Early warning method and system for freeze-thaw disaster of rock slope and medium

Abstract

The invention provides an early warning method and system for freeze-thaw disasters of a rock-soil slope and a medium, belongs to the technical field of intelligent monitoring and early warning of geological disasters, and provides the early warning method and system for the freeze-thaw disasters of the rock-soil slope aiming at the problem of early warning lag caused by the fact that micro-damage accumulation in a rock-soil body is difficult to recognize in the early stage of freeze-thaw circulation. Distributed optical fibers and wireless micro sensing nodes are arranged in the slope and on the surface of the slope to obtain temperature profile, acoustic emission, pore water pressure, micro displacement and other multi-modal data, and the multi-modal data are transmitted back to an edge computing gateway through a low-power-consumption internet of things; the gateway end operates a lightweight physical information neural network with freeze-thaw heat-force-water coupling physical constraints, outputs freeze-thaw micro-fracture active indexes and safety factor evolution, and realizes graded early warning and information reporting in combination with an adaptive threshold and 0 DEG C isothermal surface migration risks, so that the timeliness of micro-damage identification can be improved, and false alarms can be reduced.

Description

Technical Field

[0001] This invention belongs to the field of intelligent monitoring and early warning technology for geological disasters, and in particular relates to an early warning method, system and medium for freeze-thaw disasters on rock and soil slopes. Background Technology

[0002] Soil and rock slopes are widely distributed in high-altitude, cold, and seasonally frozen soil regions. Long-term exposure to repeated freeze-thaw cycles makes them prone to surface spalling, crack expansion, and even overall instability, threatening transportation, water conservancy, and other engineering projects, as well as personnel safety. The freeze-thaw disaster process is slow, insidious, and cumulative. Its mechanism typically involves repeated frost heave stress caused by pore water phase change, prompting the gradual initiation, expansion, and connection of micro-cracks within the soil and rock mass. Before significant macroscopic displacement occurs, this type of micro-damage is often difficult to detect in a timely manner using conventional monitoring methods.

[0003] Existing slope disaster monitoring and early warning systems mostly rely on GNSS, inclinometers, crack gauges, or rainfall-based threshold models. They primarily focus on the macroscopic deformation response of disasters such as landslides and collapses, but lack sufficient perception of the early internal micro-damage accumulation-intensity deterioration-critical instability evolution chain of freeze-thaw disasters, which can easily lead to delayed early warnings.

[0004] While some studies have incorporated environmental parameters such as temperature or shallow moisture into freeze-thaw risk assessment, relying solely on a single parameter and setting fixed thresholds makes it difficult to consider variations in soil and rock types, water content, and freeze-thaw history, resulting in insufficient adaptability and a high false alarm rate. Distributed temperature measurement technology focuses more on temperature field characterization and cannot directly reflect mechanical damage. Acoustic emission technology can detect micro-fractures, but point-based sensors in field scenarios have limited coverage and are susceptible to noise interference. Furthermore, purely data-driven intelligent models, lacking constraints from the freeze-thaw thermo-mechanical-hydraulic coupling mechanism, have limited generalization ability under insufficient sample size or extreme climatic conditions. Summary of the Invention

[0005] In view of this, the present invention aims to provide an early warning method, system and medium for freeze-thaw disasters on soil and rock slopes, so as to at least solve one of the problems in the background art.

