A Big Data-Based Geotechnical Engineering Monitoring Trend Prediction System

Abstract

This invention belongs to the technical field of geotechnical engineering monitoring, and relates to a geotechnical engineering monitoring trend prediction system based on big data. It includes the following modules: a data acquisition module for generating real-time front-end differential response signals; a threshold judgment module for generating local early warning trigger events; an array activation module for converting into a directional activated local high-sensitivity monitoring array; a data integration module for generating chain response data clusters; a vector analysis module for constructing spatial damage vector chains; and a decision output module for outputting structured early warning decision instructions. This invention addresses the problem that traditional monitoring points are often isolated, lacking the ability to intuitively reveal the propagation path and speed of damage through intelligent, interconnected spatial array sensing driven by physical events.

Description

Technical Field

[0001] This invention belongs to the technical field of geotechnical engineering monitoring, and relates to a geotechnical engineering monitoring trend prediction system based on big data. Background Technology

[0002] In geotechnical engineering projects such as tunnels, slopes, and mines, the instability and failure of soil and rock masses often exhibit sudden and insidious characteristics, posing a significant challenge to traditional monitoring methods. The core issue currently lies in the difficulty of effectively identifying local deformation signals representing precursors to disasters from a background of strong environmental noise, and in assessing their spatial evolution trends in real time. The solutions commonly used in the industry primarily rely on two types of methods. One involves deploying high-precision sensor networks, such as total stations, inclinometers, or fiber optic sensing systems, to collect periodic or continuous data, subsequently transmitting massive amounts of data to backend servers for centralized processing and analysis. The other method is based on numerical simulation, combined with geological survey data, to predict the stability of engineering structures. These traditional methods constitute the mainstream of current geotechnical safety monitoring.

[0003] Existing technologies have recognized the importance of data fusion and dynamic analysis, and have attempted to combine monitoring and simulation to enhance early warning capabilities. For example, invention patent application number 202511067633.9 discloses a geotechnical engineering stability early warning system method that combines slope deformation monitoring and numerical simulation. By fusing multi-source real-time monitoring data with numerical simulation, a dynamically updated slope deformation evolution analysis system is constructed, ultimately achieving automated hierarchical early warning based on early warning indicators. This technical solution, through the linkage of monitoring and simulation, aims to improve the response speed, accuracy of judgment, and scientific rigor of risk assessment in early warning, providing intelligent and dynamic technical support for slope safety prevention and control.

[0004] After in-depth research into existing technical solutions, those skilled in the art have found that they still have some inherent drawbacks in addressing the aforementioned core challenges of early warning for geotechnical engineering disasters. Centralized data processing leads to system response delays, failing to meet the stringent timeliness requirements of disaster early warning. Massive data transmission and complex model calculations consume significant computing resources and are susceptible to transmission interference. More importantly, these methods struggle to effectively isolate common-mode environmental noise at the physical level, easily overwhelming precursor signals and causing missed or false alarms. Furthermore, traditional monitoring points are often isolated, lacking the spatial array sensing capabilities for intelligent linkage based on physical events, and thus failing to intuitively reveal the spread path and speed of damage. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a geotechnical engineering monitoring trend prediction system based on big data.

[0006] A geotechnical engineering monitoring trend prediction system based on big data includes the following modules:

[0007] The data acquisition module deploys a dual-core differential probe in the monitoring area and calculates the physical difference between the agile core monitoring value and the reference core monitoring value in real time through an analog subtraction circuit to generate a front-end differential response signal.

[0008] The threshold judgment module compares the front-end differential response signal with the preset physical trigger threshold, and generates a local warning trigger event when the physical trigger threshold is exceeded.

[0009] The array activation module sends activation commands to the probes located at downstream nodes along a preset stress transmission path based on local early warning trigger events, so as to form a local high-sensitivity monitoring array.

[0010] The data integration module collects and integrates the triggering time, geographic coordinates, and differential signal amplitude of the triggered probes in the local high-sensitivity monitoring array to generate a chain response data cluster.

[0011] The vector analysis module constructs a spatial destruction vector chain based on the triggering timing and spatial location relationship of probes in the chain response data cluster.

[0012] The decision output module matches the disaster level rule base with the length and spread speed of the spatial damage vector chain and the signal amplitude in the chain response data cluster, and outputs structured early warning decision instructions.

