Beidou-based side slope geological disaster monitoring and early warning method and system

Abstract

The invention relates to the technical field of side slope geological disaster monitoring and early warning, and discloses a Beidou-based side slope geological disaster monitoring and early warning method and system.The Beidou-based side slope geological disaster monitoring and early warning method comprises the steps that multi-scale geological signals based on Beidou system time service are acquired, the multi-scale geological signals comprise environmental noise signals, acoustic emission signals and micro-seismic signals; processing the multi-scale geological signals, and extracting environmental noise interference characteristics representing the change of a microstructure in the slope, acoustic emission signal characteristics representing the microscopic damage process of a material and micro-seismic signal characteristics representing the deformation of a macrostructure; based on the environmental noise interference features, the acoustic emission features and the micro-seismic features, constructing a slope instability risk assessment model, and outputting a risk assessment result; performing early warning according to a risk assessment result; according to the invention, high-precision, low-cost and full-coverage monitoring and early warning can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of slope geological disaster monitoring and early warning, and more specifically, to a Beidou-based slope geological disaster monitoring and early warning method and system. Background Art

[0002] Slope geological disasters, such as landslides and collapses, pose a serious threat to the safety of life, property, and infrastructure. Before a disaster occurs, the slope rock and soil typically undergo a gradual process from the development of microscopic cracks to macroscopic deformation and failure. This process is accompanied by the generation of various physical signals such as acoustic emissions and microseismic activity.

[0003] Traditional slope monitoring methods rely on single-scale monitoring methods (such as surface deformation monitoring or macroseismic monitoring), which makes it difficult to capture the complete disaster chain and results in ineffective early warnings. While some active source detection methods (such as artificially stimulated seismic waves) can obtain underground structural information, they require high sensor deployment density, are expensive, and have limited coverage, making them difficult to implement on a large scale.

[0004] In addition, when multiple sensors are collaboratively monitored, the time synchronization accuracy is not high (for example, if it only relies on the Network Time Protocol (NTP) synchronization, the accuracy is in the millisecond to second level), which seriously affects the precise identification of weak precursor signals, the accuracy of earthquake source positioning, and the reliability of multi-source data fusion analysis. This is a key technical problem that needs to be urgently solved in the current field of slope geological disaster monitoring and early warning. Summary of the Invention

[0005] The present invention provides a Beidou-based slope geological disaster monitoring and early warning method and system, which solves the technical problems in related technologies such as single slope monitoring scale, high cost, limited coverage, and insufficient sensor time synchronization accuracy leading to inaccurate positioning and delayed early warning.

[0006] The present invention provides a Beidou-based slope geological disaster monitoring and early warning method, comprising the following steps: Acquire multi-scale geological signals based on BeiDou system timing, including environmental noise signals, acoustic emission signals, and microseismic signals; Process multi-scale geological signals to extract environmental noise interference features that characterize microstructural changes within the slope, acoustic emission signal features that characterize the microscopic material damage process, and microseismic signal features that characterize macrostructural deformation. Based on the environmental noise interference characteristics, acoustic emission characteristics and microseismic characteristics, a slope instability risk assessment model is constructed and the risk assessment results are output; Based on the risk assessment results, a multi-level warning threshold system is established. When the risk assessment results reach or exceed the preset threshold, the corresponding level of warning is triggered.

[0007] In a preferred embodiment, the step of extracting the environmental noise interference characteristics that characterize the microstructural changes inside the slope includes: The cross-correlation function is used to process the environmental noise signal and extract the Green's function of the underground medium; A time differential interferometry monitoring algorithm is constructed to identify small changes in the internal structure of the slope and obtain the interference characteristics of environmental noise by calculating the differences in the Green's functions of the underground medium obtained at different times.

[0008] In a preferred embodiment, the step of processing the multi-scale geological signals to extract the acoustic emission signal features and the microseismic signal features representing the macroscopic structural deformation includes: The mixed signal of acoustic emission signal and microseismic signal is decoupled by blind source separation algorithm to separate independent acoustic emission signal and microseismic signal.

[0009] In a preferred embodiment, the step of processing the multi-scale geological signals to extract microseismic signal features further includes: Wavelet packet transform and adaptive threshold algorithm are used to extract the arrival time, amplitude, duration and main frequency of the separated microseismic signals as microseismic features.

[0010] In a preferred embodiment, the step of constructing a slope instability risk assessment model includes: Construct an acoustic emission and microseismic signal correlation analysis model, combine acoustic emission characteristics, microseismic characteristics and environmental factors, and output a comprehensive risk index; Through the improved double-difference positioning method, the arrival time data of acoustic emission signals and microseismic signals are used to achieve earthquake source positioning.

