the model calculation results and the actual slope state is less than a preset threshold, thus solving the problem of traditional static modeling and inability to dynamically adapt to actual changes. Based on the mechanical solution results, the grid cell shape is updated, and the position and motion state of the rock blocks are calibrated by combining the rock block movement features extracted from the visual image, so as to achieve overall-local collaborative updates and ensure the dynamic consistency between the digital twin and the physical slope. Five parameters—peak vibration acceleration, cumulative displacement, stress concentration factor, rock movement speed, and overall stability factor—are extracted and directly correlated with the core influencing factors of rockfall risk, providing targeted data support for early warning decisions and avoiding redundant and ineffective parameters.

[0025] Preferably, a rockfall risk assessment index system is constructed, and the rockfall risk level is judged based on the core state parameters according to the rockfall risk assessment index system. Based on the judgment results, early warning information is generated, including: Based on geological survey data and historical rockfall case data of the target early warning slope, the risk classification thresholds corresponding to each level of indicators in the risk assessment indicator system are determined. The core status parameters are converted into quantitative scores. Based on the risk grading thresholds corresponding to each level of indicators, the quantitative scores corresponding to the core status parameters are used to determine the rockfall risk level, thereby obtaining risk degree information. The risk degree information includes safe level, low risk level, medium risk level, and high risk level. If the risk level is classified as medium risk, a warning message is generated to prompt the administrator to initiate the encrypted monitoring process. If the risk level information is high-risk, an emergency warning message is generated, which is used to prompt the administrator to initiate the emergency response process.

[0026] Specifically, the rockfall risk assessment index system is constructed as follows: Based on core state parameters, a hierarchical assessment system of "first-level indicators - second-level indicators" is established. The first-level indicators include slope stability indicators, rock movement risk indicators, and environmental induced indicators. The second-level indicators correspond to the core state parameters (peak value of maximum vibration acceleration, cumulative displacement, stress concentration factor, rock movement speed, and overall slope stability coefficient). The physical meaning of each second-level indicator and its weight influencing rockfall risk are clarified (determined through the analytic hierarchy process (AHP), such as a weight of 0.3 for the overall slope stability coefficient, 0.25 for the rock movement speed, 0.2 for the stress concentration factor, 0.15 for the cumulative displacement, and 0.1 for the peak value of maximum vibration acceleration). Based on geological survey data (lithology, shear strength of structural surfaces), historical rockfall case data, and industry standards for the target early warning slope, combined with the limit state calculation results of the slope mechanical twin model, the risk classification thresholds for each secondary indicator are determined and divided into four levels: "safe, low risk, medium risk, and high risk" according to the degree of risk. For example: the overall stability coefficient of the slope ≥1.3 is safe, 1.1-1.3 is low risk, 0.9-1.1 is medium risk, and less than 0.9 is high risk; the rock block movement speed <0.1m / s is safe, 0.1-0.3m / s is low risk, 0.3-0.5m / s is medium risk, and greater than 0.5m / s is high risk. Each core status parameter is standardized according to its corresponding grading threshold and converted into a quantitative score of 0-100 (the closer the indicator is to the risk limit, the higher the score); based on the preset indicator weights, the weighted summation method is used to calculate the comprehensive rockfall risk score.

[0027] Risk level ranges are determined based on the comprehensive score results. For example, a comprehensive score <30 indicates a safe level, 30-50 indicates a low-risk level, 50-70 indicates a medium-risk level, and a score greater than 70 indicates a high-risk level. If a single secondary indicator reaches the high-risk threshold (such as the overall slope stability coefficient <0.9), it is directly determined to be a high-risk level, regardless of the comprehensive score. Differentiated early warning information is generated for different risk levels. The safety level only records status data and does not trigger an early warning. The low-risk level generates a prompt early warning information, which includes the current core status parameters and trend analysis, reminding you to conduct regular inspections. The medium-risk level generates a warning early warning information, which clarifies the high-risk indicators and the area where rockfalls may occur, and initiates intensified monitoring. The high-risk level generates an emergency early warning information, which includes the probability of rockfall risk, the expected impact range, and emergency response suggestions (such as personnel evacuation and traffic control), and is simultaneously pushed to relevant responsible parties through multiple channels (SMS, platform push, and audible and visual alarms). Based on the updated core status parameters output in real time by the digital twin module, repeat steps 3-5 to dynamically adjust the comprehensive risk score and risk level, and synchronously update the warning information to ensure the real-time nature and accuracy of the warning, until the risk level drops to a safe range.

