Method and system for monitoring internal stability of rock slope

Abstract

The embodiment of the invention provides a method for monitoring the internal stability of a rock slope, and the method comprises the steps: constructing a multi-dimensional sensing network, and carrying out the processing of the collected geological structure and environment parameters of the slope, and obtaining a multi-source data set with time-space alignment; carrying out at least one of signal identification processing, fracture evolution inversion and multi-source correlation modeling on the multi-source data set to obtain a slope stability characteristic index; and determining a disaster response instruction based on the slope stability characteristic index.

Description

Technical Field

[0001] This specification relates to the field of microseismic monitoring, and in particular to a method and system for monitoring the internal stability of rock slopes. Background Technology

[0002] As my country's infrastructure construction extends into the complex geological regions of western China, the stability monitoring of rock slopes has become a core challenge in ensuring the safety of major projects. Taking the Sichuan-Tibet Railway as an example, steep rock slopes along its route commonly exhibit high-altitude, long-distance geological hazard risks. For instance, the No. 1 landslide on the northeast side of Batang Station is a typical large-scale, high-altitude rock landslide with a thickness of 15-40 meters and a volume exceeding 6 million cubic meters. It not only directly threatens the safety of engineering facilities but also poses a significant risk to the lives and property of surrounding residents. The stability problems of such slopes exhibit a significant "three-dimensional" nature—surface displacement is the direct manifestation of deformation, while the fracturing and evolution of deep rock masses is the core driving factor for instability.

[0003] However, traditional monitoring technologies are limited by the limitations of single methods, making it difficult to construct a real-time, full-chain monitoring system for "surface-subsurface" deformation: although surface monitoring technologies, represented by GNSS, can achieve millimeter-level displacement accuracy, they only reflect the planar and elevation changes of the slope surface and cannot capture the spatiotemporal evolution of deep rock mass fractures; although microseismic monitoring technologies can identify underground rupture events through seismic wave signals, they lack real-time spatial coupling with surface deformation data, making it difficult to establish a causal relationship between "rupture-displacement".

[0004] In addition, in the high mountains and valleys and cold temperate climate zones along the Sichuan-Tibet Railway, traditional equipment faces problems such as unstable power supply, communication interruption, and insufficient resistance to corrosion, resulting in a significant decrease in data continuity and reliability. Moreover, the existing system relies on offline calculation, and the cycle from data collection to early warning issuance can be as long as several hours or even several days, which cannot meet the timeliness requirements in emergency engineering scenarios.

[0005] Therefore, how to conduct real-time and accurate monitoring of the surface and internal stability of rock slopes is a key scientific problem that must be solved. Summary of the Invention

[0006] This specification provides one or more embodiments of a method for monitoring the internal stability of a rock slope, comprising: constructing a multidimensional sensing network to process the collected slope geological structure and environmental parameters to obtain a spatiotemporally aligned multi-source dataset; performing at least one of signal recognition processing, fracture evolution inversion, and multi-source correlation modeling on the multi-source dataset to obtain slope stability characteristic indicators; and determining disaster response instructions based on the slope stability characteristic indicators.

[0007] In some embodiments, the multidimensional sensing network includes: a surface displacement monitoring module configured to acquire slope displacement data; an underground rupture monitoring module configured to acquire rock mass rupture signals; an environmental parameter monitoring module configured to acquire rainfall data and ambient temperature and humidity; and a data transmission and preprocessing module configured to perform edge preprocessing and transmission on the acquired data.

[0008] In some embodiments, the signal recognition processing includes: obtaining P / S wave arrival time markers based on the original microseismic waveforms, device spatial coordinates, and environmental noise baselines in the multi-source dataset through wavelet denoising, bandpass filtering, CNN feature extraction, Conv-LSTM time series modeling, Unet semantic segmentation, and post-processing optimization.

[0009] In some embodiments, the rupture evolution inversion includes: acquiring travel time data, the travel time data including the P / S wave arrival time markers, the spatial coordinates of the equipment, and the initial wave velocity model; constructing an objective function, determining the travel time residuals based on the travel time data and the objective function; and performing iterative optimization based on the travel time residuals to obtain the optimal source parameters.

[0010] In some embodiments, the rupture evolution inversion further includes: in response to detecting a displacement jump or rainfall reaching a preset condition, optimizing the parameters of the initial wave velocity model to obtain updated travel time data, updating the travel time residuals based on the updated travel time data, and updating the optimal source parameters based on the updated travel time residuals.

[0011] In some embodiments, the multi-source correlation modeling includes: determining a spatiotemporal correlation matrix based on the rainfall data, ambient temperature and humidity, slope displacement data, and optimal source parameters, using spatiotemporal correlation analysis; determining a physical correlation model based on the spatiotemporal correlation matrix, the optimal source parameters, the rainfall data, ambient temperature and humidity, and slope displacement data; and determining slope stability characteristic indicators based on the spatiotemporal correlation matrix and the physical correlation model.

[0012] In some embodiments, determining the disaster response instruction based on the slope stability characteristic index includes: determining the disaster response instruction based on the slope stability characteristic index, signal segmentation delay processing, and early warning rule judgment, wherein the disaster response instruction includes optimizing the equipment deployment plan and the equipment maintenance scheduling plan.

[0013] This specification provides one or more embodiments of a monitoring system for the internal stability of a rock slope, comprising: a sensing module configured to construct a multi-dimensional sensing network to process the collected slope geological structure and environmental parameters to obtain a spatiotemporally aligned multi-source dataset; an analysis module configured to perform at least one of signal recognition processing, fracture evolution inversion, and multi-source correlation modeling on the multi-source dataset to obtain slope stability characteristic indicators; and a decision module configured to determine disaster response instructions based on the slope stability characteristic indicators.

[0014] This specification provides one or more embodiments of a device for monitoring the internal stability of a rock slope, including a processor for executing a method for monitoring the internal stability of a rock slope.

[0015] This specification provides one or more embodiments of a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions from the storage medium, the computer executes a method for monitoring the internal stability of a rock slope. Attached Figure Description

[0016] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0017] Figure 1 This is an exemplary schematic diagram of a monitoring system for the internal stability of a rock slope according to some embodiments of this specification;

[0018] Figure 2 This is an exemplary flowchart of a method for monitoring the internal stability of a rock slope according to some embodiments of this specification;

[0019] Figure 3 This is an exemplary schematic diagram of a multidimensional sensing network according to some embodiments of this specification;

[0020] Figure 4 This is an exemplary flowchart of the system structure topology of the joint monitoring system shown in some embodiments of this specification;

[0021] like Figure 5 The diagram shown is a schematic diagram of a millimeter-level displacement monitoring data cloud platform as illustrated in some embodiments of this specification;

[0022] Figure 6 The diagram shown is a plan view of the measuring points in the demonstration area for monitoring the deformation of the steep slope at Batang Station on the Sichuan-Tibet Railway, as illustrated in some embodiments of this specification.