[0006] To achieve the above objectives, the technical solution of the present invention is implemented as follows: An early warning method for freeze-thaw disasters on soil and rock slopes includes: A multimodal sensing network is deployed inside and on the surface of the target soil and rock slope. The multimodal sensing network includes distributed optical fiber sensors deployed along the depth direction of the slope and wireless micro-sensor nodes set in the monitoring sensitive area. The monitoring sensitive area is determined based on the slope freeze-thaw cycle characteristics, water enrichment or seepage channel distribution, potential sliding surface or structural weak zone location, and deformation concentration area. The distributed optical fiber sensors are used to simultaneously acquire temperature profile data and distributed acoustic emission signals, and the wireless micro-sensor nodes are used to acquire local temperature, humidity, pore water pressure, and micro-displacement data. The multimodal sensing data collected by the multimodal sensing network is uploaded to the edge computing gateway through the low-power wide-area IoT communication module; Lightweight physical information neural network inference is executed in the edge computing gateway. The lightweight physical information neural network uses the thermo-hydraulic multi-field coupling control equation of the freeze-thaw process of soil and rock as physical constraints, and takes temperature gradient, acoustic emission energy index, acoustic emission event frequency, pore water pressure change rate and micro-displacement change rate as inputs, and outputs the freeze-thaw micro-fracture activity index and the evolution trend of dynamic safety factor of soil and rock slope. The warning level is determined and warning information is generated based on the temporal evolution trend of the freeze-thaw microfracture activity index and the comparison results between the freeze-thaw microfracture activity index and the adaptive warning threshold. The warning information is sent to the remote management platform via a communication link, triggering the on-site warning execution agency to issue a notification.

[0007] Furthermore, the freeze-thaw micro-fracture activity index is obtained by weighted fusion of the cumulative acoustic emission energy per unit time, the frequency of acoustic emission events per unit time, and the temperature gradient perpendicular to the slope direction. The weighting coefficients of the weighted fusion are dynamically adjusted according to the soil type, water content, and current freeze-thaw cycle status.

[0008] Furthermore, the dynamic adjustment of the weighting coefficients includes: establishing a baseline feature distribution using historical data from the disaster-free phase of the site, and correcting the weighting coefficients online based on the baseline feature distribution to suppress false alarms caused by changes in the environmental background.

[0009] Furthermore, the adaptive warning threshold is dynamically updated based on the current number of freeze-thaw cycles, the daily temperature difference change index within a preset time window, and the soil saturation, thereby obtaining a set of thresholds corresponding to different warning levels.

[0010] Furthermore, the adaptive early warning threshold is updated online according to a preset update cycle. The online update dynamically adjusts the triggering conditions for each early warning level based on the current freeze-thaw cycle status, daily temperature difference change index, and soil saturation. The online update of the adaptive warning threshold includes maintaining the baseline characteristics of the Freeze-Thaw Micro-Crack Activity Index (FMAI), updating the baseline characteristics when a preset stability condition is met, and recalculating the triggering conditions for each warning level based on the updated baseline characteristics. The warning level determination adopts a management logic that separates trigger conditions and rollback conditions, and the warning level rollback is only allowed when a preset stability condition is met, so as to suppress frequent switching of warning levels between adjacent levels. When the freeze-thaw microfracture activity index FMAI fails to fall back to the baseline characteristic after multiple consecutive freeze-thaw cycles, it is determined that irreversible damage has occurred to the soil and rock mass and the warning level is automatically upgraded. By combining distributed temperature sensors to identify the location of the 0°C isotherm and combining meteorological forecast information to predict the risk of future freezing depth development, the warning level will be upgraded in conjunction with the risk indicators when they meet the preset upgrade conditions.

[0011] Furthermore, the determination of the early warning level includes: triggering a first early warning level when the freeze-thaw micro-fracture activity index continuously rises and exceeds a first threshold over multiple consecutive inference cycles; triggering a second early warning level when the freeze-thaw micro-fracture activity index exceeds a second threshold and the pore water pressure change rate or micro-displacement change rate meets the upgrade conditions; and triggering a third early warning level when the freeze-thaw micro-fracture activity index exceeds a third threshold and the dynamic safety factor of the soil and rock slope is lower than a preset safety threshold.

[0012] Furthermore, the multimodal sensing data is also used to identify the location and migration speed of the zero-degree Celsius isotherm, and combined with meteorological forecast data, to predict the risk of freezing depth development within a preset prediction time window. When the predicted freezing depth reaches a preset proportional threshold of the potential sliding surface burial depth, the warning level is automatically upgraded.