[0011] In a preferred embodiment, the step of constructing a slope instability risk assessment model further comprises: A model of the evolution of microscopic damage to macroscopic instability is constructed, and the probability of instability is predicted by combining microscopic damage parameters characterized by acoustic emission characteristics, mesoscopic damage parameters characterized by microseismic characteristics, and environmental factors.

[0012] In a preferred embodiment, the step of providing an early warning includes: Based on the sudden changes in acoustic emission characteristics, the activity trend of microseismic characteristics, the comprehensive risk index and the probability of instability, a multi-parameter early warning model is constructed, and an early warning is triggered when the monitoring parameters reach the preset threshold.

[0013] In a preferred embodiment, a Beidou-based slope geological disaster monitoring and early warning method further includes: A low-power edge computing processing unit is configured for real-time processing of multi-scale geological signals and anomaly detection in early warning.

[0014] In a preferred embodiment, a Beidou-based slope geological disaster monitoring and early warning method further includes: A three-dimensional visualization system for multi-scale slope damage is constructed to visualize the earthquake source location results, slight changes in the slope's internal structure, and instability risk assessment results in three dimensions.

[0015] In a preferred embodiment, a Beidou-based slope geological disaster monitoring and early warning system is used to implement a Beidou-based slope geological disaster monitoring and early warning method, including: The signal acquisition module is used to obtain multi-scale geological signals based on BeiDou system timing. The multi-scale geological signals include environmental noise signals, acoustic emission signals, and microseismic signals. The signal processing module is used to process multi-scale geological signals and extract the environmental noise interference characteristics that characterize the microstructural changes within the slope, the acoustic emission signal characteristics that characterize the microscopic damage process of the material, and the microseismic signal characteristics that characterize the macrostructural deformation; The risk assessment module builds a slope instability risk assessment model based on environmental noise interference characteristics, acoustic emission characteristics, and microseismic characteristics, and outputs risk assessment results; Early warning module, used to issue early warnings based on risk assessment results; The visualization module is used to display risk assessment results and warning information in three-dimensional visualization.

[0016] The beneficial effects of the present invention are: Multi-scale monitoring capabilities have been achieved: through the synchronous collection of acoustic emission and microseismic signals through Beidou high-precision timing, combined with environmental noise interference technology, it is possible to capture the complete disaster process from micro-cracks to macro-deformation. Compared with traditional single-dimensional monitoring, it significantly improves the comprehensiveness and foresight of monitoring, and can advance the effective warning time.

[0017] Improved monitoring cost-effectiveness and coverage: By utilizing the principle of environmental noise interference, it reduces dependence on active sources and a large number of sensors, significantly reduces system deployment and maintenance costs, and can achieve full coverage monitoring of the internal structure of the slope.

[0018] Improved positioning accuracy and signal detection sensitivity: Based on Beidou's nanosecond timing, high-precision synchronization of multi-sensor data is ensured, thereby significantly improving the source positioning accuracy of acoustic emissions and microseismic events, as well as the identification resolution of small changes in internal structures; through multi-source signal processing and fusion analysis, the signal detection threshold is effectively lowered, and weaker precursor signals can be captured.

[0019] Enhanced the accuracy and reliability of early warning: By constructing a multi-parameter fusion risk assessment model and an instability evolution law model, and comprehensively analyzing multi-scale monitoring information, the accuracy of disaster prediction is improved and the false alarm rate is effectively reduced; combined with edge computing and three-dimensional visualization technology, the timeliness of early warning and the intuitiveness of decision-making are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of a Beidou-based slope geological disaster monitoring and early warning method of the present invention. DETAILED DESCRIPTION

[0021] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. Furthermore, features described in some examples may be combined in other examples.

[0022] At least one embodiment of the present invention discloses a method for monitoring and early warning of slope geological disasters based on Beidou, such as Figure 1 As shown, the following steps are included: Step 1: Acquire multi-scale geological signals based on BeiDou system timing. The multi-scale geological signals include environmental noise signals, acoustic emission signals, and microseismic signals. The specific steps include: Step 1.1, deploy the Beidou timing receiver; Select a suitable location in or around the monitoring area to install the Beidou timing receiver, and ensure that it can stably receive Beidou satellite signals and output high-precision synchronous clock signals.

[0023] Step 1.2, arranging a broadband sensor array; According to the geometric dimensions and geological structural characteristics of the monitored target slope, a sensor array consisting of multiple broadband sensors (e.g., geophones, accelerometers, and acoustic emission probes) is arranged on the slope surface or through shallow drilling.

[0024] In some embodiments, other types of sensors such as fiber grating sensors and inclinometers may also be included to obtain richer monitoring data.

[0025] The frequency response range of the sensor needs to cover the expected frequency band of the target signal, for example, 1 Hz to 300 kHz, to achieve simultaneous capture of low-frequency microseismic and high-frequency acoustic emission signals.

