A multi-stage early warning triggering method and monitoring system for slope stability

Abstract

This invention provides a multi-level early warning triggering method and monitoring system for slope stability. The method deploys high-frequency sensors on the slope to simultaneously collect triaxial acceleration and tilt rate of change signals, constructs a deformation energy index (DEI), and extracts its time growth rate and fluctuation entropy. Based on the combined characteristics of DEI amplitude, growth rate, and entropy, it dynamically determines four levels of early warning status: blue, yellow, orange, and red, achieving early and accurate early warning throughout the entire process from stability, creep, acceleration to imminent slippage. The system includes a high-frequency sensing unit, an edge processing unit, a multi-modal early warning output unit, and a cloud management platform, supporting local intelligent discrimination and regional risk heat map generation. This invention requires no geological model or prior parameters, has strong anti-interference capabilities, a low false alarm rate, and can issue early warnings several hours to several days in advance, making it suitable for various engineering and emergency scenarios.

Description

Technical Field

[0001] This invention belongs to the field of geotechnical engineering technology, and more specifically, relates to a multi-level early warning monitoring system and monitoring system for slope stability. Background Technology

[0002] Slope instability is a common geological hazard, often triggering secondary disasters such as landslides and collapses, causing severe economic losses and casualties. Traditional slope stability monitoring and early warning methods mainly rely on apparent parameters such as displacement, crack width, and acceleration, and trigger alarms by setting fixed thresholds. However, these methods have significant drawbacks: First, the early warning is significantly delayed, often only issuing an alarm after macroscopic deformation has occurred, making it difficult to capture early signs of instability. Second, the false alarm rate is high; environmental disturbances such as temperature expansion and contraction, rainfall erosion, or construction vibrations are easily misjudged as risk events, leading to unnecessary emergency responses. Third, there is a lack of physical mechanism support; most methods are based on empirical thresholds or simple multi-parameter weighted fusion, which cannot truly reflect the evolution of the internal mechanical state of the slope, making it difficult to provide interpretable and guiding early warning information for engineering treatment. Fourth, the adaptability is poor; static thresholds are difficult to adaptively adjust with dynamic environmental conditions such as rainy seasons and freeze-thaw cycles, resulting in decreased early warning reliability.

[0003] In recent years, with the development of sensor technology and data processing capabilities, a number of new monitoring methods have been proposed. For example, CN120183133B discloses a real-time monitoring and early warning method and system for mountain engineering slopes based on digital twins, which outputs stability assessment data and performs risk classification by constructing a multi-physics coupled model; CN120997998A targets river slopes, uses Bayesian inversion to dynamically update model parameters, and generates a comprehensive risk score through a multi-index fusion algorithm. Although the above methods have improved the accuracy of early warning to some extent, they still have significant limitations: First, they are highly dependent on complex numerical modeling and simulation calculations, requiring the construction of a fine digital twin and the running of high-intensity finite element analysis, which not only consumes a lot of computational resources but also requires high professional skills from technical personnel; second, some methods rely on long-term historical monitoring data for model calibration or threshold calibration, making them difficult to apply to slope engineering projects lacking survey data or newly put into use; finally, their early warning criteria usually undergo multiple layers of intermediate variable transformation (such as risk scores, disaster indices, etc.), resulting in long logical chains and ambiguous physical meanings, leading to weak system interpretability and limited engineering practicality. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-level early warning triggering method and monitoring system for slope stability. By acquiring high-frequency signals of slope micro-vibration and micro-tilt angle changes, a deformation energy index (DEI) is constructed. Based on the changes in its amplitude, time growth rate, and fluctuation entropy, early, accurate, and interpretable graded early warnings for the entire process of slope instability are achieved, overcoming the shortcomings of traditional methods such as strong lag, high false alarm rate, and reliance on models.