[0023] Figure 7This is a three-dimensional top view showing the distribution of landslides on both sides of Batang Station as illustrated in some embodiments of this specification;

[0024] Figure 8 These are GNSS displacement monitoring results of slope #3 as shown in some embodiments of this specification;

[0025] Figure 9 These are GNSS displacement monitoring results of slope #5 as shown in some embodiments of this specification;

[0026] Figure 10 This is a comparison chart of GNSS displacement monitoring data for slopes #3 and #5 as shown in some embodiments of this specification;

[0027] Figure 11 These are the consistency test results of the seismic detectors shown in some embodiments of this specification;

[0028] Figure 12 These are the results of pre-field conformity tests for the microseismic monitoring equipment shown in some embodiments of this specification;

[0029] Figure 13 This is a three-dimensional result diagram of the source location of the underground microseismic event on the high slope of Deda Township, as shown in some embodiments of this specification.

[0030] Figure 14 This is a layout diagram of a general GNSS monitoring equipment as shown in some embodiments of this specification. Detailed Implementation

[0031] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0032] It should be understood that the terms “system,” “device,” “unit,” and / or “module” used herein are one way to distinguish different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0033] As indicated in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0034] Flowcharts are used in this specification to illustrate the operations performed by the system according to embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0035] In recent years, research in the field of rock slope monitoring has made some progress: Ultra-wideband (UWB) ranging technology has improved the accuracy of surface displacement monitoring to the millimeter level through triangulation calculation; the dynamic calculation frequency of Global Navigation Satellite System (GNSS) can reach 1Hz; the sampling rate of microseismic equipment has been optimized to above 250Hz and combined with array deployment to control the rupture location accuracy within 10m; and designs such as hybrid power supply of solar energy and batteries and low-power 4G modules have also improved the equipment's endurance in scenarios without mains power.

[0036] However, these technologies still have fundamental flaws: independent operation of surface and underground monitoring equipment leads to inconsistent spatial benchmarks and significant spatiotemporal registration errors in multi-source data; there is a lack of differentiated deployment strategies for steep slopes with "steep upper slopes and gentle lower slopes," deep rupture signals are easily interfered with by surface noise, and the microseismic event identification rate is about 60%; data fusion remains at the level of "result superposition" and has not formed a joint inversion model based on physical mechanisms, making it difficult to reveal the gradual evolution law of slope instability.

[0037] Currently, with the continuous advancement of universal GNSS monitoring technology, its static planar displacement accuracy can reach ±(2.5mm+0.5×10D), enabling real-time capture of millimeter-level surface deformation on slopes. Ultra-wideband (UWB) positioning technology, integrated with an inertial navigation module, achieves a horizontal positioning accuracy of <0.01m for millimeter-level surface displacement ranging monitoring stations, allowing for high-precision three-dimensional displacement calculations in complex terrain by penetrating non-metallic obstructions. Microseismic monitoring equipment, through a 32-bit high-resolution acquisition host and a 5Hz detector, records underground rock fracture signals in real-time at a 250Hz sampling rate, achieving precise timing of microseismic events by combining BeiDou satellite time synchronization (time synchronization accuracy <20 microseconds). Simultaneously, millimeter-level displacement ranging and microseismic joint monitoring technology, through ultra-wideband positioning, multi-sensor integration, and 4G data transmission, can achieve synchronous acquisition of surface and underground data, and the simulated annealing algorithm can realize the spatiotemporal positioning of microseismic sources.

[0038] Therefore, it is necessary to construct a three-dimensional monitoring network that combines GNSS surface displacement monitoring, UWB millimeter-level ranging, and microseismic underground rupture monitoring to provide a real-time, high-precision, and integrated solution for rock slope stability assessment and microseismic location.

[0039] Figure 1 This is an exemplary schematic diagram of a monitoring system for the internal stability of a rock slope according to some embodiments of this specification.

[0040] like Figure 1 As shown, the monitoring system 100 for the internal stability of a rock slope includes a sensing module 110, an analysis module 120, and a decision module 130. In some embodiments, the sensing module 110, the analysis module 120, and the decision module 130 can be communicatively connected to each other. In some embodiments, some or all of the sensing module 110, the analysis module 120, and the decision module 130 can be configured in a processor.

[0041] In some embodiments, the sensing module is configured to construct a multidimensional sensing network to process the collected slope geological structure and environmental parameters to obtain a spatiotemporally aligned multi-source dataset.

[0042] In some embodiments, the multidimensional sensing network includes: a surface displacement monitoring module configured to acquire slope displacement data; an underground rupture monitoring module configured to acquire rock mass rupture signals; an environmental parameter monitoring module configured to acquire rainfall data and ambient temperature and humidity; and a data transmission and preprocessing module configured to perform edge preprocessing and transmission on the acquired data.

[0043] In some embodiments, the analysis module is configured to perform at least one of signal recognition processing, rupture evolution inversion, and multi-source correlation modeling on the multi-source dataset to obtain slope stability characteristic indices.

[0044] In some embodiments, the signal recognition processing includes: obtaining P / S wave arrival time markers based on the original microseismic waveforms, device spatial coordinates, and environmental noise baselines in the multi-source dataset through wavelet denoising, bandpass filtering, CNN feature extraction, Conv-LSTM time series modeling, Unet semantic segmentation, and post-processing optimization.

[0045] In some embodiments, the rupture evolution inversion includes: acquiring travel time data, the travel time data including the P / S wave arrival time markers, the spatial coordinates of the equipment, and the initial wave velocity model; constructing an objective function, determining the travel time residuals based on the travel time data and the objective function; and performing iterative optimization based on the travel time residuals to obtain the optimal source parameters.

[0046] In some embodiments, the rupture evolution inversion further includes: in response to detecting a displacement jump or rainfall reaching a preset condition, optimizing the parameters of the initial wave velocity model to obtain updated travel time data, updating the travel time residuals based on the updated travel time data, and updating the optimal source parameters based on the updated travel time residuals.

[0047] In some embodiments, the multi-source correlation modeling includes: determining a spatiotemporal correlation matrix based on the rainfall data, ambient temperature and humidity, slope displacement data, and optimal source parameters, using spatiotemporal correlation analysis; determining a physical correlation model based on the spatiotemporal correlation matrix, the optimal source parameters, the rainfall data, ambient temperature and humidity, and slope displacement data; and determining slope stability characteristic indicators based on the spatiotemporal correlation matrix and the physical correlation model.

[0048] In some embodiments, the decision module is configured to determine disaster response instructions based on the slope stability characteristic indicators.

[0049] In some embodiments, the decision module is further configured to: determine a disaster response instruction based on the slope stability characteristic index, signal segmentation delay processing, and early warning rule judgment, wherein the disaster response instruction includes an optimized equipment deployment scheme and an equipment maintenance scheduling scheme.

[0050] For further explanation of the above modules, please refer to the corresponding content below.

[0051] It should be noted that the above description of the monitoring system and its modules for the internal stability of rock slopes is for ease of description only and should not be construed as limiting this specification to the scope of the embodiments described. It is understood that those skilled in the art, after understanding the principles of the system, may arbitrarily combine the various modules or construct subsystems connected to other modules without departing from these principles. In some embodiments, Figure 2The perception module 110, analysis module 120, and decision-making module 130 disclosed herein can be different modules within a single system, or a single module can implement the functions of two or more of the aforementioned modules. For example, the modules can share a single storage module, or each module can have its own separate storage module. Such variations are all within the scope of protection of this specification.