[0013] In a further embodiment of the present invention, the data acquisition module is specifically configured to perform the following operations:

[0014] The reference core is anchored in the deep, stable rock and soil within the monitoring area to collect background environmental baseline change data, including temperature and vibration.

[0015] The agile core is encapsulated in a prefabricated agile medium, and the agile medium is deployed in the shallow stress concentration area of ​​the monitoring area to deploy the dual-core differential probe.

[0016] The agile medium is composed of brittle cement mortar or conductive gel, and its physical strength is configured to be lower than that of the surrounding natural rock and soil. Under the stress of the precursor to the disaster, it will preferentially produce micro-fractures or micro-deformations compared to the surrounding rock and soil.

[0017] In a further embodiment of the present invention, the data acquisition module is specifically configured to perform the following operations:

[0018] Connect the outputs of the reference core and the agile core to the two inputs of the analog subtraction circuit, respectively.

[0019] An analog subtraction circuit is used to perform continuous subtraction of analog voltage signals at two input terminals without analog-to-digital conversion.

[0020] The output is an analog voltage waveform that has been canceled out by common environmental noise interference, generating a real-time front-end differential response signal.

[0021] In a further embodiment of the present invention, the threshold determination module is specifically configured to perform the following operations:

[0022] The front-end differential response signal is input to the hardware comparator circuit, whose reference voltage is set to a preset physical trigger threshold.

[0023] The hardware comparator continuously compares the real-time voltage of the front-end differential response signal with the physical trigger threshold;

[0024] When the voltage of the front-end differential response signal exceeds the physical trigger threshold, the hardware comparator generates a high-level local warning trigger event.

[0025] In a further embodiment of the present invention, the array activation module is specifically configured to perform the following operations:

[0026] The physical addresses of downstream nodes are pre-stored in the probe control circuit to construct a potential stress transfer path mapping;

[0027] When a local warning trigger event occurs, the control circuit bypasses the central server and sends a digital instruction packet containing a wake-up command to the probe of the downstream node corresponding to the physical address via a dedicated communication line as a bypass activation command.

[0028] In a further embodiment of the present invention, the array activation module is specifically configured to perform the following operations:

[0029] Based on the geotechnical engineering mechanics model, the physical address of the downstream node is set for each probe in advance, forming a potential stress transmission path;

[0030] When a master node probe generates a local warning trigger event, its control circuit bypasses the central server and sends a bypass activation command containing wake-up and trigger threshold reduction instructions to the probes of its downstream nodes through a dedicated communication line.

[0031] Upon receiving this instruction, the probes of the downstream nodes are awakened from sleep mode, and the physical trigger threshold of their internal hardware comparators is temporarily lowered, transforming them into a locally highly sensitive monitoring array with directional activation.

[0032] In a further embodiment of the present invention, the data integration module is specifically configured to perform the following operations:

[0033] Each activated probe in the local high-sensitivity monitoring array independently runs its front-end differential circuit and hardware comparator to determine the deformation trend of its respective region;

[0034] The main control unit listens to and records the timestamp of each probe in the array that generates a local early warning trigger event, and simultaneously retrieves its pre-stored geographic coordinates and the amplitude of the differential response signal at the moment of triggering.

[0035] The timestamps, geographic coordinates, and differential response signal amplitude data of the triggered probes are structurally integrated to generate a chain response data cluster containing spatiotemporal evolution information.

[0036] In a further embodiment of the present invention, the vector analysis module is specifically configured to perform the following operations:

[0037] Extract the geographic coordinates of the triggered probes from the chain response data cluster and map them in three-dimensional space;

[0038] Based on the order in which each probe is triggered, the coordinates of adjacent triggered probes are connected to form a vector path;

[0039] The total length of the vector path and the trigger time difference between the first and last nodes are calculated to obtain the propagation speed and construct a spatial destruction vector chain.

[0040] In a further embodiment of the present invention, the vector analysis module is specifically configured to perform the following operations:

[0041] The total length of the vector path is obtained by calculating and summing the straight-line distances between adjacent coordinate points. The propagation speed is then obtained by dividing the total length by the trigger time difference between the first and last nodes.