[0013] Furthermore, when the freeze-thaw micro-fracture activity index fails to return to the baseline level after multiple freeze-thaw cycles on the soil and rock slope, it is determined that irreversible damage has occurred to the soil and rock mass, and the warning level is raised to the second or third warning level.

[0014] Furthermore, the temperature profile data of the distributed optical fiber sensor is obtained by demodulating the Raman backscattering signal. The demodulation includes solving for the intensity ratio of the anti-Stokes component to the Stokes component, and inverting the temperature distribution along the fiber length direction accordingly.

[0015] Furthermore, the distributed acoustic emission signal is acquired through distributed optical fiber acoustic emission detection, and frequency band filtering and time-frequency feature extraction are performed on the distributed acoustic emission signal to distinguish between freeze-thaw micro-fracture signals and environmental noise.

[0016] Furthermore, the method includes preprocessing the acoustic emission signal, which includes: segmenting the acoustic emission signal into segments according to a preset time window and performing time alignment, correcting the baseline drift of the acoustic emission signal, and filtering the acoustic emission signal in combination with a preset effective frequency band to distinguish between freeze-thaw micro-fracture signals and environmental noise. The energy characterization quantity is calculated based on the preprocessed acoustic emission signal, and the candidate acoustic emission events are judged by the preset energy threshold and persistence condition to obtain the start and end times and event energy of the candidate events. Based on the statistical results of the candidate acoustic emission events within a preset statistical window, the cumulative acoustic emission energy per unit time and the frequency of acoustic emission events are obtained, and the cumulative acoustic emission energy per unit time and the frequency of acoustic emission events are used as the input of the lightweight physical information neural network model; The candidate acoustic emission events are subjected to a consistency check, which includes: checking the spatial consistency of the candidate events based on the information of adjacent measurement points along the depth direction of the distributed optical fiber, and correlating the candidate events with at least one of temperature gradient, pore water pressure change rate and micro displacement rate to improve the reliability of freeze-thaw micro-fracture event identification. When a candidate event is inconsistent with the spatial consistency or the correlation verification result, the candidate event is marked as low confidence and its contribution to the cumulative acoustic emission energy per unit time or the frequency of the acoustic emission event is reduced, or the candidate event is removed from the statistics. The edge computing gateway generates a data quality identifier for the acquisition and processing of acoustic emission signals. The data quality identifier is used to indicate sampling loss or communication abnormality. When the data quality identifier meets the preset abnormality conditions, the abnormality identifier and early warning information are sent to the remote management platform together. Furthermore, the wireless micro-sensor node integrates a temperature and humidity sensor, a pore water pressure sensor, and a displacement measurement unit, which provides millimeter-level displacement change data.

[0017] Furthermore, the edge computing gateway includes an artificial intelligence acceleration module and supports remote model updates for lightweight physical information neural networks. The remote model updates are used to load model parameter sets that match the geotechnical engineering characteristics of different regions.

[0018] Furthermore, the wireless micro-sensor node is powered by a combination of a solar power unit and an energy storage unit, and is equipped with a sleep and wake-up mechanism. When the temperature change rate or pore water pressure suddenly increases beyond the corresponding preset wake-up threshold, it switches from sleep state to wake-up state and reports the data.