[0026] Step 1.3, realize time synchronization of sensor network; The synchronous clock signal output by the Beidou timing receiver is accurately distributed to the data acquisition unit connected to each sensor in the sensor array.

[0027] Hardware synchronization solutions or software correction algorithms can be used to ensure that the timestamp accuracy of the data collected by each sensor reaches the 10 nanosecond (ns) level.

[0028] This high-precision time synchronization is a necessary prerequisite for the subsequent successful application of environmental noise interference models to extract Green's functions and conduct multi-signal joint positioning analysis.

[0029] Step 1.4, collecting multi-scale geological signals; The sensor network system is activated to continuously or according to a preset period collect the environmental background noise of the monitoring area, the microseismic signals caused by internal geological activities, and the high-frequency acoustic emission signals generated by the microscopic fracture of the material.

[0030] The collected raw waveform data and its corresponding high-precision timestamp information are stored for subsequent data processing and analysis.

[0031] Step 2: Process the multi-scale geological signals to extract the environmental noise interference characteristics that characterize the microstructural changes within the slope, the acoustic emission signal characteristics that characterize the microscopic damage process of the material, and the microseismic signal characteristics that characterize the macrostructural deformation; The specific steps include: Step 2.1, extracting the Green's function of the underground medium using the cross-correlation function; For the collected long-term continuous environmental noise data (input data type), select the signals recorded by two sensors at different locations (for example, sensor A and sensor B) By calculating the cross-correlation function between them, the empirical Green's function of the underground medium between the two sensors (specific output result) is extracted.

[0032] The cross-correlation function can be specifically expressed as: ; in, represents the time delay, that is, the time offset applied to signal B when calculating the cross-correlation function; and They represent the start and end time of the integration period, respectively, and define the time window range for calculating the cross-correlation; represents the time series signal recorded by the sensor at position A; Represents the time series signal recorded by the sensor at position B and with a time delay applied It represents the cross-correlation function between the signals recorded by sensors at positions A and B, reflecting the wave field propagation characteristics between the two points.

[0033] Through this method, passive environmental noise data can be used to estimate the propagation characteristics of the underground wave field, providing key basic data for subsequent change detection analysis.

[0034] Step 2.2: Apply blind source separation algorithm to decouple the mixed signal of acoustic emission and microseismic; For mixed broadband signal data containing both acoustic emission signals and microseismic signals, blind source separation algorithm can be used for processing.

[0035] For example, independent component analysis (ICA) or principal component analysis (PCA) may be used, or in some implementations, algorithms such as non-negative matrix factorization (NMF) may also be considered.

[0036] Assume that the observed mixed signal vector is The original source signal vector is The matrix describing the signal mixing process is The ambient noise is Then the hybrid model can be mathematically expressed as: ; in, Represents the mixed signal matrix collected by the sensor, each row represents the time series record of a sensor; represents the mixing matrix, which describes how the source signals are mixed into the various sensors; Represents the source signal matrix, which includes independent source signals such as acoustic emission signals and microseismic signals; represents the noise matrix, which represents the environmental noise and instrument noise during the monitoring process. This equation is a mathematical expression of the blind source separation problem, which is used to separate the mixed multi-scale geological signals into independent acoustic emission and microseismic signal components.

[0037] Estimate the mixing matrix by applying the selected blind source separation algorithm The inverse matrix or pseudo-inverse matrix of can effectively separate the relatively independent acoustic emission signal components and microseismic signal components, which facilitates the subsequent feature extraction and refined analysis of the two signals.

[0038] Step 2.3: Use wavelet packet transform and adaptive threshold algorithm to extract microseismic signal features; The separated microseismic signals are analyzed in the time-frequency domain using the wavelet packet transform algorithm, and are effectively decomposed into sub-signals in different frequency bands.

[0039] For each sub-signal component, an adaptive threshold algorithm (for example, based on energy proportion or kurtosis index changes) is used to accurately identify and extract core characteristic parameters of microseismic events, such as the arrival time of P-wave or S-wave, maximum amplitude, effective duration, and dominant frequency.

[0040] Step 2.4, construct a time differential interferometry monitoring algorithm to identify small changes in the internal structure of the slope; Based on different period Green's functions (input data type), by calculating the different time periods (e.g. Moment and Green's function obtained at time The difference between them can be used to effectively identify the slight changes in the physical properties of the medium inside the slope (specific output results).

[0041] The time differential interferometry monitoring algorithm can be specifically expressed as: ; in, Indicates from time At the time The change in the cross-correlation function between sensors A and B during the period; Indicates the initial time The cross-correlation function between sensors A and B when ; Indicates subsequent time The cross-correlation function between sensors A and B when ; represents the time delay variable used to calculate the cross-correlation function. Significant changes can usually indicate changes in the adjustment of the internal stress state of the slope, the development or expansion of microcracks, etc.