[0028] In the above embodiments, the grading threshold is determined based on the geological survey data of the target slope and historical rockfall cases, avoiding the problem that the general threshold is not suitable for specific slopes, and the evaluation results are more in line with the actual scenario; the medium / high risk levels are distinguished, and corresponding encrypted monitoring and emergency response differentiated early warning information is generated, and clear operation guidelines for administrators are provided.

[0029] Preferably, the edge transmission module includes a multi-protocol communication unit, which includes a LoRa-WAN communication subunit, a 5G communication subunit, and an Ethernet communication subunit; Data transmission between the sensing and acquisition module and the cloud server is established through multiple transmission protocols, including: The LoRa-WAN communication subunit is used to transmit data collected by MEMS inertial sensors, MEMS pressure sensors, laser rangefinders, and environmental sensors via the LoRa-WAN protocol. The 5G communication subunit is used to transmit data collected by the visual camera device via the 5G transmission protocol; The Ethernet communication subunit is used to output the processed standard multi-dimensional physical parameters to the cloud server via the TCP / IP protocol suite.

[0030] In the above embodiments, the LoRa-WAN protocol is adapted to low-power, long-distance sensor data, the 5G protocol supports high-bandwidth visual image transmission, and Ethernet ensures stable output of standardized data to the cloud through the TCP / IP protocol, solving the problem of poor adaptability of traditional single protocols and covering different scenarios from remote slopes to urban slopes. Multi-protocol redundancy backup avoids data interruption due to single protocol failure, and Ethernet transmission of standardized data relies on the reliability of the TCP / IP protocol to ensure that the data arrives at the cloud intact.

[0031] like Figure 3 As shown in the figure, this embodiment of the invention also provides a geological early warning method for rockfall slopes based on digital twins, including the following steps: S1. Collect initial multi-dimensional physical parameters of the slope geological environment; S2. Standardize the initial multidimensional physical parameters to obtain standard multidimensional physical parameters; S3. Construct a slope mechanical twin model, update the state of the slope mechanical twin model through the standard multi-dimensional physical parameters, obtain the slope digital twin model and output the core state parameters; S4. Construct a rockfall risk assessment index system, determine the rockfall risk level of the core state parameters based on the rockfall risk assessment index system, and generate early warning information based on the judgment results.

[0032] Preferably, the initial multidimensional physical parameters include vibration acceleration data, internal stress data of the slope, laser displacement data, environmental rainfall, and visual images of the slope surface; The initial multidimensional physical parameters are subjected to data standardization processing, including: The vibration acceleration data is denoised using the Kalman filter algorithm to obtain standard vibration acceleration data. The internal stress data, laser displacement data, and environmental rainfall data of the slope were normalized using the min-max normalization algorithm to obtain standard internal stress data, standard laser displacement data, and standard environmental rainfall data. The slope surface visual image is subjected to grayscale processing and dehazing processing to obtain an optimized slope surface visual image. Based on the ROI region of interest extraction algorithm, the slope body is extracted from the optimized slope surface visual image to obtain the slope body region image. Standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall, and slope area images are correlated, and the correlated data is encapsulated to obtain standard multi-dimensional physical parameters.

[0033] Preferably, constructing a slope mechanics twin model includes: Import the slope point cloud data captured by drone, and fit the slope surface features based on the slope point cloud data and the RANSAC algorithm to obtain a three-dimensional geometric twin model that matches the actual slope, and determine the spatial coordinate mapping relationship of each region of the actual slope. Based on the spatial morphology of the three-dimensional geometric twin model, the three-dimensional geometric twin model is divided into finite element meshes, and lithological basic mechanical parameters are assigned to the mesh units after division. Then, based on the structural surface distribution characteristics of the target early warning slope area, structural surface mechanical parameters are assigned to the mesh units and unit interfaces in the area where the structural surface is located, thus obtaining a preliminary slope mechanical twin model. Boundary conditions and load adaptation settings are applied to the preliminary slope mechanical twin model to obtain the slope mechanical twin model.

[0034] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0035] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and unit can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0036] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0037] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0038] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0039] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A geological early warning device for rockfall slopes based on digital twins, characterized in that, include: The sensing and acquisition module is deployed in the target early warning slope area to collect initial multi-dimensional physical parameters of the slope's geological environment; The edge transmission module, integrated in the edge control cabinet, is used to establish data transmission between the sensing and acquisition module and the cloud server through multiple transmission protocols. It is also used to standardize the initial multi-dimensional physical parameters to obtain standard multi-dimensional physical parameters and output them to the cloud server. The digital twin module, deployed on the cloud server, is used to construct a slope mechanical twin model and update the state of the slope mechanical twin model through the standard multi-dimensional physical parameters to obtain the slope digital twin model and output the core state parameters. The early warning decision module, deployed on the cloud server, is used to construct a rockfall risk assessment index system, determine the rockfall risk level of the core status parameters based on the rockfall risk assessment index system, and generate early warning information based on the judgment results.