[0052] Figure 2 This is an exemplary flowchart of a method for monitoring the internal stability of rock slopes according to some embodiments of this specification. Figure 2 As shown, process 200 includes the following steps. In some embodiments, process 200 may be executed by a monitoring system 200 for the internal stability of a rock slope or by a processor; the following description uses processor-based execution as an example.

[0053] Step 201: Construct a multi-dimensional sensing network to process the collected slope geological structure and environmental parameters to obtain a spatiotemporally aligned multi-source dataset.

[0054] Figure 3 This is an exemplary schematic diagram of a multidimensional sensing network 300 according to some embodiments of this specification.

[0055] like Figure 3 As shown, the multidimensional sensing network 300 includes a surface displacement monitoring module 310, an underground rupture monitoring module 220, an environmental parameter monitoring module 330, and a data transmission and preprocessing module 340.

[0056] In some embodiments, the surface displacement monitoring module is configured to acquire slope displacement data.

[0057] In some embodiments, the surface displacement monitoring module may include a universal GNSS station, and the processor may collect three-dimensional displacement data of the slope surface based on the universal GNSS station. The three-dimensional displacement of the slope surface includes horizontal displacement (e.g., in the X / Y direction) and vertical displacement (in the Z direction).

[0058] In some embodiments, a universal GNSS station can be based on an accuracy of ±(2.5mm+0.5×10). -6 D) Sampling rate: 1Hz for data acquisition and real-time three-dimensional coordinates.

[0059] In some embodiments, the universal GNSS station can also preprocess the acquired data, such as by filtering the acquired data with satellite signals (to eliminate multipath effects) and performing dynamic calculations (RTK / PPP mode).

[0060] In some embodiments, the surface displacement monitoring module may include a millimeter-level displacement station, and the processor may collect millimeter-level three-dimensional displacement of the slope based on the millimeter-level displacement station. The millimeter-level three-dimensional displacement includes UWB ranging values ​​(distance between the reference point and the monitoring point) and data such as acceleration or angular velocity from the inertial navigation module.

[0061] In some embodiments, the millimeter-level displacement station can acquire data and obtain real-time three-dimensional coordinates based on working parameters such as accuracy: horizontal error <1cm, vertical error <3cm, sampling rate: 10Hz, and penetration capability: able to penetrate non-metallic obstructions (trees, netting).

[0062] In some embodiments, the millimeter-level displacement station can also preprocess the collected data, such as performing multi-source positioning fusion processing (UWB + Bluetooth + inertial navigation) and displacement vector calculation (triangulation adjustment) on the collected data.

[0063] In some embodiments, the underground fracture monitoring module is configured to acquire rock mass fracture signals.

[0064] In some embodiments, the underground rupture monitoring module may include a microseismic monitoring station, and the processor may acquire rock mass rupture signals based on the microseismic monitoring station. The rock mass rupture signals include seismic wave voltage signals and event trigger times.

[0065] In some embodiments, the microseismic monitoring station can acquire data based on operating parameters such as sampling rate: 250Hz, detector frequency response: 5Hz-1kHz, and array layout: 3×3 grid (average spacing 200m) to obtain rock mass fracture signals.

[0066] In some embodiments, the microseismic monitoring station can also preprocess the acquired data, such as performing signal gain control (adaptive amplification of weak signals) and local caching (temporary storage on a flash memory card, with a storage capacity of 2MB triggering transmission).

[0067] In some embodiments, the environmental parameter monitoring module is configured to acquire rainfall data and ambient temperature and humidity.

[0068] In some embodiments, the environmental parameter monitoring module may include a rain sensor, a temperature / humidity sensor, etc.

[0069] In some embodiments, the rain sensor can collect rainfall intensity, such as by collecting data based on operating parameters with a resolution of ±1 mm, and obtain the cumulative rainfall (mm / h).

[0070] In some embodiments, the temperature / humidity sensor can be the ambient temperature and humidity.

[0071] In some embodiments, the data transmission and preprocessing module is configured to perform edge preprocessing and transmission on the acquired data.

[0072] In some embodiments, the data transmission and preprocessing module includes a passage function and an edge preprocessing function. The passage function can be implemented based on a 4G real-time transmission communication network, and in extreme weather emergencies, it can be implemented based on BeiDou short message service, etc.

[0073] In some embodiments, edge preprocessing includes coordinate transformation of GNSS coordinates (e.g., converting WGS84 coordinates to a local coordinate system) to achieve a unified spatial reference. For example, GNSS reference point coordinates (e.g., 99°24′2.84″E, 30°16′56.50″N) can be used as the spatial origin for all devices to eliminate registration errors.

[0074] In some embodiments, edge preprocessing includes temperature compensation of the UWB ranging values ​​(e.g., eliminating thermal expansion errors) to improve ranging accuracy.

[0075] In some embodiments, edge preprocessing includes preliminary filtering of the microseismic waveform (e.g., based on 50Hz power frequency noise suppression) to reduce invalid data uploads.

[0076] In some embodiments, edge preprocessing includes packaging environmental data hourly or periodically (e.g., every 10 minutes) to reduce communication load.

[0077] In some embodiments, the processor can perform spatial benchmarking and other processing on data such as slope geological structure (attitude, location of fracture zones) and environmental parameters (temperature, rainfall, topography) to obtain a spatiotemporally aligned multi-source dataset. For example, three types of equipment, such as GNSS+UWB displacement stations+microseismic stations, are deployed at key points on potential slip surfaces, with the benchmark points placed on stable bedrock (such as the bottom of a gully), forming a non-coplanar triangular monitoring network.

[0078] The spatiotemporally aligned multi-source dataset includes surface displacement (GNSS: 1Hz three-dimensional coordinates; UWB: 10Hz millimeter-level ranging), underground rupture signals (microseismic: 250Hz voltage waveform), and environmental data (rainfall: 1 / 10min).

[0079] In some embodiments, the individual module units in the multidimensional sensing network can be powered by solar energy and employ dual-mode communication (4G / BeiDou).

[0080] In some embodiments, the multi-source dataset obtained in step 201 has a unified spatiotemporal label, which can eliminate the spatial registration error of traditional methods.

[0081] Step 202: Perform at least one of the following on the multi-source dataset: signal recognition processing, rupture evolution inversion, and multi-source correlation modeling, to obtain slope stability characteristic indicators.

[0082] In some embodiments, the signal recognition processing includes: obtaining P / S wave arrival time markers based on the original microseismic waveforms, device spatial coordinates, and environmental noise baselines in the multi-source dataset through wavelet denoising, bandpass filtering, CNN feature extraction, Conv-LSTM time series modeling, Unet semantic segmentation, and post-processing optimization.

[0083] In some embodiments, signal identification processing refers to the intelligent identification of microseismic signals. In some embodiments, the objects of signal identification processing include the original microseismic waveform, equipment spatial coordinates (from the common point layout in step 201), and environmental noise baseline (generated based on the rainfall / temperature data in step 201). The specific processing steps include:

[0084] Wavelet denoising includes: selecting a db4 wavelet basis (matching the micro-vibration frequency of 20-100Hz), performing 5-level decomposition, processing high-frequency noise based on hard thresholding, and finally reconstructing the signal. Based on wavelet denoising, the signal-to-noise ratio (SNR) of the signal is increased from 5dB to 15dB.