[0042] In a further embodiment of the present invention, the decision output module is specifically configured to perform the following operations:

[0043] The total length of the spatial destruction vector chain and the destruction propagation speed are used as input parameters to query the corresponding basic disaster type and risk level in the disaster level rule base;

[0044] Calculate the average amplitude of the signal from each probe in the cascade response data cluster;

[0045] The risk level is dynamically adjusted based on the average value; the higher the value, the higher the urgency level.

[0046] By integrating and revising the risk level, disaster type, and spatial coverage of the vector path, a structured early warning decision instruction is generated that includes the disaster type, scope of impact, and urgency level.

[0047] In summary, the present invention has the following beneficial technical effects:

[0048] 1. This invention achieves the extraction and locking of the uneven deformation trend of soil and rock through front-end physical differential and hardware logic decision. Its analog subtraction circuit cancels the environmental common-mode noise such as temperature and vibration at the physical layer, so that the effective signal of the advanced deformation of the responsive medium is amplified and identified.

[0049] 2. By using an event-based bypass activation mechanism, the static sensor network is transformed into a dynamic, directional, and highly sensitive monitoring array. When a node detects an abnormal trend, it can autonomously and quickly activate downstream nodes along a preset mechanical path and temporarily enhance their sensitivity. The network's self-organizing capability enables monitoring resources to intelligently focus on potential risk diffusion directions.

[0050] 3. A spatial damage vector chain capable of quantitatively describing the disaster spread process was constructed. By integrating the spatiotemporal information of the triggered probes, a vectorized model containing the spread direction, path length, and expansion speed was automatically generated, providing data support for understanding the physical mechanisms of disasters.

[0051] 4. Decision matching is based on a clear rule base, and structured early warning instructions are output. This reduces the complexity and time-consuming numerical model inversion. By using vector chain parameters and signal amplitudes to match rules, the disaster type, level and scope of impact can be quickly determined and disposal suggestions can be given. This shortens the chain from data to decision and improves the efficiency and pertinence of emergency response. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings are used to provide a further understanding of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This application discloses a schematic diagram of the framework in the embodiments;

[0054] Figure 2 This discloses a flowchart of an embodiment of this application. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] The following is in conjunction with the appendix Figure 1 - Figure 2 A preferred description of the present invention is provided below.

[0057] See attached document Figure 1 - Figure 2A geotechnical engineering monitoring trend prediction system based on big data includes the following modules:

[0058] A dual-core differential probe is deployed in the monitoring area. The dual-core differential probe has a reference core anchored to the stable rock and soil, and an agile core placed in an agile medium with a physical strength lower than that of the surrounding rock and soil. The reference core is anchored in the deep stable rock and soil inside the monitoring area to collect background environmental benchmark change data including temperature and vibration. The physical difference between the agile core monitoring value and the reference core monitoring value is calculated in real time through an analog subtraction circuit. After filtering out common environmental noise at the physical level, a real-time front-end differential response signal is generated.

[0059] The agile medium is composed of brittle cement mortar or conductive gel, and its physical strength is configured to be lower than that of the surrounding natural rock and soil. Under the stress of the precursor to the disaster, it will preferentially produce micro-fractures or micro-deformations compared to the surrounding rock and soil.

[0060] In a further embodiment of the present invention, the data acquisition module is specifically configured to perform the following operations:

[0061] Connect the outputs of the reference core and the agile core to the two inputs of the analog subtraction circuit, respectively.

[0062] An analog subtraction circuit is used to perform continuous subtraction of analog voltage signals at two input terminals without analog-to-digital conversion.

[0063] The output is an analog voltage waveform that has been canceled out by common environmental noise interference, generating a real-time front-end differential response signal.

[0064] Specifically, in key geotechnical engineering areas to be monitored, such as the tunnel face or slope surfaces at risk of sliding, where advanced prediction is required, the mechanical structure of a dual-core differential probe is fixed at the location. This probe contains two independent sensing cores.

[0065] One of the sensing cores serves as a reference core. Its end is firmly anchored and driven into the deep internal area and the relatively stable rock and soil layer through drilling or hydraulic devices.

[0066] The sensing elements built into the reference core begin to continuously collect physical quantities at its location, including changes in the temperature of the soil and rock mass and fluctuations caused by environmental vibrations. These data constitute a baseline change data stream that reflects the overall background environment.