[0005] A multi-level early warning triggering method for slope stability includes the following steps:

[0006] S1: Collect triaxial acceleration signals and tilt rate of change signals using high-frequency sensors deployed on the slope; S2: Calculate the deformation kinetic energy per unit mass based on the triaxial acceleration signals; S3: Generate a deformation energy index (DEI) by combining the deformation kinetic energy per unit mass with the tilt rate of change signals; S4: Calculate the time growth rate and fluctuation entropy of the deformation energy index (DEI); S5: Determine the corresponding warning level based on the amplitude, time growth rate, and fluctuation entropy of the deformation energy index (DEI), and output a warning signal corresponding to the warning level.

[0007] Furthermore, the deformation energy index DEI is calculated according to the following formula:

[0008]

[0009] in, , which is the kinetic energy per unit mass of deformation. α represents the rate of change of tilt angle, and β represents the preset weighting coefficients.

[0010] Furthermore, the fluctuation entropy is obtained by calculating the sample entropy of the deformation energy index (DEI) sequence within the sliding window.

[0011] Furthermore, the warning levels include four levels: Blue warning: triggered when the deformation energy index (DEI) is below the first threshold and the time growth rate is close to zero, and the fluctuation entropy is low; Yellow warning: triggered when the deformation energy index (DEI) is between the first and second thresholds, the time growth rate is greater than zero, and the fluctuation entropy is stable; Orange warning: triggered when the deformation energy index (DEI) is above the second threshold, the time growth rate rises rapidly, and the fluctuation entropy increases; Red warning: triggered when the deformation energy index (DEI) suddenly jumps to its peak and then drops sharply by more than a preset proportion, and the fluctuation entropy increases significantly.

[0012] Furthermore, the first threshold and the second threshold are dynamically set based on historical deformation energy index (DEI) data.

[0013] A multi-level early warning and monitoring system for slope stability includes:

[0014] A high-frequency sensing unit is used to collect triaxial acceleration signals and tilt rate of change signals of the slope, with a sampling frequency of not less than 10 Hz; an edge processing unit is deployed on the slope site and is communicatively connected to the high-frequency sensing unit, and is equipped with a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements the method as described in any one of claims 1 to 6; an early warning output unit is connected to the edge processing unit and is used to issue early warning signals according to the determined early warning level; a cloud management platform is connected to the edge processing unit through a wireless communication network and is used to receive and store historical deformation energy index (DEI) sequences and provide regional risk analysis services.

[0015] Furthermore, the high-frequency sensing unit includes a MEMS triaxial accelerometer and a high-precision inclinometer, and the early warning output unit includes at least one of an audible and visual alarm device, a mobile terminal notification module, or an emergency control system interface.

[0016] Furthermore, the cloud management platform is equipped with a regional risk analysis module, which is used to aggregate the deformation energy index (DEI) evolution data of multiple slopes and generate a regional landslide risk distribution map.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] (1) By using high-frequency sampling of no less than 10 Hz, the high-frequency vibration energy signal generated by micro-fractures and particle slippage inside the slope can be effectively captured. It can identify the abnormal growth of deformation energy index (DEI) in the early stage of creep before the macro displacement has developed significantly. It can issue an early warning several hours to several days earlier than the traditional method based on displacement threshold or low-frequency monitoring, thus gaining valuable time for emergency response.

[0019] (2) By introducing fluctuation entropy as a criterion for system orderliness, it is possible to effectively distinguish between random vibrations caused by environmental disturbances such as temperature changes and vehicle traffic and orderly deterioration caused by the accumulation of internal damage in the slope. At the same time, red alerts must simultaneously meet the triple conditions of "DEI jump - sudden drop - entropy increase", which greatly suppresses false alarms caused by fluctuations in a single parameter.

[0020] (3) This method is based entirely on field measured dynamic signals to construct the deformation energy index. It does not involve geotechnical parameters, finite element modeling, safety factor calculation or historical disaster data. It is particularly suitable for new engineering slopes, emergency rescue scenarios or remote areas where detailed survey data is lacking. It has a low deployment threshold and strong promotion potential. Attached Figure Description

[0021] Figure 1 This is a partial cross-sectional three-dimensional structural diagram of the connection structure of an external connection cable for an electronic device according to the present invention.