[0085] Bandpass filtering includes: based on a 10-200Hz Butterworth filter, power frequency interference (50Hz) and low-frequency environmental vibration are removed, ultimately achieving a purified waveform, such as retaining only the characteristic peaks of the broken waveform.

[0086] CNN feature extraction includes: extracting short-term amplitude abrupt changes (corresponding to the initial arrival of rock mass fracture) based on 5×1 convolution kernel 1, and capturing long-term frequency decay (identifying P / S wave differences) based on 20×1 convolution kernel 2, ultimately outputting a 128-dimensional time-frequency feature vector.

[0087] Conv-LSTM time series modeling includes: processing the input feature vector (e.g., LSTM cells memorizing the first 60 sampling points) to obtain contextual features, such as associating the physical propagation process from P-wave to S-wave, in order to solve the problem of traditional methods missing the S-wave.

[0088] Unet semantic segmentation includes: extracting abstract patterns based on encoder downsampling and compression features; fusing shallow details (such as first-arrival spikes) with deep semantics based on skip connections; and restoring temporal resolution based on decoder upsampling to output the class probability (noise / P-wave / S-wave) for each sample point, ultimately obtaining a three-channel probability sequence. As an example only, the three-channel probability sequence [0.02, 0.91, 0.07] represents a 91% probability of P-wave.

[0089] Post-processing optimizations include: Gaussian smoothing (e.g., window size 5ms) to eliminate isolated noise points; morphological operations such as removing spurious P-wave segments with a duration of <2ms and filling S-wave gaps, ultimately outputting accurate P-wave first arrival times and S-wave first arrival times.

[0090] In some embodiments, the rupture evolution inversion includes: acquiring travel time data, the travel time data including the P / S wave arrival time markers, the spatial coordinates of the equipment, and the initial wave velocity model; constructing an objective function, determining the travel time residuals based on the travel time data and the objective function; and performing iterative optimization based on the travel time residuals to obtain the optimal source parameters.

[0091] In some embodiments, the objective function is a function used to quantify the deviation between candidate source solutions and the true solution. The inputs to the objective function include travel time data, geophone coordinates, and an initial wave velocity model. The output of the objective function includes the travel time residual E(m) (a scalar value in milliseconds), which can characterize the error level of the current source parameters m = (x, y, z, t0), where (x, y, z) represents the three-dimensional coordinates of the source (in meters), and t0 represents the time of origin (in seconds). Based on the objective function, physical observations (P / S wave arrival time difference) can be transformed into an optimizable mathematical problem, making source location a matter of finding the parameter combination that minimizes E(m).

[0092] In some embodiments, the objective function can be expressed as: in, m represents the current source parameters m = (x, y, z, t0); Let be the arrival time (i.e., first arrival time) of the P-wave and S-wave recorded by the i-th detector, determined based on travel time data. To calculate the theoretical arrival times of the P-wave and S-wave of the i-th detector based on the current source parameter m and wave velocity model, d i This represents the distance from the earthquake source location (x, y, z) to the i-th detector (x). i ,y i ,z i The distance between ) and v p ,v s This represents the propagation velocity of P-waves and S-waves in the rock mass. The initial value is given by a three-dimensional wave velocity model constructed from geological survey data, and is corrected in real time during the iterative optimization process according to dynamic parameter correction rules (such as correction based on displacement and rainfall). t0 represents the time of earthquake occurrence.

[0093] In some embodiments, different weights can be assigned to different detectors based on the signal-to-noise ratio (SNR), such that the higher the SNR, the greater the weight. The objective function can then be expressed as: in, It is the weight of the i-th detector, which can be determined based on the signal-to-noise ratio of the P-wave and S-wave detected by the i-th detector.

[0094] In some embodiments, iterative optimization refers to VFSA (Very Fast Simulated Annealing, minute-level source location) iterative optimization, which can iteratively optimize the current source parameters m = (x, y, z, t0).

[0095] In some embodiments, the iterative optimization process includes: based on the current solution m old The random perturbation -R: a uniformly distributed random number from 0 to 1 is used to calculate the residual and then the objective function value E of the new solution. new =E(m) new ); Determine E new With E old The size of E new Less than E old If the current source parameters are accepted with a probability of 100%, then the occurrence probability of the current source parameters is considered to be P, where P = e^(-1 / 2). (-(Enew-Eold) / T) T is the current exponential decay temperature, where T {k+1} =0.95×T k The iteration ends when the termination condition is met. The termination condition includes either the residual change rate being less than a preset value (e.g., 0.1%) or the number of iterations being greater than a preset number (e.g., 200).

[0096] Wherein, based on the current solution m old The random perturbation - R: a uniformly distributed random number from 0 to 1, can be used to calculate the residual, which may include v p ,v s The figures were reduced by 50% and 2% respectively, as part of the updated v p ,v s The objective function is then updated and recalculated.

[0097] The VFSA iterative optimization described above drives parameter updates by repeatedly evaluating the objective function value, and determines the search direction by the probability of E(m) increasing or decreasing, thus avoiding getting trapped in local optima.

[0098] In some embodiments, the rupture evolution inversion further includes: in response to detecting a displacement jump or rainfall reaching a preset condition, optimizing the parameters of the initial wave velocity model to obtain updated travel time data, updating the travel time residuals based on the updated travel time data, and updating the optimal source parameters based on the updated travel time residuals. A displacement jump may include a displacement change exceeding 3 mm, or accumulated rainfall exceeding a preset value, such as 10 mm.

[0099] In some embodiments, parameter optimization of the initial wave velocity model includes soil deformation correction optimization and pore water pressure optimization, as shown by the corresponding correction formulas v. p' =v p ×(1-0.05·Δd), v s' =v s ×(1-0.02·R); where Δd is the displacement change in mm, and R is the daily rainfall in mm; v p v p' The propagation velocities of P-waves in the rock mass before and after correction are v0 and v1, respectively. s v s' These represent the propagation velocities of the S-wave in the rock mass before and after the correction, respectively.

[0100] In some embodiments, soil deformation correction is triggered in response to a sudden displacement jump, and pore water pressure optimization is triggered if accumulated rainfall exceeds a preset value. This allows for updating the intrinsic parameters of the objective function based on real-time data, ensuring model synchronization with the environment.

[0101] In some embodiments, the multi-source correlation modeling includes: determining a spatiotemporal correlation matrix based on the rainfall data, ambient temperature and humidity, slope displacement data, and optimal source parameters, using spatiotemporal correlation analysis; determining a physical correlation model based on the spatiotemporal correlation matrix, the optimal source parameters, the rainfall data, ambient temperature and humidity, and slope displacement data; and determining slope stability characteristic indicators based on the spatiotemporal correlation matrix and the physical correlation model.

[0102] In some embodiments, the spatiotemporal correlation matrix includes parameters such as displacement point ID, number of associated events, displacement change, and cumulative energy, which can be determined based on spatiotemporal correlation analysis. In some embodiments, the spatiotemporal correlation matrix also includes the frequency N of microseismic events and the energy release rate dE / dt.

[0103] In some embodiments, the processor can determine the parameters of the physical correlation model based on the displacement-energy empirical formula and the slip surface attitude inversion, thereby determining the physical correlation model. The parameters of the physical correlation model may include the slip surface attitude, energy coefficient α, etc.