2. The rockfall slope geological early warning device according to claim 1, characterized in that, The sensing and acquisition module includes a MEMS inertial sensor, a MEMS pressure sensor, a laser rangefinder, an environmental sensor, and a visual camera device. Initial multi-dimensional physical parameters of the slope geological environment were collected, including: The MEMS inertial sensor is provided in multiple locations, which are deployed in potential rockfall areas and key structural surfaces of the slope to collect vibration acceleration data of the slope body along the x, y, and z axes. The MEMS pressure sensor is provided in multiple ways, and the multiple MEMS pressure sensors are buried inside the slope at different depths to collect stress data inside the slope. The laser ranging sensor is provided in multiple units, which are deployed on the top and both sides of the slope to collect laser displacement data on the slope surface. The environmental sensors are provided in multiple locations, which are deployed in an open area around the slope to collect environmental rainfall data. The visual camera device is provided in multiple locations, which are deployed at different angles on the slope to collect visual images of the slope surface. This yields the initial multidimensional physical parameters.

3. The rockfall slope geological early warning device according to claim 2, characterized in that, The edge transmission module includes a data standardization processing unit, which is used to perform data standardization processing on the initial multi-dimensional physical parameters, including: The vibration acceleration data is denoised using the Kalman filter algorithm to obtain standard vibration acceleration data. The internal stress data, laser displacement data, and environmental rainfall data of the slope were normalized using the min-max normalization algorithm to obtain standard internal stress data, standard laser displacement data, and standard environmental rainfall data. The slope surface visual image is subjected to grayscale processing and dehazing processing to obtain an optimized slope surface visual image. Based on the ROI region of interest extraction algorithm, the slope body is extracted from the optimized slope surface visual image to obtain the slope body region image. Standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall, and slope area images are correlated, and the correlated data is encapsulated to obtain standard multi-dimensional physical parameters.

4. The rockfall slope geological early warning device according to claim 3, characterized in that, Constructing a slope mechanics twin model includes: Import the slope point cloud data captured by drone, and fit the slope surface features based on the slope point cloud data and the RANSAC algorithm to obtain a three-dimensional geometric twin model that matches the actual slope, and determine the spatial coordinate mapping relationship of each region of the actual slope. Based on the spatial morphology of the three-dimensional geometric twin model, the three-dimensional geometric twin model is divided into finite element meshes, and lithological basic mechanical parameters are assigned to the mesh units after division. Then, based on the structural surface distribution characteristics of the target early warning slope area, structural surface mechanical parameters are assigned to the mesh units and unit interfaces in the area where the structural surface is located, thus obtaining a preliminary slope mechanical twin model. Boundary conditions and load adaptation settings are applied to the preliminary slope mechanical twin model to obtain the slope mechanical twin model.

5. The rockfall slope geological early warning device according to claim 4, characterized in that, The standard multidimensional physical parameters include standard vibration acceleration data, standard internal stress data of the slope, standard laser displacement data, standard environmental rainfall, and images of the slope area. The slope mechanical twin model is updated using the standard multidimensional physical parameters to obtain a digital twin model of the slope and output state parameters, including: Using the timestamp of GPS and Beidou dual-mode clock synchronization as a reference, the standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall and slope area images are synchronized in time. Stiffness matrix of slope mechanical twin model constructed based on lithological foundation mechanical parameters of grid cells [ K ] and quality matrix [ M ]; Based on the mechanical equilibrium equations, the synchronized standard vibration acceleration data is used as the motion excitation and substituted into the mechanical equilibrium equations as the acceleration vector. The synchronized standard environmental rainfall is converted into an additional self-weight load on the slope and superimposed on the total load vector. ( The stress and displacement results of the slope mechanical twin model are obtained by solving the equations using the Newmark-β method. The mechanical equilibrium equations are expressed as follows: ， in, It is a displacement vector. For velocity vectors, For acceleration vectors, This is the total load vector. Here is the damping matrix. Using the synchronized standard slope internal stress data as mechanical feedback, the deviation between the model-calculated stress and the actual stress is compared, and the stiffness matrix is ​​iteratively corrected. K The local parameters of the slope mechanics twin model are determined by repeating the solution process until the deviation is less than the preset threshold, thus obtaining the iterative solution results of the twin model. Based on the iterative solution results of the slope mechanical twin model, the corresponding mesh elements of the three-dimensional geometric twin model are driven to perform morphological updates, thereby obtaining a digital twin of the slope. The rock block movement features are obtained from the slope area image, and the spatial position and movement state of the rock blocks in the slope digital twin model are updated based on the rock block movement features. The core state parameters used to characterize slope stability are extracted from the updated digital twin model of the slope and output to the early warning decision module. The core state parameters include the peak value of maximum vibration acceleration, cumulative displacement, stress concentration factor, rock block movement speed and overall slope stability coefficient.