[0042] In a specific application, for example, first in the initial state ( The environmental noise data is collected over a long period of time (such as hours or days) and the reference Green's function is obtained by cross-correlation calculation. .

[0043] During the monitoring period (e.g. At this moment, which can be several hours or days later, collect the same length of environmental noise data again and calculate the current Green's function By comparing the waveform differences of these two Green's functions (e.g., calculating the decrease in their correlation coefficient, or the change in travel time difference), the changes in the medium inside the slope can be quantified.

[0044] If a sensor pair observes If there is a significant and persistent change in the wave velocity, it indicates that the physical properties of the area passed by the path may have changed, such as changes in water content, stress concentration, or change in wave velocity caused by the expansion of microcracks.

[0045] Step 3: Based on the environmental noise interference characteristics, acoustic emission characteristics, and microseismic characteristics, a slope instability risk assessment model is constructed and the risk assessment results are output; The specific steps include: Step 3.1, constructing an acoustic emission microseismic signal correlation analysis model; Based on the extracted acoustic emission characteristic parameters (e.g., ring count, energy value), microseismic characteristic parameters (e.g., event magnitude, occurrence frequency) (input data type), and the obtained relevant environmental influencing factors (e.g., rainfall and temperature periodic changes in the monitoring area) (input data type), a comprehensive risk index assessment model (specific output results) is constructed.

[0046] The model can be specifically expressed as: ; in, Indicates The comprehensive risk index evaluation value at a specific moment is used to quantify the overall instability risk level of the slope at a specific moment; Indicates The comprehensive characterization value of the acoustic emission parameters at the moment reflects the intensity of micro-crack development activities, which is usually calculated by comprehensively calculating the number of acoustic emission events, energy size and frequency characteristics; Indicates that a specific lag time is taken into account of The comprehensive characterization value of microseismic parameters at a certain moment reflects the intensity of macroscopic deformation activity and is usually calculated by comprehensively calculating the frequency of microseismic events, magnitude, spatial distribution characteristics and other indicators. is the lag time parameter, which represents the time delay effect of microseismic activity relative to acoustic emission activity; It means that Quantitative assessment indicators of environmental factors at all times, including the combined impact of external triggering factors such as rainfall, groundwater level changes, and temperature changes; They represent the weight coefficients of acoustic emission parameters, microseismic parameters and environmental factors, respectively, reflecting the relative importance of various monitoring signals to the overall risk assessment. These coefficients can be determined by analyzing historical monitoring data and drawing on the experience of experts in the field, and they meet the requirements of To ensure the normalization of the evaluation results.

[0047] This model can be used to dynamically assess the overall risk level development trend of the slope and provide a quantitative basis for early warning decisions. The specific construction process of the acoustic emission microseismic signal correlation analysis model may include: Select characteristic parameters that can represent the level of acoustic emission activity, such as the number of events per unit time, average energy, peak amplitude, etc., and synthesize them into .

[0048] Select characteristic parameters representing the level of microseismic activity, such as the number of microseismic events per unit time, average magnitude, b-value, etc., and consider the potential time lag between them and the occurrence of acoustic emission events Comprehensive .

[0049] Environmental factors It can include hourly rainfall, daily cumulative rainfall, temperature gradient, etc.

[0050] Using historical monitoring data (including acoustic emission, microseismic and environmental data) and corresponding slope deformation or instability records, multiple regression analysis, time series analysis or machine learning methods (such as logistic regression, decision tree) are used to fit parameters. and time lag .

[0051] In practical applications, the model can output a normalized risk index, which can trigger an early warning when the index exceeds the preset threshold.

[0052] Step 3.2: Apply the improved double-difference positioning algorithm to achieve sub-meter earthquake source positioning; Using the arrival time data of high-precision synchronized acoustic emission and microseismic signals (input data type), an improved double-difference positioning algorithm is used to accurately calculate the spatial position of the earthquake source (specific output result).

[0053] The core of the double-difference positioning algorithm lies in iteratively minimizing the residual between the theoretically calculated arrival time difference and the actual observed arrival time difference, thereby simultaneously inverting the earthquake source location parameters and the velocity structure model of the underground medium, thereby significantly improving the positioning accuracy.

[0054] High-precision time synchronization is essential for achieving sub-meter earthquake source location. The improved double-difference location algorithm described here can be optimized for the specific characteristics of slope monitoring. For example, given the dramatic lateral variations in velocity structure in slope areas, a three-dimensional non-uniform velocity model can be introduced to replace traditional one-dimensional or two-dimensional uniform / layered models.

[0055] In addition, considering the differences in arrival time accuracy between high-frequency acoustic emission signals and low-frequency microseismic signals, different weights can be assigned to the arrival time residuals of different types of signals in the objective function.