[0104] For example, the cumulative energy ΣE and cumulative rainfall R of the associated events obtained based on the spatiotemporal correlation matrix are input into the displacement rate calculation formula: displacement rate V=α×ΣE+β×R, to quantify the physical relationship between displacement and micro-seismic events / rainfall, where α and β are preset coefficients, such as α=0.003mm / J, β=0.15mm / mm.

[0105] For example, the processor can obtain the slip surface orientation (including strike φ / dip angle δ) based on the source cluster coordinates and GNSS displacement vector direction through point cloud fitting and directional constraint processing. The source cluster coordinates are determined based on the locations of multiple microseismic events. When fitting the point cloud, the source cluster can be fitted into a plane using the least squares method. When constraining the direction, it can be ensured that the angle between the normal of the fitted plane and the GNSS displacement direction is <15°.

[0106] In some embodiments, the processor can process the spatiotemporal correlation matrix, displacement time series, physical model parameters, and environmental thresholds (such as critical rainfall Rc = 50 mm / 24 h) to obtain the instability probability, which serves as a characteristic index of slope stability.

[0107] For example, feature extraction is performed on the aforementioned data to determine instability characteristics. The extracted features include displacement acceleration *a*, which characterizes the degree of slope deformation acceleration; event cluster intensity *Nt*, which characterizes the activity of rock mass fracturing; and energy release rate *Pe*, which characterizes the rate of fracturing expansion, such as *a = d*. 2 S / dt 2 ; Nt=ΔN / Δt; Pe=dE / dt.

[0108] The processor can further process the instability features based on a machine learning model to obtain the instability probability. The instability features of the input model can be represented based on the feature vector [a, Nt, Pe].

[0109] In some embodiments, the machine learning model can be trained using multiple first training samples with a first label. For example, multiple first training samples with a first label can be input into an initial model, a loss function can be constructed using the first labels and the results of the initial model, and the parameters of the initial model can be iteratively updated based on the loss function using gradient descent or other methods. When a preset condition is met, the model training is complete, and a trained model is obtained. The preset condition may be that the loss function converges, the number of iterations reaches a threshold, etc.

[0110] The first training sample may include sample instability features, which can be obtained from historical data; the first label includes the sample instability probability corresponding to the sample instability feature, which can be obtained through manual annotation. For example, in historical data, the processor can obtain the sample instability features, and technicians can measure the number of times instability occurs among the historical occurrences of the sample instability features, and use the ratio of the number of times instability occurs to the total number of occurrences as its sample instability probability.

[0111] Step 203: Determine the disaster response command based on the slope stability characteristic index.

[0112] In some embodiments, the decision module is further configured to: determine a disaster response instruction based on the slope stability characteristic index, signal segmentation delay processing, and early warning rule judgment, wherein the disaster response instruction includes an optimized equipment deployment scheme and an equipment maintenance scheduling scheme.

[0113] In some embodiments, the processor can perform signal segmentation and delay processing based on an online solution engine. In some embodiments, the warning rules include triggering a yellow warning when a sudden jump in horizontal displacement exceeds a preset value (e.g., 3 mm) or when the frequency of microseismic events in the same area exceeds a preset frequency (e.g., 3 times / hour). In some embodiments, the warning rules include triggering a yellow warning when vertical displacement acceleration exceeds a preset acceleration value (e.g., 2 mm / h) or when the source energy suddenly increases to a preset energy (e.g., 10). 3 A red alert is triggered when J).

[0114] It should be noted that the above description of process 200 is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to process 200 under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0115] Example 1

[0116] This embodiment discloses a joint monitoring system based on a high slope stability monitoring information platform, such as... Figure 4 The diagram shows the system topology of the joint monitoring system. The joint monitoring system includes three key pieces of equipment: a universal GNSS monitoring station, a millimeter-level surface displacement ranging monitoring station, and a microseismic underground rock mass fracture monitoring station. Various monitoring devices deployed in the geological disaster monitoring area upload data collected by transmitters to their respective cloud platforms via 4G mobile communication. The data collection and management platform's database software processes the data, performing parsing, tagging, and synthesis before writing the data into the corresponding databases. The data collection and management platform can also read telemetry data from the databases, display real-time telemetry data in reports, and perform statistical analysis and report printing functions.

[0117] In some embodiments, the universal GNSS monitoring station can be the DM-GNSSA300 model, which can achieve millimeter-level deformation monitoring accuracy, can monitor the instantaneous deformation of the monitored body in real time, can monitor the displacement of the monitored body in both horizontal and vertical directions, and can effectively analyze the displacement deformation trend of the monitored body to achieve the purpose of prevention and early warning.

[0118] In some embodiments, the universal GNSS monitoring station has a high-precision deformation monitoring data processing algorithm, which can obtain millimeter-level real-time deformation monitoring accuracy, enabling users to grasp the instantaneous deformation of the monitored object in real time, thus meeting the user's need to grasp the instantaneous deformation of the monitored object in real time.

[0119] In some embodiments, the universal GNSS monitoring station uses GNSS technology to continuously, automatically, and in real time collect monitoring data of the monitored object, sends it to the data processing center through a communication link for high-precision data processing and analysis, and then uploads the data results to a web server, so that users can understand the real-time deformation of the monitored object anytime and anywhere through terminals such as computers and mobile phones.

[0120] In some embodiments, the millimeter-level surface displacement ranging and monitoring station can be the JW-UWB500-MM model, based on electromagnetic field and microwave principles, integrating multiple positioning technologies such as ultra-wideband (UWB), Bluetooth (BT), and accelerometers to achieve high-precision ranging and displacement monitoring. Simultaneously, the system integrates wireless data transmission, real-time monitoring of distance, location, network, and battery status information for all users in the system, and provides rapid and timely alarm functions. This data can be viewed in real-time via computer or mobile phone, offering convenience and speed; it can also be uploaded to a designated server in an encrypted manner, ensuring stability and reliability. The system's measurement accuracy is close to that of traditional optical methods, and it has advantages such as fast monitoring frequency, high accuracy, low power consumption, real-time viewing, and alarm functions. The electromagnetic waves used can penetrate and diffract generally non-metallic obstructions (flowers, trees, nets, etc.) and are unaffected by severe weather (fog, rain, snow, sandstorms), offering significant advantages. The overall system is divided into a ranging hardware layer, a data transmission and analysis layer, and a backend application layer.

[0121] The ranging hardware consists of three parts: a reference point, measurement points, and a central control point. The reference point is placed in geologically stable areas such as bedrock and bushes near the landslide. The measurement point is placed at an appropriate location on the landslide, forming a three-dimensional network with the reference point to obtain its own three-dimensional position information. Ultra-wideband positioning, inertial navigation modules, and wireless data transmission modules are used to monitor its displacement changes in real time. The central control station has flexible placement and can communicate with the reference point and measurement point in real time in three directions, providing control. The reference point and monitoring point can be flexibly combined according to the monitoring target and environmental conditions to achieve real-time monitoring of multiple monitoring points. Each hardware device is powered by a built-in battery, with ultra-low power consumption, and can work continuously for 3 months without an external power supply. An external solar power board is provided for long-term uninterrupted continuous operation. Data transmission can default to GPRS wireless transmission, directly interfacing with the cloud platform without the need for data cables, facilitating outdoor installation. It can also be modified to use Wi-Fi, LoRa, and wired transmission as needed. The backend application layer is generally deployed in the cloud for easy integration and access, enabling functions such as device management, displacement and settlement monitoring, etc. The system can provide an open API interface and SDR development tools, enabling integration with the customer's existing system.