[0067] The term "deep interior" refers to the requirement that the rock and soil mass at this location maintains a long-term stable mechanical state or an extremely low rate of change under the influence of geological history and current engineering activities. It can serve as a reliable benchmark for monitoring background noise. Its determination is not simply about pursuing absolute depth, but rather about selecting rock and soil masses located under intact bedrock, thick-layered stable sedimentary layers, or known inactive faults or weak interlayers based on engineering geological survey reports, borehole core data, and geological profiles. At the same time, it should be far away from the current excavation face, loading area, or potential slip surface to ensure that it is not directly affected by the catastrophic process of the monitored target.

[0068] For example, in slope monitoring, it should be located at a sufficient distance below the potential sliding surface; in tunnel monitoring, it should be located outside the plastic zone or strongly disturbed zone in front of the tunnel face.

[0069] Another sensing core, known as the agility core, is pre-encapsulated in a housing made of a specific material before deployment. This housing is the agility medium, which is prepared from brittle cement mortar mixed in a specific ratio or a special gel with conductive properties.

[0070] The agile medium, which encapsulates the agile core, is deployed in the same monitoring area through casting or embedding in locations known to be prone to stress concentration based on geomechanical analysis. Due to the material composition of the agile medium, its physical strength after curing is designed to be lower than that of the surrounding natural rock and soil. Therefore, when stress begins to accumulate and change within the rock and soil due to precursors of a disaster, the agile medium, with its lower strength, will respond preferentially to the surrounding rock and soil. This manifests as micro-cracks that are invisible to the naked eye or overall shape deformation. The sensing elements inside the agile core monitor these changes in physical state within the agile medium in real time.

[0071] The electrical signal output terminals of the reference core and the agile core are connected to the two input ports of the analog subtraction circuit integrated at the front end of the dual-core differential probe via wires. This analog subtraction circuit consists of an operational amplifier and a resistor network, operates in the pure analog signal domain, and does not contain any digital conversion components. The circuit continuously receives two analog voltage signals from the reference core and the agile core, and performs a real-time physical subtraction operation, that is, subtracts the real-time voltage signal of the reference core from the real-time voltage signal of the agile core.

[0072] This subtraction process physically cancels out the environmental noise components shared by the two signals, such as vibrations from distant vehicles or temperature differences between day and night, because such noise affects both cores simultaneously.

[0073] After the subtraction operation, the circuit outputs a single analog voltage signal that has been filtered out of common noise. This signal is the real-time front-end differential response signal, and its amplitude reflects the degree of advanced deformation of the agile medium relative to the stable rock and soil.

[0074] A dual-core differential probe is a hardware device that integrates two independent sensing units and front-end processing circuits. Its structural feature is that the two cores are physically separate but circuitically connected.

[0075] The reference core is the sensing core in the probe used to establish a monitoring benchmark. Its setting is based on the need to be anchored in the rock and soil body that is determined to be deep and stable by engineering mechanics in order to obtain environmental background data that is not disturbed by local deformation.

[0076] Agile medium is a prefabricated material structure with an agile core. Its functional characteristics are that its physical strength is lower than that of the surrounding natural rock and soil and it is sensitive to stress. The material is set as brittle cement mortar or conductive gel based on more than 500 sets of indoor geotechnical tests to select the proportion that can produce measurable changes in electrical or mechanical parameters under stress of less than 5 kPa.

[0077] It should be noted that the physical strength of the agile medium specifically refers to the uniaxial compressive strength of the material. The uniaxial compressive strength of the agile medium should be configured to be 30%-70% lower than that of the surrounding natural rock and soil. The specific ratio is determined through indoor sensitivity tests. For example, a series of agile medium specimens with different strengths are prepared and loaded together with rock and soil specimens sampled in the field under a pressure machine. The stress-strain curves are recorded. Under the same stress increment, the strength ratio corresponding to the agile medium mix that first shows a clear inflection point in strain increment is determined.

[0078] Next, specific examples of the composition, preparation, and performance characterization of agile media are provided:

[0079] Option A, such as brittle cement mortar:

[0080] Composition and weight ratio: 100 parts ordinary silicate cement, 150-180 parts standard sand, 40-50 parts water, and 2-5 parts sensitizer. The sensitizer is hollow glass microspheres with a particle size of 5-20 μm. The specified parts are by weight.