[0022] Figure 2 This is a partially exploded cross-sectional view of the connection structure of an external connection cable for an electronic device according to the present invention. Detailed Implementation

[0023] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0024] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0025] Example 1:

[0026] like Figure 1 As shown, a multi-level early warning triggering method for slope stability includes the following steps:

[0027] S1. High-frequency signal acquisition: High-frequency sensors are deployed at key locations on the slope (such as the top of the slope, near potential slip surfaces, and at the toe of the slope) to simultaneously acquire triaxial acceleration signals a. x (t), a y (t), a z (t) and tilt angle change rate signal The high-frequency sensor has a sampling frequency of not less than 10 Hz, preferably 20–100 Hz, to effectively capture the high-frequency vibration energy generated by micro-fractures and particle slippage in the soil and rock mass.

[0028] S2. Calculation of Kinetic Energy per Unit Mass Deformation: Based on Newton's second law, half of the sum of the squares of the three-axis accelerations is defined as the kinetic energy per unit mass deformation.

[0029]

[0030] This kinetic energy reflects the instantaneous motion intensity of the micro-units inside the slope due to stress adjustment.

[0031] S3. Deformation Energy Index (DEI): Taking into account both translational energy and rotational tendency, the deformation energy index is defined as follows:

[0032]

[0033] Here, α and β are preset weighting coefficients used to balance the dimensions and sensitivity of acceleration and tilt angle signals. In the absence of prior information, α can be set to 1 and β to 0.5; alternatively, they can be dynamically adjusted based on on-site calibration.

[0034] S4. Feature Extraction:

[0035] The time growth rate r(t) is obtained by estimating the DEI sequence through sliding difference or first derivative and is used to characterize the energy accumulation rate.

[0036] Fluctuation entropy S(t): The complexity of calculating the DEI sequence within a sliding window of length N using the sample entropy algorithm. Lower sample entropy values ​​indicate a more ordered system, such as stable creep; higher values ​​indicate a more chaotic system, such as near-slip mutation. Specifically, the preferred parameters for sample entropy calculation are an embedding dimension m=2, a similarity tolerance r=0.2×std(DEI), and a window length of 30~60 seconds.

[0037] S5. Multi-level early warning judgment and output: Based on the combination of the amplitude of DEI, the growth rate r(t), and the fluctuation entropy S(t), a four-level early warning is triggered:

[0038] Blue alert (Level IV slope stability): DEI < D0, |r(t)| < ε1, entropy S(t) < S0; at this time, the slope is in a state of slight vibration caused by thermal noise or environmental disturbances (such as wind vibration, vehicle traffic), with no obvious trend of energy accumulation, and the system behavior is highly ordered. The system only records data and does not trigger external alarms; the "normal" status can be displayed on the management platform.

[0039] Yellow alert for Level III creep initiation: D0≤DEI < D1r(t)>ε1, fluctuation entropy S(t)∈[S0,S1], and the change is stable (|dS / dt| is small); at this time, micro-cracks begin to expand or particles rearrange inside the slope, and energy accumulates slowly, but the overall creep is still in a controllable stage and has not yet entered accelerated failure. The system edge endpoints light up with yellow LEDs; a text message / APP notification is sent to designated management personnel: "The slope has entered a slight deterioration stage, please strengthen patrols"; the cloud platform marks the slope as a "focused object" and increases the data upload frequency.

[0040] Orange Alert (Level II Accelerated Deformation): DEI ≥ D1, r rises rapidly with its derivative dr / dt > ε2, fluctuation entropy S(t) > S1 and shows an upward trend (dS / dt > 0); at this point, micro-fractures connect to form a macroscopic slip surface, the slope enters a nonlinear accelerated deformation stage, and the risk of instability increases significantly. On-site audible and visual alarms emit intermittent beeps and flashing orange lights; warning information is automatically sent to the emergency command center and maintenance unit; if the system is connected to traffic control, speed limit or lane closure commands can be triggered; the cloud-based regional risk assessment module is activated to analyze whether surrounding slopes are affected by cascading effects.