[0056] In specific applications, a certain number of "master events" with relatively reliable initial positioning results are first collected. Then, the relative time differences between these master events and the event to be located are used for iterative optimization to obtain more accurate relative positioning results. For example, on a slope with a known potential sliding surface, this improved algorithm can accurately locate small rupture events detected near the sliding surface, thereby more accurately determining the risk of instability.

[0057] Step 3.3, construct an analysis model of the evolution of microscopic damage to macroscopic instability; By comprehensively utilizing the micro-scale damage parameters reflected by acoustic emission signals (e.g., crack density index, material damage rate), the meso-scale damage parameters reflected by microseismic signals (e.g., spatial distribution pattern of microfracture events, energy release rate), and related environmental influencing factors, a predictive analysis model is constructed to describe the evolution law of the slope from the micro-damage accumulation process to the macro-instability state.

[0058] The model can be expressed mathematically as: ; in, In the future The probability value of instability occurring within the time period, that is, the prediction at the current moment After The probability of geological disasters such as landslide or collapse occurring on the slope within a certain period of time; Indicates The characteristic parameters of the microscopic damage state at each moment mainly include the frequency, energy, spatial distribution of acoustic emission events and other indicators reflecting the development of microscopic cracks; Indicates The characteristic parameters of the mesoscopic damage state at the time of the earthquake mainly include the frequency, magnitude, and concentration of the earthquake source locations of microseismic events, which are indicators reflecting the deformation of larger-scale structures. It means that The comprehensive influence of environmental factors at the moment, including rainfall, groundwater level changes, temperature changes and other external environmental factors. Function The specific form of can be constructed through training through advanced machine learning algorithms, for example, support vector machine (SVM) models, various neural network models, or in some implementations, integrated learning models such as gradient boosting decision tree (GBDT) and random forest can be used, or it can be established based on more classic physical damage mechanics models.

[0059] The construction of this evolution law analysis model can specifically include: Defining micro-damage parameters For example, the density and growth rate of microcracks inside the material can be quantified based on the ring count, energy accumulation value, or b-value of the acoustic emission event.

[0060] Defining mesoscopic damage parameters For example, the formation and development of larger-scale ruptures can be characterized based on the spatial clustering characteristics of microseismic events, the rupture type reflected by the focal mechanism solution, or the stress drop in a specific area.

[0061] Environmental factors These may include accumulated rainfall, pore water pressure, temperature cycles, etc. Subsequently, historical data containing these three types of parameters and actual instability events (or significant deformation events) are collected and used to train the selected machine learning model or calibrate the evolution equation based on damage mechanics theory (such as a model based on continuum damage mechanics, which considers the evolution of damage variables with time and stress).

[0062] In a specific application scenario, for example, for a slope that has been identified to have creep deformation, the model can input the real-time monitoring of acoustic emission activity enhancement Increase), micro-earthquakes concentrated towards the slope foot changes) and continued heavy rainfall ( Increase) and other information, output the probability of accelerated deformation or overall instability of the slope within the next 24 hours or 72 hours Provide quantitative basis for early warning and response measures.

[0063] Step 4: Based on the risk assessment results, a multi-level warning threshold system is established. When the risk assessment results reach or exceed the preset threshold, the corresponding level of warning is triggered; The specific steps include: Step 4.1, construct a multi-parameter early warning model; Based on the mutation characteristics of the acoustic emission parameter sequence (for example, the sharp increase of ring count and energy value), the spatiotemporal characteristics of microseismic activity (for example, the frequency of events, the non-steady-state trend of energy release, the migration and concentration of the dominant earthquake source location), and the comprehensive risk index and the probability of instability Construct a multi-parameter threshold warning model.

[0064] The core of this model is to set corresponding warning trigger thresholds for different single parameters or combined parameters. When the real-time monitoring value exceeds the preset threshold, the system will automatically trigger the corresponding level of warning signal.

[0065] The construction of this multi-parameter early warning model can adopt a hierarchical early warning strategy.

[0066] For example, a Level 1 warning (concern level) may be based on a single parameter (e.g., the number of acoustic emission rings exceeds three times the average value in a short period of time) or a comprehensive risk index. A slight increase.

[0067] Level 2 warning (attention level) may require multiple parameters to be abnormal at the same time, or the probability of instability Reach a lower threshold.

[0068] Level 3 warning (danger level) may be based on drastic changes in key parameters (such as rapid concentration of microseismic activity in a specific area and significant increase in energy) and the probability of instability. Reaching a higher threshold.

[0069] These thresholds can be determined based on a retrospective analysis of historical disaster cases, combined with expert experience and numerical simulation results. In actual slope monitoring, when the edge computing unit or central server runs the model and detects that the warning conditions are met, it immediately issues a warning message to notify relevant management personnel.