6. The rockfall slope geological early warning device according to claim 5, characterized in that, A rockfall risk assessment index system is constructed. Based on this index system, the rockfall risk level of the core state parameters is determined, and early warning information is generated based on the assessment results, including: Based on geological survey data and historical rockfall case data of the target early warning slope, the risk classification thresholds corresponding to each level of indicators in the risk assessment indicator system are determined. The core status parameters are converted into quantitative scores. Based on the risk grading thresholds corresponding to each level of indicators, the quantitative scores corresponding to the core status parameters are used to determine the rockfall risk level, thereby obtaining risk degree information. The risk degree information includes safe level, low risk level, medium risk level, and high risk level. If the risk level is classified as medium risk, a warning message is generated to prompt the administrator to initiate the encrypted monitoring process. If the risk level information is high-risk, an emergency warning message is generated, which is used to prompt the administrator to initiate the emergency response process.

7. The rockfall slope geological early warning device according to any one of claims 1 to 6, characterized in that, The edge transmission module further includes a multi-protocol communication unit, which includes a LoRa-WAN communication subunit, a 5G communication subunit, and an Ethernet communication subunit. Data transmission between the sensing and acquisition module and the cloud server is established through multiple transmission protocols, including: The LoRa-WAN communication subunit is used to transmit data collected by MEMS inertial sensors, MEMS pressure sensors, laser rangefinders, and environmental sensors via the LoRa-WAN protocol. The 5G communication subunit is used to transmit data collected by the visual camera device via the 5G transmission protocol; The Ethernet communication subunit is used to output the processed standard multi-dimensional physical parameters to the cloud server via the TCP / IP protocol suite.

8. A geological early warning method for rockfall slopes based on digital twins, characterized in that, Includes the following steps: Collect initial multi-dimensional physical parameters of the slope geological environment; The initial multidimensional physical parameters are standardized to obtain standard multidimensional physical parameters. A slope mechanics twin model is constructed, and the state of the slope mechanics twin model is updated using the standard multidimensional physical parameters to obtain a slope digital twin model and output the core state parameters. A rockfall risk assessment index system is constructed. Based on the rockfall risk assessment index system, the rockfall risk level of the core state parameters is judged, and early warning information is generated based on the judgment results.

9. The geological early warning method for rockfall slopes according to claim 8, characterized in that, The initial multidimensional physical parameters include vibration acceleration data, internal stress data of the slope, laser displacement data, environmental rainfall, and visual images of the slope surface; The initial multidimensional physical parameters are subjected to data standardization processing, including: The vibration acceleration data is denoised using the Kalman filter algorithm to obtain standard vibration acceleration data. The internal stress data, laser displacement data, and environmental rainfall data of the slope were normalized using the min-max normalization algorithm to obtain standard internal stress data, standard laser displacement data, and standard environmental rainfall data. The visual images of the slope surface are processed by grayscale conversion and dehazing to obtain optimized visual images of the slope surface. Based on the ROI region of interest extraction algorithm, the slope body is extracted from the optimized visual images of the slope surface to obtain the slope body region image. Standard vibration acceleration data, standard slope internal stress data, standard laser displacement data, standard environmental rainfall, and slope area images are correlated, and the correlated data is encapsulated to obtain standard multi-dimensional physical parameters.

10. The geological early warning method for rockfall slopes according to claim 9, characterized in that, Constructing a slope mechanics twin model includes: Import the slope point cloud data captured by drone, and fit the slope surface features based on the slope point cloud data and the RANSAC algorithm to obtain a three-dimensional geometric twin model that matches the actual slope, and determine the spatial coordinate mapping relationship of each region of the actual slope. Based on the spatial morphology of the three-dimensional geometric twin model, the three-dimensional geometric twin model is divided into finite element meshes, and lithological basic mechanical parameters are assigned to the mesh units after division. Then, based on the structural surface distribution characteristics of the target early warning slope area, structural surface mechanical parameters are assigned to the mesh units and unit interfaces in the area where the structural surface is located, thus obtaining a preliminary slope mechanical twin model. Boundary conditions and load adaptation settings are applied to the preliminary slope mechanical twin model to obtain the slope mechanical twin model.

Images