[0041] Red Alert (Level I Precipitated Slippage): Triggering condition: Meeting any one of the following conditions is sufficient:

[0042] Scenario A (Sudden Energy Release): The DEI jumps to a local peak in a short period of time and then drops sharply by more than 30%, that is:

[0043]

[0044] Meanwhile, the fluctuation entropy increased by more than 50% within 10 minutes;

[0045] Case B (Sustained High Energy): DEI is consistently higher than D1 and r(t) > 0.5 DEI / s for more than 30 minutes, and the entropy value > 0.8.

[0046] At this point, the internal shear zone of the slope has been penetrated, leading to localized or overall instability. Energy is being released rapidly, and a landslide is highly likely to occur within minutes to hours. On-site, a continuous high-volume alarm is activated, accompanied by flashing red lights; pre-set emergency phone numbers are automatically dialed, and a high-priority alarm is pushed to the government emergency platform; external systems are coordinated: road gates are closed, power / gas pipelines are cut off, and broadcast evacuation instructions are initiated; edge nodes immediately cache raw data from 5 minutes before and after the event, prioritizing its upload once communication is restored.

[0047] To adapt to environmental changes such as different seasons, rainfall, and freeze-thaw cycles, D0 and D1 are updated adaptively using a sliding statistical window. Specifically, at 0:00 every day, the system automatically calculates the mean μ and standard deviation σ of the DEI sequence over the past 7 days (168 hours) and updates them as follows: D0 = μ + 2σ, D1 = μ + 3σ. In case of extreme weather (such as hourly rainfall > 20 mm), a "rain mode" can be temporarily activated to increase D0 and D1 by 10% to 20% to avoid misjudgment of high-frequency vibrations caused by rainfall.

[0048] In this plan, during the initial operation phase (the first 7 days), if there is insufficient historical data, preset empirical values ​​will be used (e.g., D0 = 0.1, D1 = 0.2, with units depending on sensor calibration).

[0049] In this solution, the early warning signal is reliably transmitted through multi-channel redundant output, including...

[0050] Local output: Four-color LED light + graded buzzer (blue: silent; yellow: slow flashing; orange: fast flashing; red: constant light + high-pitched sound);

[0051] Remote output: Mobile terminals: APP push, SMS, voice calls; Emergency system: dry contact switch signals (NO / NC), Modbus TCP commands; Public platform: JSON format reporting conforming to the "Geological Disaster Monitoring Data Interface Specification".

[0052] Example 2

[0053] like Figure 2As shown, a multi-level early warning and monitoring system for slope stability includes:

[0054] High-frequency sensing unit: Composed of a MEMS triaxial accelerometer and a high-precision inclinometer, it supports IP67 protection and has built-in temperature compensation to ensure long-term stability in the field.

[0055] Edge processing unit: Deployed on the slope site, it adopts a low-power ARM Cortex-M series microcontroller, runs a lightweight DEI calculation and early warning judgment program, and has local storage (≥72 hours of data cache) and breakpoint resume capability;

[0056] Early warning output unit: Supports multi-modal output, including on-site sound and light alarm (four-color LED + buzzer), 4G / LoRa wireless push to mobile terminal, and dry contact signal output to traffic control or emergency broadcast system;

[0057] The cloud management platform receives DEI feature data uploaded by each edge node through a secure communication protocol (such as MQTT over TLS), provides functions such as historical backtracking, threshold template distribution, and multi-slope clustering analysis, and integrates a regional risk analysis module to generate a heat map of regional landslide risk distribution based on a spatiotemporal clustering algorithm.

[0058] Example 3: Deployment and Operation of a Multi-Level Early Warning System for Steep Accumulation Slopes along Highways in Southwest Mountainous Areas

[0059] I. Project Background and System Deployment

[0060] The K85+200 section of a certain expressway is a steep artificial fill slope, 22 m high, with a gradient of 1:1.3, mainly composed of gravelly soil and silty clay. Historically, shallow landslides have occurred there. To improve disaster prevention capabilities, the monitoring system described in this invention has been deployed.