[0070] Step 4.2: Construct a mapping correlation model between internal structural changes and instability risks; Combining the areas of internal structural change and the degree of change identified by the time differential interferometry monitoring algorithm with the spatial distribution characteristics of the precisely located earthquake sources, a mapping correlation model describing the relationship between internal structural changes and potential instability risks is constructed.

[0071] For example, areas with dense earthquake sources or areas with significant changes in Green's function can be deeply correlated with historical landslide hazard data or related numerical simulation analysis results to more accurately identify potential high-risk areas.

[0072] Specifically, the establishment of this mapping association model can first conduct a statistical analysis of historical instability cases to find the spatial correspondence between the medium change area (such as the wave velocity reduction area) indicated by the time differential interferometry monitoring algorithm and the concentrated area of ​​earthquake sources (especially shear-type microseisms) within a specific time window before the instability.

[0073] For example, if it is observed many times that instability often occurs in a weak interlayer at a certain depth, and the Green's function of the interlayer area changes significantly before the instability, and the microseismic events are also concentrated in this area, then the "specific depth medium change + earthquake source concentration in this area" model can be established. high risk of instability”.

[0074] This mapping relationship can also be automatically learned from large amounts of monitoring data through machine learning (such as association rule mining). In practical applications, when new monitoring data meets the established mapping rules, the system can determine that the corresponding area has an increased risk of instability.

[0075] Step 4.3: Configure the low-power edge computing processing unit. In order to achieve near real-time processing of monitoring data and rapid response to warning information, a low-power edge computing processing unit is configured in the embodiment of the present application.

[0076] This unit is usually deployed in the vicinity of the monitoring site and is mainly responsible for the preliminary processing of the raw monitoring data collected by the sensor network (for example, signal filtering, trend noise reduction), the extraction of key features (such as microseismic feature extraction), the calculation of relevant models (such as risk index models, instability probability models, and some lightweight computing modules in early warning models), and the preliminary detection and timely reporting of abnormal events.

[0077] The introduction of edge computing units can significantly reduce the amount of long-distance data transmission, effectively reduce the computing resource pressure on the central server, and significantly improve the timeliness of early warning responses. The edge computing processing unit can include a data receiving module, a data processing module, a model calculation module, an early warning judgment module, and a wireless / wired communication module.

[0078] Step 4.4, constructing a three-dimensional geological information model of the slope; Based on the acquired geological survey basic data (such as drill hole histogram data, engineering geological profile) and digital elevation model (DEM) data, a three-dimensional geological structure information model that can reflect the specific morphology of the slope in the monitoring area is constructed.

[0079] The model should be able to clearly reflect key geological information such as the stratigraphic lithologic composition of the slope area, major discontinuous structural surfaces (e.g., existing faults, large joints and fissures), etc.

[0080] Step 4.5: Combine multi-scale damage information and realize 3D visualization; The core information, including the spatial location information of the precisely located acoustic emission and microseismic events, the regional information of internal structural changes, and the damage evolution state parameters evaluated by the damage evolution analysis model, are accurately superimposed on the constructed three-dimensional geological information model.

[0081] By employing graphical techniques such as visual symbology (e.g., using spheres of varying sizes and colors to represent seismic sources of varying energies), color rendering schemes (e.g., using a color spectrum to represent damage severity or rate of change), or transparency adjustment, the spatial distribution characteristics of the slope's internal structure, from microscopic cracks (typically characterized by acoustic emission events) to macroscopic potential slip surfaces (typically characterized by dense clusters of microseismic events or areas of significant structural change), and their temporal evolution (specific output) are intuitively displayed. In some implementations, virtual reality (VR) or augmented reality (AR) technologies can be incorporated to provide a more immersive interactive visualization experience.

[0082] Step 4.6, identify potential risk areas and support dynamic update function; Based on the analysis results of the mapping correlation model between internal structural changes and instability risks and the judgment results of the early warning model, potential high-risk areas or possible unstable sliding paths are clearly identified in the 3D visualization system interface through highlighting and other methods.

[0083] The results displayed by the visualization system can be dynamically updated in real time based on the latest monitoring data and analysis results, thereby providing a continuous and reliable basis for key decision-making links such as effective disaster prevention and scientific delineation of emergency evacuation areas.

[0084] Application examples of this embodiment: Application scenarios: This implementation was applied to a steep slope adjacent to a mountain highway. This slope had a history of small rockfalls and surface slides, posing a potential threat to traffic safety. The slope's rock mass is primarily moderately weathered sandstone with well-developed joints and fissures, making it susceptible to rainfall infiltration during the rainy season, leading to reduced stability. To ensure safe highway operations, long-term automated monitoring and early warning of this slope were required.