[0081] Preparation method: After dry mixing evenly according to the above ratio, add water and stir until the consistency is reached. Pour into the mold containing the agile core. Standard curing time is 28 days. Curing requirements are as follows: temperature 20±2 ℃, humidity >95%.

[0082] Key performance indicators: The uniaxial compressive strength is controlled within the range of 3-10 MPa 28 days after molding. The strength can be adjusted within this range by adjusting the water-cement ratio and the amount of sensitizer.

[0083] Option B, such as conductive gel:

[0084] Composition: Polyvinyl alcohol hydrogel is used as the matrix, and conductive fillers such as carbon nanotubes, graphene fragments or carbon black are added at a mass percentage of 10%-15%.

[0085] Preparation method: PVA is dissolved in hot water, conductive filler is added and ultrasonically dispersed, injected into a mold containing agile core, and cross-linked and molded by repeated freeze-thaw cycles.

[0086] The key performance indicators are that the elastic modulus of the gel is controlled within the range of 0.1-1.0 MPa, and its resistivity has a good linear or piecewise linear relationship with strain.

[0087] The uniaxial compressive strength range of shallow soil and rock masses in the monitoring area is determined by drilling undisturbed samples in the monitoring area or by referring to the geological survey report. For example, 5-20 MPa.

[0088] Select or formulate an agile medium according to the above range, so that its uniaxial compressive strength [ ]satisfy This means that the strength of the agile medium is no more than 50% of the minimum estimated strength of the surrounding soil and rock mass. This proportionality factor, such as 0.5, comes from a large number of indoor comparative experiments and can ensure that under stress, for example, when the stress of the surrounding soil and rock mass reaches 5% of its strength, the agile medium has entered the plastic deformation or micro-fracture stage.

[0089] The agile core is a sensing core encapsulated inside an agile medium to sense changes in the medium's own state. Similar to the reference core, its placement is based on shallow stress concentration areas determined by geological survey reports in order to capture the pre-adjustment of the stress field.

[0090] In engineering structures, due to factors such as geometry, geological conditions, or external loads, the stress level in shallow stress concentration zones is significantly higher than that in surrounding areas. Shallow stress concentration zones are most likely to yield, fracture, or undergo significant deformation first. The methods for determining shallow stress concentration zones are based on engineering geomechanical models, numerical simulations, or theoretical calculations to predict and identify areas with high stress concentration coefficients and low safety reserves within the monitoring area, such as the toe of a slope, the top of a slope, the crown of a tunnel, the shoulder of a tunnel, and the core soil of the tunnel face. Alternatively, through engineering experience and on-site geological surveys, areas with known geological defects such as fault fracture zones, densely jointed zones, weak interlayer outcrops, and areas with severe seepage can be identified as shallow stress concentration zones.

[0091] The analog subtraction circuit is a pure hardware circuit module deployed locally on the probe. It performs continuous-time arithmetic subtraction on two analog voltage signals and outputs a differential signal. Its circuit parameters, such as the amplification factor, are determined based on the output range of the matching sensing element and are based on standard circuit design manuals.

[0092] The front-end differential response signal is a continuously changing analog voltage output by the analog subtraction circuit. Its data structure is a voltage waveform that changes over time, and its physical meaning represents the intensity information of the advanced deformation of the agile medium after filtering out common noise.

[0093] The signal conditioning circuitry inside the Agile Core, such as a constant current source, converts the resistance change into a proportional analog voltage change output. Therefore, the Agile Core's output voltage signal... It reflects the degree of accumulation of micro-fractures within the agile medium.

[0094] The reference core uses the same sensing element structure, but is encapsulated in a high-strength protective sleeve and anchored to a stable rock layer. The requirements for a stable rock layer are that it is located in a complete and continuous bedrock, or in a thick layer of sedimentary rock with uniform lithology and undeveloped structural planes, while avoiding known active faults, fracture zones, weak interlayers, strong weathering zones, and large karst caves and other unfavorable geological bodies.

[0095] Its main sensitivity is to temperature changes, which affect the resistance of the conductive adhesive itself and overall vibration. If no micro-cracks occur, its output voltage... It mainly includes environmental noise.