[0019] Furthermore, this solution discloses an early warning system for freeze-thaw disasters on soil and rock slopes, including: A multimodal sensor network includes a distributed optical fiber sensor and multiple wireless micro-sensor nodes. The distributed optical fiber sensor is used to simultaneously acquire temperature profile data and distributed acoustic emission signals, and the wireless micro-sensor nodes are used to acquire local temperature, humidity, pore water pressure and micro-displacement data. The Internet of Things (IoT) communication module is used to transmit the multimodal sensing data collected by the multimodal sensing network back over a long distance with low power consumption. An edge computing gateway is connected to the IoT communication module. The edge computing gateway has a built-in lightweight physical information neural network and is used to infer the multimodal sensing data using the thermo-hydraulic multi-field coupling control equation of the freeze-thaw process of soil and rock as a physical constraint, and output the freeze-thaw micro-fracture activity index and the evolution trend of the dynamic safety factor of soil and rock slope. The early warning release module is used to determine the early warning level based on the freeze-thaw micro-fracture activity index and the adaptive early warning threshold, generate early warning information and send it to the remote management platform, and drive the on-site early warning execution mechanism to provide prompts. The power supply module is used to provide power to the wireless micro-sensor node and the edge computing gateway.

[0020] Furthermore, the distributed optical fiber sensor is deployed along the borehole structure and supports distributed temperature measurement and distributed acoustic emission detection, so as to achieve synchronous sensing of temperature field and damage field on the same deployment carrier.

[0021] Furthermore, the wireless micro-sensor node includes a temperature and humidity sensor, a pore water pressure sensor, and a displacement measurement unit, and uploads the collected data to the edge computing gateway according to a preset sampling period.

[0022] Furthermore, the IoT communication module uses the LoRaWAN protocol or the NBIoT protocol to transmit the multimodal sensing data back.

[0023] Furthermore, the early warning release module supports pushing tiered early warning information to the monitoring center via cellular communication, BeiDou short message communication, or satellite communication.

[0024] Furthermore, the power supply module includes a solar power supply unit and an energy storage unit. The wireless micro-sensor node is configured with a sleep and wake-up mechanism and automatically wakes up and reports data when the rate of temperature change or the pore water pressure suddenly increases beyond a preset wake-up threshold.

[0025] Compared with existing technologies, the early warning method for freeze-thaw disasters on soil and rock slopes described in this invention has the following advantages: (1) This invention constructs a multimodal sensing network that integrates distributed optical fiber and wireless micro-sensor nodes, and simultaneously acquires damage-related information such as temperature profile and acoustic emission. Combined with response quantities such as pore water pressure and micro-displacement, it enables earlier identification of micro-fracture activity in the early stage of freeze-thaw, which helps to detect risk evolution in advance before macroscopic deformation becomes significant. (2) The present invention deploys a lightweight physical information neural network at the edge computing gateway, uses the freeze-thaw thermo-mechanical-water multi-field coupling control equation as physical constraint, integrates the multimodal input and output freeze-thaw micro-fracture activity index and safety factor evolution trend, takes into account both mechanism consistency and data adaptability, can improve the inference stability under complex working conditions and reduce the risk of false alarms; (3) The present invention adopts an adaptive threshold early warning strategy, which dynamically adjusts the triggering conditions of graded early warning based on factors such as freeze-thaw cycle status, daily temperature difference changes and soil saturation, and can be linked to the risk of 0℃ isothermal surface migration and freezing depth development to improve the adaptability of the early warning strategy to environmental and working condition changes. (4) This invention uses low-power wide-area IoT communication for data backhaul, and combines solar power supply and multi-link early warning release method, which is suitable for long-term autonomous monitoring deployment in areas without mains power, weak network or remote high-altitude cold areas, and is more feasible. Attached Figure Description

[0026] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a schematic diagram of the overall architecture as described in an embodiment of the present invention; Figure 2 This is a schematic diagram of the rock and soil slope layout according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the typical curves showing the evolution of the freeze-thaw microfracture activity index (FMAI) and slope safety factor over time, as described in an embodiment of the present invention. Figure 4 This is a schematic diagram of the method described in an embodiment of the present invention.