[0061] High-frequency sensing unit: Two monitoring points are set up at the top of the slope, two at the middle of the slope (potential slip surface depth of about 8 m), and two at the foot of the slope, for a total of 6 monitoring points. Each point is equipped with: a MEMS triaxial accelerometer (model: ICM-42688-P, range ±4g, noise density 70 μg / Hz); and a high-precision digital inclinometer (resolution 0.001°, temperature drift <0.003° / ℃). All sensors are encapsulated in an IP68 protective box and connected to the edge node via an RS485 bus. The sampling frequency is uniformly set to 20 Hz.

[0062] Edge processing unit: Employs an industrial-grade ARM Cortex-M7 microcontroller (480 MHz clock speed, 1 MB RAM), deployed within a sloped, rainproof cabinet. Built-in programs implement all the aforementioned steps, supporting: real-time DEI calculation (α=1, β=0.5); sliding window sample entropy calculation (window length 60 s, m=2, r=0.2×std); dynamic threshold updates (based on the previous 7 days' data at 00:00 daily); and local storage of ≥72 hours of raw data (MicroSD card).

[0063] Warning output unit: On-site four-color LED light poles + buzzer (blue / yellow / orange / red frequency division control); remote 4G DTU module, pushing to the "Slope Guardian" APP and provincial emergency platform; linkage: dry contact output to road section information board and speed limit sign controller.

[0064] Cloud management platform: Deploy web applications on Alibaba Cloud ECS servers to receive DEI feature data uploaded by each edge node, provide historical backtracking and threshold template distribution, and integrate a regional risk analysis module to perform cluster analysis on 12 slopes in the same road section.

[0065] II. Method Execution Flow and Key Technical Details

[0066] 1. Data Acquisition and Preprocessing: Data is collected synchronously every 50 ms. x ,a y ,a z (Unit: m / s²) and θ˙ (Unit: ° / s); the acceleration signal is low-pass filtered at 5 Hz to retain the effective quality signal.

[0067] 2. Calculation of the deformation energy index (DEI), including...

[0068] Kinetic energy per unit mass of deformation:

[0069]

[0070] DEI synthesis (after dimensionless normalization):

[0071]

[0072] Note: β=0.5 is determined by on-site calibration because the amplitude of the rate of change of tilt angle is usually on the order of 1 / 2 of the acceleration.

[0073] 3. Feature Extraction

[0074] Time growth rate r(t): Calculated using the 5-point central difference method, with noise smoothing;

[0075] Fluctuation entropy S(t): The sample entropy is calculated for a 60-second DEI sequence, reflecting the orderliness of the system.

[0076] 4. Multi-level early warning judgment:

[0077] Dynamic threshold initialization (days 1–7): D0 = 0.12, D1 = 0.18 (based on empirical values ​​for similar slopes);

[0078] Adaptive update (starting from day 8): Calculate the mean μ and standard deviation σ of the DEI over the past 7 days at 0:00 daily, and update: D0 = μ + 2σ, D1 = μ + 3σ

[0079] The early warning rules are implemented as shown in the table below:

[0080]

[0081] III. Retrospective of Typical Warning Events (June 12-15, 2025)

[0082] June 12: Continuous rainfall (cumulative 60 mm), DEI slowly rises to 0.14, r=0.055, S=0.31 → Yellow alert triggered. Maintenance personnel confirmed on-site that the slope was seeping water but no cracks were found.

[0083] June 14: DEI suddenly increased to 0.25, dr / dt=0.015, S=0.58 → Orange alert, traffic restrictions initiated.

[0084] June 15, 07:20: DEI reached a peak of 0.32, then plummeted to 0.19 at 07:28 (↓41%), S jumped to 0.82 → Red alert automatically triggered.

[0085] 07:35: Road closure completed;

[0086] 08:10: A shallow landslide of approximately 200 m³ occurred, with no casualties.

[0087] Compared with traditional methods: the maximum displacement on that day was only 4.1 mm (inclinometer data), which is far below the conventional 10 mm threshold, and the traditional method failed to issue a warning.