[0096] The algorithm formula for the analog subtraction circuit is:

[0097]

[0098] It is a differential voltage signal, that is, the voltage value obtained after processing by the analog subtraction circuit. This value mainly reflects the non-common-mode variation component in the output voltage of the Agile Core caused by micro-fractures.

[0099] It is the output voltage signal of the Agile Core. The signal conditioning circuit inside the Agile Core, such as the constant current source, converts the resistance change into a proportional analog voltage change output, reflecting the degree of accumulation of micro-fractures inside the Agile medium.

[0100] It is the output voltage of the reference core. The reference core adopts the same sensing element structure as the agile core, but it is encapsulated in a high-strength protective sleeve and anchored to a stable rock layer. It mainly senses temperature changes, such as those affecting the resistance of the conductive adhesive itself, and overall vibrations. If no micro-cracks are generated, its output voltage mainly includes environmental noise.

[0101] in The adjustable gain is used to match the common-mode noise amplitude of the two signals. This is achieved by adjusting... It can maximally offset common-mode drift caused by temperature and overall vibration, making Mainly reflects The non-common-mode variation component caused by micro-fractures.

[0102] The specific formula for calculating the differential signal is as follows:

[0103]

[0104] It is the voltage value of the front-end differential response signal, representing the physical difference between the monitored values ​​of the agile core and the reference core.

[0105] It is the output voltage signal of the agile core, reflecting the changes in the physical state inside the agile medium.

[0106] It is the output voltage signal of the reference core, which mainly includes environmental noise, such as temperature and vibration change data.

[0107] For example, consider the monitoring of the tunnel face in a certain tunnel project.

[0108] The key monitoring area was determined at the arch 10m ahead of the tunnel face. The reference core of the dual-core differential probe was drilled to a depth of 3m in the stable rock layer ahead of the tunnel face using an anchor bolting machine and then anchored.

[0109] Another agile core was pre-encapsulated in brittle cement mortar with a mix ratio of cement:sand:water:sensitizer = 1:1.5:0.45:0.02 to form a spherical agile medium with a diameter of 10 cm. It was then embedded in a shallow hole at the top of the same tunnel face, but only 0.5 m deep into the rock mass. This location was determined by numerical simulation to be a stress concentration zone.

[0110] The reference core measured a background vibration noise voltage of 0.1 V, and the agile core also output a voltage of 0.1 V under the same noise environment.

[0111] When stress adjustment occurs in the rock mass in front of the tunnel face, the agile medium first generates micro-fractures, causing the agile core output voltage to jump to 0.15 V, while the reference core voltage remains at 0.1 V. The analog subtraction circuit at the probe front end performs an operation in real time: the 0.15 V agile core voltage is subtracted from the 0.1 V reference core voltage, outputting a real-time front-end differential response signal of 0.05 V.

[0112] This 0.05 V signal is the effective signal of the advanced deformation of the responsive medium after filtering out the 0.1 V common environmental noise. It should be explained that the sensitizer is a hollow glass microsphere with a particle size of 5-20 μm, and its addition weight ratio is 0.5%-3% of the cement weight.

[0113] Its mechanism of action is to introduce a large number of weak points into the cement mortar, reducing the material's compressive strength and fracture toughness while ensuring that it generates more and more discrete microcracks under stress, thereby amplifying the resistance change signal and improving the front-end differential response signal. The signal-to-noise ratio.

[0114] The threshold judgment module inputs the real-time front-end differential response signal into the hardware comparator. By comparing the signal amplitude with the preset physical trigger threshold, the circuit state is flipped when the signal amplitude exceeds the threshold, generating a local early warning trigger event that represents the locking of the deformation trend of a single node.

[0115] In a further embodiment of the present invention, the threshold determination module is specifically configured to perform the following operations:

[0116] The front-end differential response signal is input to the hardware comparator circuit, whose reference voltage is set to a preset physical trigger threshold.

[0117] The hardware comparator continuously compares the real-time voltage of the front-end differential response signal with the physical trigger threshold;

[0118] When the voltage of the front-end differential response signal exceeds the physical trigger threshold, the hardware comparator immediately generates a high-level local warning trigger event.

[0119] Specifically, the generated real-time front-end differential response signal is physically triggered to lock onto the uneven deformation trend. The real-time front-end differential response signal output from the analog subtraction circuit of the dual-core differential probe is input to the signal input terminal of an independent hardware comparator circuit through a shielded wire. This hardware comparator is an integrated analog circuit chip that contains a high-speed voltage comparison unit.