[0122] In some embodiments, the microseismic underground rock mass fracture monitoring station mainly consists of a data acquisition host, a seismic detector, a solar panel, and a mounting bracket. Multiple devices are installed on the mountainside to be monitored in the field to form a microseismic monitoring array. The array records seismic wave signals acquired by the seismic detector sensors in real time. To balance instrument power consumption and communication bandwidth, when the newly added data reaches 2MB, the device automatically activates the 4G communication module to upload the new data to a remote cloud platform via the 4G mobile network. Personnel indoors can download all the data from the cloud platform to their local machine via the internet. Offline data processing is then performed on the raw data acquired by the field array to calculate the occurrence time, spatial location, and relative energy of underground microseismic events (i.e., the location of underground rock mass fractures) on the monitored slope.

[0123] Example 2

[0124] This embodiment specifically discloses how to achieve microseismic signal identification and time-of-arrival pickup based on Conv-LSTM-Unet and microseismic positioning based on simulated annealing.

[0125] The Conv-LSTM-Unet algorithm combines the advantages of CNN, LSTM, and Unet networks, making it particularly suitable for processing time series data. CNN is responsible for feature extraction, capturing basic information such as waveform amplitude and frequency; LSTM processes long-term dependencies in the sequence data, capturing the characteristics of signal changes over time; and Unet is used for semantic segmentation of time series data, classifying microseismic signals into categories such as noise, P-waves, and S-waves. The slope microseismic monitoring technology based on the Conv-LSTM-Unet algorithm first utilizes wavelet denoising. The signal is decomposed using wavelet basis functions selected according to the microseismic signal frequency, and then reconstructed after denoising. Next, a bandpass filter is used to set the passband frequency according to the frequency range, filtering out environmental and instrument noise and converting the signal into a two-dimensional matrix. Then, basic and multi-scale features of the microseismic signal are extracted using convolutional kernels of different sizes in a CNN layer. Temporal features are then mined through a Conv-LSTM layer. The extracted features are input into the Unet decoder, and the feature maps are fused through skip connections to improve segmentation accuracy, dividing the microseismic signal into noise, P-waves, and S-waves. Finally, Gaussian smoothing filtering is used to remove isolated noise and small fluctuations, combined with morphological operations (opening operations to remove small noise blocks and closing operations to fill holes) for optimization. Finally, waveform parameters are statistically analyzed according to actual needs to provide support for slope stability assessment.

[0126] The principle of simulated annealing is derived from the de-cooling process of metallic materials in physics. It obtains the global optimal solution to the objective function by simulating the cooling process of high-temperature metallic materials, making it a direct inversion method for solving nonlinear problems. It views the solution space of the optimization problem as the atomic state space of the material, and the objective function value as an analogy to the energy states of atoms. The algorithm starts at a relatively high "temperature" and gradually decreases the temperature during iteration to control the parameters, progressively focusing the search on a more optimal region, ultimately obtaining an optimal solution.

[0127] In statistical physics, the probability density of the energy state of matter at temperature T, where the energy is E(m), can be expressed as:

[0128]

[0129] In the formula: E(m) is the total internal energy of the substance, k is the Boltzmann constant, and T is the absolute temperature.

[0130] In some embodiments, when the energy of the new state is lower than the current state, the particle will accept the new state with probability 1; while when the energy of the new state is higher than the current state, the particle will accept the new state with a certain probability, which decreases as the temperature decreases. This study uses a perturbation function, namely the Very Fast Simulated Annealing (VFSA) algorithm, as shown in the following formula: T is the temperature, and R is a random number between 0 and 1.

[0131] In some embodiments, the core of VFSA is to iteratively optimize the location of microseismic sources and minimize the objective function. It improves the temperature update strategy and perturbation function, demonstrating stronger anti-interference capabilities in the high slopes of Deda Township. The algorithm's local fine-search capability can accurately determine the stability and potential hazardous areas of the slope and effectively process multi-source data. In the slope study of the high slopes in Deda Township, based on the geological characteristics of the rock slope, the temperature update strategy was set to T = 0.95T, and a three-dimensional wave velocity model was established, assigning wave velocity values ​​according to geological characteristics. Traditional slope source location models typically treat slopes as static media, but in reality, slopes undergo complex dynamic changes under seismic action. Therefore, the influence of dynamic parameters such as soil deformation and pore water pressure changes is considered in the model.

[0132] In microseismic location based on travel time, the time difference between the seismic wave arrival at different detectors can be expressed as:

[0133] Therefore, given the coordinates of each geophone, to locate the seismic source, it is necessary to know the arrival time of the seismic waves at different geophones and the velocity model of the survey area, and then deduce the coordinates of the seismic source and the time of origin. The residual values ​​are continuously corrected through iterative updates until the conditions are met.

[0134] Example 3

[0135] This embodiment specifically discloses a real-time monitoring and early warning data receiving system.

[0136] In some embodiments, field monitoring equipment transmits data to a server via GPRS, CDMA, or BeiDou satellite communication networks. The data is then received and stored in a database through a data receiving platform. After storage, the data is displayed on a platform using text and graphs to facilitate data querying and analysis. Once on the platform, users can view monitoring information from all stations within the work area. This includes viewing data analysis charts for that type of station, the latest uploaded data, querying historical data for any time period, equipment operating status, and various basic information about the station. A cloud data management platform allows users to view monitoring information from all stations within the work area. Figure 5 The diagram shows a cloud platform for millimeter-level displacement monitoring data.

[0137] In some embodiments, by logging into the remote cloud data management platform through FTP client software, one can view the raw seismic wave data collected and uploaded by all microseismic underground rock mass fracture monitoring stations within the work area.

[0138] Example 4

[0139] Project Application Results (Batang Station Demonstration Area)

[0140] Figure 6 The diagram shows the layout of the monitoring points in the high and steep slope deformation monitoring demonstration area of ​​Batang Station on the Sichuan-Tibet Railway. Figure 7 A three-dimensional top view showing the distribution of landslide bodies on both sides of Batang Station.

[0141] In landslide No. 1 (area 316,000 m²) 2 ) and slope body #3 (area 239,000 m²) 2 22 monitoring points were established to achieve three-dimensional monitoring of "surface displacement (3 sets of GNSS) - shallow sliding (25 sets of displacement stations) - deep rupture (20 sets of microseismic stations)".

[0142] Based on the temperature and rainfall data of the work area, from early June to September 11, 2021, the work area gradually entered the rainy season. Both slopes #3 and #5 showed a trend of gradually increasing surface displacement. The maximum horizontal displacement of slope #3 reached 5 mm, and the maximum vertical displacement reached 4 mm. Figure 9 The maximum horizontal displacement of slope #5 is 7.0 mm, and the maximum vertical displacement is 3 mm. Figure 10After September 11th, the dry winter season began, with significantly reduced rainfall. The horizontal and vertical displacements of slopes #3 and #5 returned to their early June positions. It should be further noted that slopes #3 and #5 exhibited similar deformation fluctuations over the same period, from September 1st to September 30th, 2021. Figure 8 This is a GNSS displacement monitoring result image of slope #3. Figure 9 This is a GNSS displacement monitoring result image of slope #5. Figure 10 Comparison chart of GNSS displacement monitoring data for slopes #3 and #5.