[0027] Explanation of reference numerals in the attached figures: 1-Slope; 2-Distributed optical fiber; 3-Slope crack development zone; 4-Wireless micro-sensor node; 5-Edge computing gateway; 6-Detection host; 7-Fiber optic patch cord; 8-Ethernet; 9-Indicator light; 10-Solar panel. Detailed Implementation

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0029] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0030] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" 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; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0031] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0032] This solution discloses an early warning method for freeze-thaw disasters on soil and rock slopes, including the following steps: A multimodal sensing network is deployed inside and on the surface of the target soil and rock slope. The multimodal sensing network includes distributed optical fiber sensors and multiple wireless micro-sensor nodes. The distributed optical fiber sensors are used to synchronously collect temperature distribution data and acoustic emission signals along the depth direction of the slope, and the wireless micro-sensor nodes are used to collect temperature, humidity, pore water pressure and micro-displacement data of local areas.

[0033] The data collected by the distributed fiber optic sensors and wireless micro-sensor nodes is uploaded to the edge computing gateway in real time through a low-power wide-area IoT communication module.

[0034] A lightweight physical information neural network model is run in the edge computing gateway. The model uses the thermo-mechanical-hydraulic multi-field coupled control equations of the freeze-thaw process of soil and rock as physical constraints. The inputs are real-time temperature gradient, cumulative acoustic emission energy per unit time, frequency of acoustic emission events, pore water pressure change rate and micro-displacement rate. The outputs are the freeze-thaw micro-fracture activity index FMAI and the dynamic safety factor of soil and rock slope.

[0035] Based on the temporal evolution trend of the FMAI and its comparison with the adaptive early warning threshold, the corresponding level of freeze-thaw disaster early warning is triggered, and the early warning information is sent to the remote management platform through the wireless communication link.

[0036] The Freeze-Thaw Microfracture Activity Index (FMAI) is calculated using the following formula: ; in, The cumulative energy of acoustic emission per unit time. The frequency of acoustic emission events per unit time. The temperature gradient is perpendicular to the slope direction. , , This is a weighting coefficient, which is dynamically adjusted based on soil type, water content, and the current number of freeze-thaw cycles.

[0037] The adaptive warning threshold is based on the current number of freeze-thaw cycles. The largest daily temperature range in the past 24 hours and soil saturation Dynamically updated, satisfying the following relation: ; in, , , These are empirical constants calibrated using historical disaster samples.

[0038] If the FMAI fails to return to the baseline level after three or more consecutive freeze-thaw cycles, it is determined that irreversible damage has occurred to the soil and rock mass, and an orange or red warning is immediately triggered.

[0039] When a narrow-linewidth laser pulse is injected into an optical fiber, the backscattered light contains Stokes and anti-Stokes Raman components. The intensity of the anti-Stokes component is temperature-sensitive, while the intensity of the Stokes component is largely unaffected by temperature. By demodulating the intensity ratio of these two components, the temperature distribution along the fiber length can be retrieved. The basic relationship can be expressed as: in, Let z be the absolute temperature at position z of the optical fiber. and These represent the anti-Stokes and Stokes Raman backscattered light powers at that location, respectively. This is a constant related to the optical fiber material (approximately 485 K). These are the system calibration coefficients.

[0040] The lightweight physical information neural network model introduces the constitutive relation of frost heave stress and the conservation equation of latent heat of phase change as soft physical constraints during the training phase. Its loss function is composed of a weighted average of the data fitting loss term and the physical residual loss term, ensuring that the model output conforms to the basic laws of soil and rock freeze-thaw mechanics.

[0041] The loss function is defined as: ; The second item is frost heave stress. With the effective stress balance residual, This is the coefficient of frost heave.

[0042] The location of the 0°C isotherm is identified based on data from wireless micro-sensor nodes, and the rate of frost depth development is predicted in the next 24–72 hours based on meteorological forecast data. When the predicted frost depth exceeds 80% of the potential sliding surface depth, the warning level is automatically upgraded.