[0120] Before deployment, a key voltage reference value is pre-set based on geotechnical engineering mechanics analysis and historical monitoring data. This reference value represents the critical point at which the soil and rock mass transforms from a stable state to a trend state, and is called the physical trigger threshold.

[0121] The reference voltage input of the hardware comparator circuit is set to the voltage value of the physical trigger threshold, such as 0.03 V, using a precision potentiometer or digital potentiometer.

[0122] Once the settings are complete, the hardware comparator enters a continuous working state, and its internal circuit continuously compares the instantaneous voltage of the real-time front-end differential response signal with the preset physical trigger threshold voltage at an extremely high frequency.

[0123] Most of the time, the amplitude of the front-end differential response signal is less than the physical trigger threshold, indicating that the deformation of the agile medium has not yet reached the warning critical point, and the hardware comparator maintains its default output state, such as outputting a low level.

[0124] The state of the internal circuit of the hardware comparator will instantaneously flip if and only if the internal stress adjustment of the soil and rock mass causes the agile medium where the agile core is located to undergo pre-deformation, and the degree of deformation is large enough that the voltage amplitude of the real-time front-end differential response signal generated by the data acquisition module exceeds the preset physical trigger threshold. This flipping process is implemented by analog electronic devices at the physical level without the intervention of any software program or digital judgment of the central processing unit.

[0125] After the circuit state flips, the output pin level of the hardware comparator jumps from the default low level to the high level. This high-level signal is defined as a local early warning trigger event. As a digital signal, this event indicates that the uneven deformation trend of the soil and rock mass in the area where the current monitoring node is located has been physically locked by the hardware circuit.

[0126] The hardware comparator is a dedicated integrated circuit module with two analog voltage input ports and one digital level output port. One of the two analog voltage input ports is connected to a signal, and the other is connected to a physical trigger threshold. Its function is to continuously compare the magnitudes of two input voltages and output the corresponding logic level. The circuit model and response speed are set to ensure that it can capture millisecond-level front-end differential response signal transitions.

[0127] The physical trigger threshold is a preset fixed DC voltage value, which serves as the reference voltage for the hardware comparator. Its function is to serve as a critical benchmark for judging whether a trend has occurred. The specific value of this threshold is based on the analysis of data from more than 200 historical disaster cases of similar projects. The typical inflection point voltage of the front-end differential response signal from background fluctuations to abnormal growth is extracted and used as a preliminary reference for the threshold setting.

[0128] The threshold should be adjusted according to the risk level of the specific project. A lower threshold should be used for high-risk projects to improve early warning sensitivity, while a higher threshold should be used for low-risk projects to reduce false alarms. Specific adjustments can be made through the sensitivity coefficient. accomplish, The value is selected within the range of 4.0±0.5 based on the actual situation of the project.

[0129] The local early warning trigger event is a digital level transition signal output by the hardware comparator. Its characteristic attribute is a discrete logic signal, such as a transition from 0 V to 3.3 V. Its function is to serve as a hardware-level abnormal trend confirmation mark, notifying the subsequent system to initiate corresponding actions.

[0130] It should be noted that the physical trigger threshold The setting adopts a dynamic baseline method based on background statistics, and the steps are as follows:

[0131] After system installation, the front-end differential response signal is continuously acquired during N hours (e.g., 168 hours, or one week) of continuous and stable monitoring. The data.

[0132] Calculate the learning period of this background The mean μ and standard deviation σ of the signal.

[0133] Physical trigger threshold The formula for setting it is as follows:

[0134]

[0135] in, The physical trigger threshold serves as a critical benchmark for determining whether a trend has occurred, based on the front-end differential response signal. If the threshold is exceeded, it is considered that an abnormal event may have occurred.

[0136] The front-end differential response signal is continuously acquired during N hours (e.g., 168 hours, or one week) of continuous and stable monitoring. The mean of the data reflects the average signal strength under the background noise level.

[0137] The front-end differential response signal is continuously acquired during N hours (e.g., 168 hours, or one week) of continuous and stable monitoring. The standard deviation of the data measures the degree of dispersion of the signal under background noise levels.