[0143] Through the consistency test of seismic detectors, the waveform correlation coefficient of 20 devices was >0.98. The root mean square error of GNSS monitoring data compared with manual measurement by total station was <2mm, which meets the requirements of the "Code for Geological Investigation of Railway Engineering". Figure 11 The results are from the consistency test of the seismic detector. Figure 12 This is the result of the consistency test of the microseismic monitoring equipment before it is deployed to the field.

[0144] The technical solution presented in this application achieves multi-dimensional technological breakthroughs in the field of microseismic monitoring, significantly improving system performance indicators. Compared to traditional methods, the microseismic signal recognition rate increases from 75% to 95%, and the detection accuracy improves by 20 percentage points, effectively enhancing the ability to identify potential risks. Regarding source location accuracy, the location error is drastically reduced from over 10 meters to within 5 meters, improving accuracy by 50%, providing a reliable guarantee for accurate disaster source identification. The system's real-time performance has achieved a qualitative leap, with data processing latency shortened from over 60 minutes to within 10 minutes, improving processing efficiency by 83%; the early warning response time is further reduced from over 30 minutes to within 1 minute, improving response speed by 97%, establishing a minute-level closed loop from data acquisition to early warning issuance. This series of technological improvements forms a complete technical chain of "high-precision identification - rapid calculation - real-time early warning," providing strong technical support for geological disaster monitoring and early warning. Figure 13 A three-dimensional map showing the location of the epicenter of a microseismic event on a high slope in Deda Township.

[0145] The table below shows the improvements of this approach compared to traditional methods: index Traditional methods This invention Increase Microseismic signal recognition rate 75％ 95％ +20％ Seismic source location error ＞10m ＜5m 50％ Data processing delay >60min <10min 83％ Early warning response time >30min <1min 97％

[0146] Example 5

[0147] This invention was successfully applied in the Sichuan Provincial Key Laboratory project "Intelligent Monitoring Technology for Internal Stability of Rock Mass Based on Microseismic Earthquakes and Its Application Demonstration" (Project No.: 2024KFKT002) from 2024 to 2025. During the specific implementation of the project, to address the stability monitoring needs of the steep slopes (Slope No. 3 and Landslide No. 1) on both sides of Batang Station on the Sichuan-Tibet Railway (Zhongdeda Village, Deda Township, Batang County, Ganzi Prefecture, Sichuan Province, geographical coordinates: 99°24′25.23″E, 30°16′26.23″N), an integrated monitoring system of "surface displacement-underground microseismic earthquakes" was constructed. This system enabled monitoring of the stability of the large rock slope (Slope No. 3 has a volume of 4.78 million m³). 3 The volume of landslide #1 is 6.32 million cubic meters. 3 Real-time monitoring of deep fracturing processes and surface deformation provides disaster early warning support for the construction and operation of the Sichuan-Tibet Railway.

[0148] Figure 14 This is a layout diagram of the general-purpose GNSS monitoring equipment used in this project. The specific application method in this project is as follows: Based on field data collection, potential fracture zones were identified in slope #3 (main sliding direction 32°, relative elevation difference 380m) and landslide #1 (main sliding direction 300°, relative elevation difference 296m). The bedrock is Triassic limestone interbedded with sandstone and slate, with a dip of 28°∠40°. Slope #3 and landslide #1 on both slopes have a significant impact on the safety of Batang Station.

[0149] The GNSS monitoring station reference point was set at the bottom of the gully on stable bedrock (99°24′2.84″E, 30°16′56.50″N), avoiding the river erosion area. Monitoring points were set at the middle of slope #3 (GNSS-G1, 99°24′24.31″E, 30°16′11.11″N) and slope #5 (GNSS-G2, 99°23′56.33″E, 30°16′37.49″N), with each point meeting the requirement of unobstructed 45° elevation angle. Simultaneously, 10 sets of microseismic monitoring equipment were deployed on both landslide #1 and slope #3. The monitoring stations (20 sets in total) are installed on the same pole as the millimeter-level displacement stations, forming a 3×3 array layout with an average spacing of 200m (typical location such as 3#-1, 99°24′23.5852″E, 30°16′16.8845″N), covering the front and middle edges of the slope and also taking into account deep fracture monitoring. In addition, 5 millimeter-level displacement fixed points (numbered 1-3, etc.) are set up in the stable area around the slope to construct a non-coplanar triangular monitoring network (the height difference between adjacent points is ≥50m), ensuring that the accuracy of three-dimensional displacement calculation reaches a horizontal error of <1cm and a vertical error of <3cm, forming a space-surface collaborative monitoring network.

[0150] Systematic verification ensures monitoring accuracy and stability. Sensor consistency testing involved 7 days of continuous testing using manual impact on the seismic source (energy 10⁻³ J, main frequency 20 Hz). The waveform correlation coefficient of 20 microseismic devices was >0.98, and the arrival error was <0.01 s. In data transmission testing, GNSS data was uploaded to the cloud platform in real-time via a 4G network (bandwidth 100 kbps) (delay <10 s). Microseismic data was transmitted via triggered FTP (47.108.83.9:22 port, initiated when the data size reaches 2 MB, transmission time <30 s, reception delay <5 min), with zero packet loss throughout. Positioning accuracy verification involved simulating seismic sources at depths of 20 m, 50 m, and 100 m using artificial blasting. Simulated annealing algorithms yielded positioning errors of 3.2 m, 4.5 m, and 6.8 m, respectively, meeting the requirement of <10 m deep monitoring error in the "Railway Engineering Geological Investigation Specification" (TB10012-2022). The overall system met the design specifications.

[0151] This specification provides one or more embodiments of a device for monitoring the internal stability of a rock slope, including a processor for executing a method for monitoring the internal stability of a rock slope.

[0152] The core algorithm of the intelligent monitoring technology for rock mass internal stability based on microseismic events is optimized through the Conv-LSTM-Unet model and simulated annealing algorithm. Conv-LSTM-Unet integrates the advantages of Convolutional Neural Networks (CNN), Long Short-Term Memory Networks (LSTM), and Unet networks. First, it preprocesses the signal using wavelet denoising and bandpass filters (10-200Hz). Then, CNN extracts multi-scale time-frequency features, Conv-LSTM captures temporal correlations, and the encoder-decoder structure of Unet achieves accurate segmentation of P-waves, S-waves, and noise. Gaussian filtering and morphological operations further optimize the results. The simulated annealing algorithm employs a cooling strategy (T=0.95T) and a travel time residual objective function. Combined with a three-dimensional wave velocity model and dynamic parameters (soil deformation, pore water pressure), it improves global search efficiency and enhances anti-interference capabilities through VFSA (Very Fast Simulated Annealing), achieving high-precision source localization. The two types of algorithms work together to construct an "integrated ground-deep array monitoring technology", which, combined with GNSS surface displacement data, effectively reveals the spatiotemporal patterns of rock mass fracturing and provides reliable support for early warning of high-altitude rock landslides.