[0043] The multimodal sensor network consists of distributed optical fiber sensors and wireless micro-sensor nodes. The distributed optical fiber is deployed along the borehole structure. The IoT communication module uses LoRaWAN or NB-IoT protocols to achieve low-power long-distance data backhaul. The edge computing gateway includes an AI acceleration chip and a built-in lightweight physical information neural network model. In terms of early warning release, it supports pushing three-level early warning information to the monitoring center via 4G, Beidou short message or satellite communication.

[0044] The wireless micro-sensor node is powered by a combination of solar panels and supercapacitors and is equipped with a sleep-wake mechanism. It automatically wakes up and reports data when it detects a temperature change rate exceeding 0.5°C / min or a sudden increase in pore water pressure exceeding 10 kPa.

[0045] The edge computing gateway supports remote OTA model update functionality and can load customized physical information neural network parameter sets according to the geotechnical engineering characteristics of different regions.

[0046] The sampling frequency of the acoustic emission signal is no less than 50 kHz, and the effective frequency band is 30–200 kHz, which is used to distinguish the freeze-thaw micro-crack signal from environmental noise.

[0047] In addition, the following embodiments are proposed in this solution: This embodiment takes a section of highway slope in a high-altitude and cold region as an example. The slope is located in a mountainous area at an altitude of about 3,500 meters. The winter temperature can drop to -25°C, and the average number of freeze-thaw cycles per year is about 60. Historically, there have been many slope instability events caused by freeze-thaw action.

[0048] like Figure 1 and Figure 2As shown, three monitoring holes with a depth of 3.0 meters were drilled along the vertical slope 1, located at the top, middle, and toe of the slope, respectively. An armored distributed optical fiber 2 was installed in each hole. Simultaneously, six wireless miniature sensor nodes 4 were installed in the slope crack development area 3.

[0049] The distributed optical fiber 2 is tightly bonded to the hole wall with epoxy resin coupling agent to ensure the effectiveness of strain and temperature transfer.

[0050] The distributed optical fiber 2 is connected to the FC optical fiber patch cord 7 of the detection host 6 at the other end.

[0051] The detection host 6 is the core device for realizing the transmission, reception and demodulation of distributed optical fiber 2 sensing signals. It integrates a laser, photodetector, high-speed data acquisition card and embedded processor.

[0052] The distributed optical fiber 2 simultaneously supports Raman scattering thermometry (DTS) and Rayleigh scattering acoustic emission detection (DFOS-AE), with a spatial resolution of 0.5 meters, a temperature measurement accuracy of ±0.5°C, an acoustic emission sampling frequency of 100 kHz, and an effective frequency band of 30–200 kHz, used to capture micro-crack signals induced by frost heave stress.

[0053] Each wireless miniature sensor node 4 integrates a digital temperature and humidity sensor, a miniature pore water pressure gauge, and an FMCW millimeter-wave radar displacement gauge, with measurement accuracies of ±0.2°C, ±1 kPa, and ±0.1 mm, respectively.

[0054] The edge computing gateway 5 is connected to the detection host 6 via Ethernet 8 and is used to receive the sensor data received by the detection host 6.

[0055] The sensor data is uploaded to the edge computing gateway 5 deployed under the slope 1 via the LoRaWAN protocol.

[0056] LoRaWAN is a low-power wide-area network technology that is well-suited for outdoor applications that require long-distance transmission but have strict power consumption requirements.

[0057] The edge computing gateway 5 uses an ESP32-S3 main control chip paired with a Coral Edge TPU Lite AI acceleration module to run a lightweight physical information neural network model (PINN-Lite).

[0058] The lightweight physical information neural network (PINN-Lite) model uses the embedded freeze-thaw thermo-mechanical-hydraulic multi-field coupled control equation as physical constraints, integrates multi-dimensional inputs such as temperature gradient, acoustic emission energy rate, and pore water pressure change rate, and dynamically outputs the freeze-thaw micro-fracture activity index (FMAI) and the evolution trend of slope safety factor.