[0088] IV. System Robustness and Expansion Functionality

[0089] Network outage test: Simulating a 72-hour communication interruption, the edge node successfully recorded and triggered one yellow alert locally, and automatically retransmitted the data after the network was restored;

[0090] Rain mode: When the hourly rainfall is >20 mm (accessed via meteorological API), D0 / D1 will be temporarily increased by 15% to avoid false alarms caused by rainfall erosion;

[0091] Regional risk map: The cloud platform detected that the DEI of the slopes at K85+200 and K85+500 rose simultaneously, indicating the risk of groundwater uplift in the area and guiding a full-line investigation.

[0092] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A multi-level early warning triggering method for slope stability, characterized in that, Includes the following steps: S1: Collect triaxial acceleration signals and tilt rate of change signals using high-frequency sensors deployed on the slope; S2: Calculate the deformation kinetic energy per unit mass based on the triaxial acceleration signals; S3: Combine the deformation kinetic energy per unit mass with the tilt rate of change signals to generate the deformation energy index DEI; S4: Calculate the time growth rate and fluctuation entropy of the deformation energy index DEI; S5: Determine the corresponding early warning level based on the amplitude, time growth rate and fluctuation entropy of the deformation energy index DEI, and output the early warning signal corresponding to the early warning level.

2. The multi-level early warning triggering method for slope stability as described in claim 1, characterized in that: The deformation energy index DEI is calculated according to the following formula: in, , which is the kinetic energy per unit mass of deformation. α represents the rate of change of tilt angle, and β represents the preset weighting coefficients.

3. The multi-level early warning triggering method for slope stability as described in claim 1, characterized in that: The fluctuation entropy is obtained by calculating the sample entropy of the deformation energy index (DEI) sequence within the sliding window.

4. The multi-level early warning triggering method for slope stability as described in claim 1, characterized in that: The warning levels include four levels, namely: Blue alert: Triggered when the deformation energy index (DEI) is below the first threshold and the time growth rate is close to zero, and the fluctuation entropy is low. Yellow alert: Triggered when the deformation energy index (DEI) is between the first and second thresholds, the time growth rate is greater than zero, and the fluctuation entropy is stable. Orange alert: Triggered when the deformation energy index (DEI) exceeds the second threshold, the time growth rate increases rapidly, and the fluctuation entropy increases; Red alert: Triggered when the deformation energy index (DEI) suddenly jumps to its peak and then drops sharply by more than a preset percentage, and the fluctuation entropy increases significantly.

5. The multi-level early warning triggering method for slope stability as described in claim 5, characterized in that: The first and second thresholds are dynamically set based on historical deformation energy index (DEI) data.

6. A multi-level early warning and monitoring system for slope stability, characterized in that: include: A high-frequency sensing unit is used to acquire triaxial acceleration signals and tilt rate of change signals of the slope, with a sampling frequency of not less than 10 Hz; An edge processing unit is deployed at the slope site and is communicatively connected to the high-frequency sensing unit. It is equipped with a memory and a processor. The memory stores a computer program, and when the program is executed by the processor, it implements the method as described in any one of claims 1 to 6. An early warning output unit, connected to the edge processing unit, is used to issue early warning signals according to the determined early warning level; a cloud management platform, connected to the edge processing unit via a wireless communication network, is used to receive and store historical deformation energy index (DEI) sequences and provide regional risk analysis services.

7. The multi-level early warning and monitoring system for slope stability according to claim 6, characterized in that: The high-frequency sensing unit includes a MEMS triaxial accelerometer and a high-precision inclinometer, and the early warning output unit includes at least one of an audible and visual alarm device, a mobile terminal notification module, or an emergency control system interface.

8. The multi-level early warning monitoring system for slope stability according to claim 6, characterized in that: The cloud management platform is equipped with a regional risk analysis module, which is used to aggregate the deformation energy index (DEI) evolution data of multiple slopes and generate a regional landslide risk distribution map.

Images