[0153] All field monitoring equipment transmits data to the server via GPRS, CDMA, and BeiDou satellite communication networks. The data is then received and stored in the database through a data receiving platform. After storage, the data is displayed on a platform using text and graphs to facilitate data querying and analysis. Users can view data analysis charts for this type of station, the latest uploaded data, query historical data for any time period, equipment operating status, and various basic information about the station.

[0154] The three-dimensional digital model of the high slopes on both sides of Zhongdeda Village is as follows: Figure 13 As shown, an array of 20 microseismic monitoring stations can detect microseismic events within a 2-kilometer radius. A very fast simulated annealing algorithm is used to calculate the occurrence time, spatial location, and energy magnitude of microseismic events. Taking 45 microseismic location data points recorded and processed over a three-month period from June to September 2021 as an example, microseismic events occurring within the No. 1 slope were relatively frequent.

[0155] This paper studies the application of the Conv-LSTM-Unet algorithm in intelligent identification of microseismic events, achieving accurate identification of microseismic signals triggered by internal rock rupture and sliding. Using the simulated annealing algorithm principle and the travel-time residual method to establish the objective function, a cooling strategy of T = 0.95T is set. The location of underground microseismic events on the high slope of Deda Township is spatiotemporally located, and a feasible, effective, and real-time online microseismic monitoring and rock mass stability assessment system is preliminarily constructed.

[0156] This application proposes and implements an "integrated ground and deep array monitoring technology" that combines surface displacement data with the location of internal rock mass rupture events. The use of array monitoring technology can help improve the accuracy of monitoring unstable processes inside the rock mass, and is expected to solve the current problem of monitoring and early warning of hidden, sudden, and massive high-altitude rock landslides.

[0157] In summary, the technical solution of this application provides a high-precision, high-reliability, and low-cost solution for the prevention and control of rock slope disasters. It breaks through the dimensional limitations, accuracy bottlenecks, and delay defects of traditional monitoring methods and has irreplaceable application value in major projects.

[0158] This specification provides one or more embodiments of a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions from the storage medium, the computer executes a method for monitoring the internal stability of a rock slope.

[0159] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0160] Furthermore, this specification uses specific terms to describe embodiments thereof. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Moreover, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0161] Furthermore, unless expressly stated in the claims, the order of elements and sequences, the use of numbers and letters, or other names in this specification are not intended to limit the order of the processes and methods described herein. Although various examples have been discussed in the foregoing disclosure of some embodiments of the invention that are currently considered useful, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the spirit and scope of the embodiments described herein. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely using software solutions, such as installing the described system on an existing server or mobile device.

[0162] Similarly, it should be noted that, in order to simplify the description disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments in this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0163] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed depending on the characteristics required by individual embodiments. In some embodiments, numerical parameters should take into account specified significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of range in some embodiments of this specification are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0164] For each patent, patent application, patent application publication, and other material, such as articles, books, specifications, publications, and documents, referenced in this specification, the entire contents of which are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this specification, as well as documents that limit the broadest scope of the claims in this specification (currently or subsequently appended to this specification). It should be noted that in the event of any inconsistency or conflict between the descriptions, definitions, and / or terminology used in the supplementary materials and the content of this specification, the descriptions, definitions, and / or terminology used in this specification shall prevail.

[0165] Finally, it should be understood that the embodiments in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments in this specification are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments in this specification are not limited to those explicitly described and illustrated herein.

Claims

1. A method for monitoring the internal stability of a rock slope, characterized in that, include: A multidimensional sensing network is constructed to process the collected slope geological structure and environmental parameters to obtain a spatiotemporally aligned multi-source dataset. At least one of the following methods is performed on the multi-source dataset: signal recognition processing, rupture evolution inversion, and multi-source correlation modeling, to obtain slope stability characteristic indicators; Based on the slope stability characteristic indicators, disaster response instructions are determined.

2. The method for monitoring the internal stability of a rock slope as described in claim 1, characterized in that, The multidimensional sensing network includes: The surface displacement monitoring module is configured to acquire slope displacement data; The underground fracture monitoring module is configured to acquire rock mass fracture signals; The environmental parameter monitoring module is configured to acquire rainfall data and ambient temperature and humidity. The data transmission and preprocessing module is configured to perform edge preprocessing and transmission on the acquired data.

3. The method for monitoring the internal stability of a rock slope as described in claim 1, characterized in that, The signal recognition processing includes: Based on the original microseismic waveforms, equipment spatial coordinates, and environmental noise baselines from the multi-source dataset, P / S wave arrival time markers are obtained through wavelet denoising, bandpass filtering, CNN feature extraction, Conv-LSTM time series modeling, Unet semantic segmentation, and post-processing optimization.

4. The method for monitoring the internal stability of a rock slope as described in claim 3, characterized in that, The fracture evolution inversion includes: Acquire travel time data, which includes the P / S wave arrival time marker, the device spatial coordinates, and the initial wave velocity model; Construct an objective function, and determine the travel time residuals based on the travel time data using the objective function; The optimal source parameters are obtained by iterative optimization based on the travel time residuals.

5. The method for monitoring the internal stability of a rock slope as described in claim 4, characterized in that, The fracture evolution inversion also includes: In response to the detection of a sudden displacement jump or rainfall reaching a preset condition, the parameters of the initial wave velocity model are optimized to obtain updated travel time data. The travel time residuals are then updated based on the updated travel time data, and the optimal source parameters are updated based on the updated travel time residuals.

6. The method for monitoring the internal stability of a rock slope as described in claim 3, characterized in that, The multi-source association modeling includes: Based on the rainfall data, ambient temperature and humidity, slope displacement data, and optimal seismic source parameters, a spatiotemporal correlation matrix is ​​determined based on spatiotemporal correlation analysis. Based on the spatiotemporal correlation matrix, the optimal source parameters, the rainfall data, the ambient temperature and humidity, and the slope displacement data, a physical correlation model is determined. Based on the spatiotemporal correlation matrix and the physical correlation model, slope stability characteristic indicators are determined.

7. The method for monitoring the internal stability of a rock slope as described in claim 6, characterized in that, The determination of disaster response instructions based on the slope stability characteristic indicators includes: Based on the slope stability characteristic indicators, and based on signal segmentation delay processing and early warning rule judgment, disaster response instructions are determined. The disaster response instructions include optimized equipment deployment schemes and equipment maintenance scheduling schemes.

8. A monitoring system for the internal stability of a rock slope, characterized in that, include: The perception module is configured to construct a multi-dimensional perception network to process the collected slope geological structure and environmental parameters to obtain a spatiotemporally aligned multi-source dataset. The analysis module is configured to perform at least one of the following on the multi-source dataset: signal recognition processing, rupture evolution inversion, and multi-source correlation modeling, to obtain slope stability characteristic indicators. The decision-making module is configured to determine disaster response instructions based on the slope stability characteristic indicators.

9. A computer-readable storage medium storing computer instructions, wherein when a computer reads the computer instructions in the storage medium, the computer performs the method as described in any one of claims 1-7.

10. A monitoring device for the internal stability of a rock slope, the device comprising at least a processor and at least one memory; the at least one memory being used to store computer instructions; the at least one processor being used to execute at least a portion of the computer instructions to implement the method as described in any one of claims 1 to 7.

Images