[0059] The lightweight physical information neural network (PINN-Lite) model is trained using a transfer learning strategy: it is first pre-trained on a laboratory freeze-thaw cycle test dataset (containing 12 types of soil and rock samples and 300 freeze-thaw cycles), and then fine-tuned using the first 30 days of disaster-free field data. The loss function is defined as: ; The second item is frost heave stress. With the effective stress balance residual, This is the coefficient of frost heave.

[0060] When the lightweight physical information neural network (PINN-Lite) model detects an anomaly, it will send early warning information to a remote management platform or mobile device via the BeiDou short message module so that timely measures can be taken to prevent disasters from occurring.

[0061] The system performs edge inference every 10 minutes. When the FMAI is > 0.6 for three consecutive periods and shows an upward trend, a yellow light warning is triggered. If the pore water pressure increases by more than 15 kPa or the predicted freezing depth exceeds 2.2 meters, the warning is upgraded to an orange light warning. If the FMAI remains above 1.0 and the safety factor Fs < 1.15, a red light warning is immediately issued (indicator light 9 includes yellow, orange, and red lights), and the warning information is pushed to the provincial geological disaster monitoring platform via the Beidou short message module.

[0062] The entire system is powered by 20 W solar panels and 10 F supercapacitors, and can operate continuously in an environment of -30°C without the need for external power grid support.

[0063] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for early warning of freeze-thaw disasters on soil and rock slopes, characterized in that, include: A multimodal sensing network is deployed inside and on the surface of the target soil and rock slope. The multimodal sensing network includes distributed optical fiber sensors deployed along the depth direction of the slope and wireless micro-sensor nodes set in the monitoring sensitive area. The monitoring sensitive area is determined based on the slope freeze-thaw cycle characteristics, water enrichment or seepage channel distribution, potential sliding surface or structural weak zone location, and deformation concentration area. The distributed optical fiber sensors are used to simultaneously acquire temperature profile data and distributed acoustic emission signals, and the wireless micro-sensor nodes are used to acquire local temperature, humidity, pore water pressure, and micro-displacement data. The multimodal sensing data collected by the multimodal sensing network is uploaded to the edge computing gateway through the low-power wide-area IoT communication module; Lightweight physical information neural network inference is executed in the edge computing gateway. The lightweight physical information neural network uses the thermo-hydraulic multi-field coupling control equation of the freeze-thaw process of soil and rock as physical constraints, and takes temperature gradient, acoustic emission energy index, acoustic emission event frequency, pore water pressure change rate and micro-displacement change rate as inputs, and outputs the freeze-thaw micro-fracture activity index and the evolution trend of dynamic safety factor of soil and rock slope. The warning level is determined and warning information is generated based on the temporal evolution trend of the freeze-thaw microfracture activity index and the comparison results between the freeze-thaw microfracture activity index and the adaptive warning threshold. The warning information is sent to the remote management platform via a communication link, triggering the on-site warning execution agency to issue a notification.

2. The method according to claim 1, characterized in that: The freeze-thaw microfracture activity index is obtained by weighted fusion of the cumulative acoustic emission energy per unit time, the frequency of acoustic emission events per unit time, and the temperature gradient perpendicular to the slope direction. The weighting coefficients of the weighted fusion are dynamically adjusted according to the soil type, water content, and current freeze-thaw cycle status.

3. The method according to claim 2, characterized in that: The dynamic adjustment of the weighting coefficients includes: establishing a baseline feature distribution using historical data from the disaster-free phase of the event, and correcting the weighting coefficients online based on the baseline feature distribution.

4. The method according to claim 1, characterized in that: The adaptive early warning threshold is dynamically updated based on the current number of freeze-thaw cycles, the daily temperature difference change index within the preset time window, and the soil saturation. The determination of the early warning level includes: triggering the first early warning level when the freeze-thaw micro-fracture activity index continuously rises and exceeds the first threshold in multiple consecutive inference cycles; triggering the second early warning level when the freeze-thaw micro-fracture activity index exceeds the second threshold and the pore water pressure change rate or micro-displacement change rate meets the upgrade conditions; and triggering the third early warning level when the freeze-thaw micro-fracture activity index exceeds the third threshold and the dynamic safety factor of the soil and rock slope is lower than the preset safety threshold. When the freeze-thaw micro-fracture activity index fails to fall back to the baseline level after multiple freeze-thaw cycles on a soil and rock slope, it is determined that irreversible damage has occurred to the soil and rock mass, and the warning level is raised to the second or third warning level.