[0085] Implementation process example: Example of building a multi-scale signal acquisition network based on Beidou timing: Two high-precision Beidou timing receivers were installed near the top and toe of the slope, ensuring time synchronization accuracy better than 10ns. Twelve broadband seismic sensors (frequency range 1Hz-1kHz) were deployed in three layers along the slope surface to monitor microseismic signals and extract ambient noise. Eight acoustic emission sensors (frequency range 50kHz-300kHz) were deployed in denser areas of potential fracture development. All sensors were connected to an edge computing processing unit via wired connections and synchronized using Beidou timing signals.

[0086] Example of extracting and processing multi-source signals: 24 hours of continuous environmental noise data were selected, and the cross-correlation calculation was performed on the signals recorded by any two seismometers, and the empirical Green's function between them was extracted.

[0087] For example, sensors A and B are in the initial state The cross-correlation function is After a week Time, calculate again .

[0088] Calculate the energy change: ; in, Indicates from time At the time The change in the cross-correlation function between sensors A and B during the period.

[0089] It was found that the Green's function waveforms between the sensor pairs in the upper and middle parts of the slope showed a relative change of 0.8%, indicating that the medium properties in this area may have changed.

[0090] The Independent Component Analysis (ICA) algorithm was used to successfully separate the mixed acoustic emission and microseismic signals from the acquired broadband signals. The separated microseismic signals were decomposed into eight layers using wavelet packet transform. Combined with an adaptive threshold based on the kurtosis criterion, the P waveguide was accurately picked up with an error of less than 5ms.

[0091] Example of establishing a multi-scale signal correlation analysis model: Using monitoring data from the past year (including the number and average energy of daily acoustic emission events, the number and average magnitude of daily microseismic events, and daily rainfall), we trained an acoustic emission and microseismic signal correlation analysis model: ; in, represents the comprehensive risk index, is the normalized acoustic emission parameter of the day, is the normalized microseismic parameter with a lag of 24 hours, The comprehensive risk index output by the model is The range is 0-1.

[0092] After a continuous rainfall, the number of monitored acoustic emission events increased from an average of 100 times / day to 500 times / day. 20 hours after the acoustic emission activity intensified, the frequency of microseismic events also increased from 2 times / day to 10 times / day. The improved double-difference positioning algorithm (using a three-dimensional velocity model) showed that the earthquake source was concentrated in the middle of the slope, with a positioning accuracy of 0.45 meters. The index reached 0.75.

[0093] At the same time, the evolution law model of micro damage to macro instability composed of LSTM neural network trained based on historical data Enter the current acoustic emission parameters (characterized by the number of events and energy), microseismic parameters (characterized by frequency and spatial concentration) and environmental factors rainfall), predict the probability of instability in the next 24 hours It is 65%.

[0094] in, In the future The probability value of instability occurring within the time period, that is, the probability of instability occurring at the current moment After The probability that geological disasters such as landslide or collapse may occur on the slope within a certain period of time; Indicates The characteristic parameters of the microscopic damage state at each moment mainly include the frequency and energy of acoustic emission events, which are indicators reflecting the development of microscopic cracks; Indicates The characteristic parameters of the mesoscopic damage state at the time of the earthquake mainly include the frequency of microseismic events and the concentration of earthquake source locations, which are indicators reflecting the deformation of larger-scale structures. Indicates The combined influence of environmental factors at the time, mainly referring to the external triggering factor of rainfall; A nonlinear mapping function constructed on behalf of an LSTM (Long Short-Term Memory) neural network, which is used to convert current monitoring parameters into predicted values ​​of the probability of instability at a specific point in the future.

[0095] Example of establishing a multi-parameter early warning decision system: The early warning system sets the following thresholds: Comprehensive risk index And the probability of instability When the level 2 (orange) warning is triggered; At the same time, if the time differential interference monitoring results show that the medium change rate in a specific area exceeds 0.5% / day, and the area coincides with the microseismic concentration area, the warning level will be raised to level 1 (red) warning.

[0096] In the aforementioned continuous rainfall events, due to The Green function change area coincided with the microseismic concentration area (the middle of the slope), and the medium change rate was 0.8% / day>0.5% / day. The system automatically triggered a level 1 (red) warning and notified the highway management department and emergency response personnel via SMS and App push.

[0097] Technical effect verification: Through the long-term application of the above method and system on the highway slope (for example, continuous monitoring for more than 2 years), the following main technical effects have been verified: High-precision positioning and multi-scale monitoring capabilities: Based on BeiDou's 10ns timing, the source location accuracy of multiple microseismic events has been verified to be better than 0.5 meters after field verification. This improves positioning accuracy by at least an order of magnitude compared to conventional methods that rely solely on network synchronization (typically with positioning accuracy ranging from several to tens of meters). By simultaneously monitoring acoustic emission and microseismic signals, the system captured the expansion of microscopic cracks, which showed a significant increase in acoustic emission activity, 75 hours before a major landslide. It also detected a sharp increase in the frequency and energy of microseismic events 30 hours before the landslide, demonstrating comprehensive monitoring capabilities from the microscopic to the macroscopic level.