5. The method according to claim 1, characterized in that: The multimodal sensing data is also used to identify the location and migration speed of the zero-degree Celsius isotherm, and combined with meteorological forecast data, to predict the risk of freezing depth development within a preset prediction time window. When the predicted freezing depth reaches a preset proportional threshold of the potential sliding surface burial depth, the warning level is automatically upgraded.

6. The method according to claim 1, characterized in that: The temperature profile data of the distributed optical fiber sensor is obtained by demodulating the Raman backscatter signal. The demodulation includes solving for the intensity ratio of the anti-Stokes component to the Stokes component, and inverting the temperature distribution along the fiber length accordingly. The distributed acoustic emission signal is acquired through distributed optical fiber acoustic emission detection, and frequency band filtering and time-frequency feature extraction are performed on the distributed acoustic emission signal.

7. The method according to claim 1, characterized in that: The wireless micro-sensor node integrates a temperature and humidity sensor, a pore water pressure sensor, and a displacement measurement unit, which provides millimeter-level displacement change data. The wireless micro-sensor node is powered by a combination of a solar power unit and an energy storage unit, and is equipped with a sleep and wake-up mechanism. When the temperature change rate or pore water pressure suddenly increases beyond the corresponding preset wake-up threshold, it switches from sleep mode to wake-up mode and reports the data.

8. An early warning system for freeze-thaw disasters on soil and rock slopes, characterized in that, include: A multimodal sensor network includes a distributed optical fiber sensor and multiple wireless micro-sensor nodes. The distributed optical fiber sensor is used to simultaneously acquire temperature profile data and distributed acoustic emission signals, and the wireless micro-sensor nodes are used to acquire local temperature, humidity, pore water pressure and micro-displacement data. The Internet of Things (IoT) communication module is used to transmit the multimodal sensing data collected by the multimodal sensing network back over a long distance with low power consumption. An edge computing gateway is connected to the IoT communication module. The edge computing gateway has a built-in lightweight physical information neural network and is used to infer the multimodal sensing data using the thermo-hydraulic multi-field coupling control equation of the freeze-thaw process of soil and rock as a physical constraint, and output the freeze-thaw micro-fracture activity index and the evolution trend of the dynamic safety factor of soil and rock slope. The early warning release module is used to determine the early warning level based on the freeze-thaw micro-fracture activity index and the adaptive early warning threshold, generate early warning information and send it to the remote management platform, and drive the on-site early warning execution mechanism to provide prompts. The power supply module is used to provide power to the wireless micro-sensor node and the edge computing gateway.

9. The system according to claim 8, characterized in that: The distributed optical fiber sensors are deployed along the borehole structure and support distributed temperature measurement and distributed acoustic emission detection. The wireless micro-sensing node includes a temperature and humidity sensor, a pore water pressure sensor, and a displacement measurement unit, and uploads the collected data to the edge computing gateway according to a preset sampling period. The IoT communication module uses the LoRaWAN protocol or NBIoT protocol to transmit the multimodal sensing data back. The early warning release module supports pushing tiered early warning information to the monitoring center via cellular communication, BeiDou short message communication, or satellite communication. The power supply module includes a solar power supply unit and an energy storage unit. The wireless micro-sensor node is equipped with a sleep and wake-up mechanism and automatically wakes up and reports data when the rate of temperature change or the pore water pressure suddenly increases beyond a preset wake-up threshold.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the method described in any one of claims 1-7.

Images