[0098] Disaster prediction accuracy and warning timeliness: During the two-year monitoring period, the system successfully warned of three small landslides and multiple rockfalls, achieving an overall prediction accuracy of 88% (defined as the number of successful warnings divided by (the number of successful warnings + the number of missed warnings)). By capturing earlier precursor information (such as increased acoustic emission activity), the average warning time was approximately 48 to 72 hours earlier than traditional methods that rely solely on surface deformation monitoring. The false alarm rate (defined as the number of false alarms divided by (the number of successful warnings + the number of false alarms)) was kept below 7%. The technical performance is shown in Table 1: Table 1: Comparison of technical effects; ; The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A Beidou-based slope geological disaster monitoring and early warning method, characterized in that: The following steps are involved: Acquire multi-scale geological signals based on BeiDou system timing, including environmental noise signals, acoustic emission signals, and microseismic signals; Process multi-scale geological signals to extract environmental noise interference features that characterize microstructural changes within the slope, acoustic emission signal features that characterize the microscopic material damage process, and microseismic signal features that characterize macrostructural deformation. Based on the environmental noise interference characteristics, acoustic emission characteristics and microseismic characteristics, a slope instability risk assessment model is constructed and the risk assessment results are output; Based on the risk assessment results, a multi-level warning threshold system is established. When the risk assessment results reach or exceed the preset threshold, the corresponding level of warning is triggered.

2. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: The step of extracting the environmental noise interference characteristics that characterize the microstructural changes inside the slope includes: The cross-correlation function is used to process the environmental noise signal and extract the Green's function of the underground medium; A time differential interferometry monitoring algorithm is constructed to identify small changes in the internal structure of the slope and obtain the interference characteristics of environmental noise by calculating the differences in the Green's functions of the underground medium obtained at different times.

3. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: Extracting acoustic emission signal features and microseismic signal features representing macroscopic structural deformation in the step of processing multi-scale geological signals includes: The mixed signal of acoustic emission signal and microseismic signal is decoupled by blind source separation algorithm to separate independent acoustic emission signal and microseismic signal.

4. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: The step of processing the multi-scale geological signals to extract microseismic signal features further includes: Wavelet packet transform and adaptive threshold algorithm are used to extract the arrival time, amplitude, duration and main frequency of the separated microseismic signals as microseismic features.

5. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: The steps of constructing the slope instability risk assessment model include: Construct an acoustic emission and microseismic signal correlation analysis model, combine acoustic emission characteristics, microseismic characteristics and environmental factors, and output a comprehensive risk index; Through the improved double-difference positioning method, the arrival time data of acoustic emission signals and microseismic signals are used to achieve earthquake source positioning.

6. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: The step of constructing the slope instability risk assessment model further comprises: A model of the evolution of microscopic damage to macroscopic instability is constructed, and the probability of instability is predicted by combining microscopic damage parameters characterized by acoustic emission characteristics, mesoscopic damage parameters characterized by microseismic characteristics, and environmental factors.

7. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: The steps of providing early warning include: Based on the sudden changes in acoustic emission characteristics, the activity trend of microseismic characteristics, the comprehensive risk index and the probability of instability, a multi-parameter early warning model is constructed, and an early warning is triggered when the monitoring parameters reach the preset threshold.

8. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: Also includes: A low-power edge computing processing unit is configured for real-time processing of multi-scale geological signals and anomaly detection in early warning.

9. The method for monitoring and early warning of slope geological disasters based on Beidou according to claim 1, characterized in that: Also includes: A three-dimensional visualization system for multi-scale slope damage is constructed to visualize the earthquake source location results, slight changes in the slope's internal structure, and instability risk assessment results in three dimensions.

10. A BeiDou-based slope geological disaster monitoring and early warning system, used to implement the BeiDou-based slope geological disaster monitoring and early warning method according to any one of claims 1 to 9, characterized in that: include: The signal acquisition module is used to obtain multi-scale geological signals based on BeiDou system timing. The multi-scale geological signals include environmental noise signals, acoustic emission signals, and microseismic signals. The signal processing module is used to process multi-scale geological signals and extract the environmental noise interference characteristics that characterize the microstructural changes within the slope, the acoustic emission signal characteristics that characterize the microscopic damage process of the material, and the microseismic signal characteristics that characterize the macrostructural deformation; The risk assessment module builds a slope instability risk assessment model based on environmental noise interference characteristics, acoustic emission characteristics, and microseismic characteristics, and outputs risk assessment results; Early warning module, used to issue early warnings based on risk assessment results; The visualization module is used to display risk assessment results and warning information in three-dimensional visualization.

Images