[0138] For sensitivity coefficients, the initial sensitivity coefficients are... A value of 4.0 is recommended. This value is based on the statistical properties of the standard normal distribution and can be fine-tuned within the range of 4.0 ± 0.5 according to the stationarity of the background noise. This is particularly relevant in high-risk modes, such as... =3.5 and high stability mode, for example =4.5 Choose from two preset options. If you choose... =4, meaning when When the signal value exceeds four standard deviations from the background noise level, an abnormal event is considered to have occurred. The value can be manually fine-tuned according to the project risk level, taking a smaller value for high-risk projects and a larger value for low-risk projects.

[0139] Under the assumption that the signal follows a normal distribution, The value determines the degree of deviation of the physical trigger threshold relative to the noise distribution. When the value is 3, the false alarm rate is theoretically about 0.3%. When the value is 4, the false alarm rate is less than 0.01%.

[0140] Taking into account both the sensitivity and reliability of early warning systems, the sensitivity coefficient... The initial recommended value is set to 4.0. In actual engineering, it can be fine-tuned within the range of 3.5 (high risk and high sensitivity) to 4.5 (low risk and high stability) depending on the specific risk level.

[0141] Example calculation: Assuming background learning, =0.001 V, = 0.005 V, take =4.0, then =0.001+4.0×0.005=0.021 V.

[0142] Background learning can be performed periodically, such as monthly, or after changes during the construction phase, to update the information. , and To adapt to the long-term, slow changes in environmental background noise.

[0143] For example, in the case of the data acquisition module output, after the analog subtraction circuit outputs a real-time front-end differential response signal of 0.05 V, this signal is sent to the hardware comparator. Assume that the physical trigger threshold is preset to 0.03 V based on the tunnel's geological characteristics. The specific voltage value of 0.03 V is determined based on historical data analysis, engineering risk levels, and the dynamic baseline method. The aim is to accurately capture abnormal increases in soil and rock deformation trends while avoiding false alarms.

[0144] The hardware comparator continuously compares the data: the 0.05 V real-time signal is greater than the 0.03 V physical trigger threshold. Therefore, the circuit state of the hardware comparator instantly flips, and its output pin voltage jumps from 0V to a high level of 3.3 V. This 3.3 V high-level signal is the generated local early warning trigger event, confirming that at this monitoring point on the tunnel face arch, the uneven deformation trend of the soil and rock mass, indicated by the advanced reaction of the agile medium, has reached the critical state for early warning.

[0145] The array activation module captures local early warning trigger events and, based on these events, triggers the probes to send bypass activation commands to the probes of their adjacent downstream nodes along a preset stress transmission path, thereby transforming the passive, isolated dormant probe network into a directionally activated local high-sensitivity monitoring array.

[0146] In a further embodiment of the present invention, the array activation module is specifically configured to perform the following operations:

[0147] The physical addresses of downstream nodes are pre-stored in the probe control circuit to construct a potential stress transfer path mapping;

[0148] When a local warning trigger event occurs, the control circuit bypasses the central server and sends a digital instruction packet containing a wake-up command to the probe of the downstream node corresponding to the physical address via a dedicated communication line as a bypass activation command.

[0149] Specifically, an example of a dedicated communication line:

[0150] The probes are interconnected through low-power, low-data-rate wireless self-organizing network modules, such as radio frequency chips using LoRa modulation technology, to form a mesh network. Each probe has a unique short network address, such as a physical address. Under normal circumstances, the main controller and most circuits of the probe enter a deep sleep mode, and only the wake-up receiving circuit of the wireless module and the hardware comparator output listening circuit maintain extremely low power consumption.

[0151] The local warning trigger event output by the hardware comparator is connected to the external interrupt pin of the main controller and the transmit enable pin of the wireless module. Once triggered, the main controller is woken up by the interrupt and controls the wireless module to broadcast or unicast a fixed, short wake-up beacon frame in maximum power and reliable mode, based on the pre-stored downstream node address. This frame contains: source address, command code such as 0x01 indicating wake-up and entry into high-sensitivity mode.

[0152] Reference voltage of hardware comparator A digital potentiometer, connected to and controlled by the main controller via an I2C bus, is used. When the probes of downstream nodes are in deep sleep mode, their wireless modules' Worm Response (WoR) circuits continuously monitor the channel. Upon receiving a wake-up beacon frame with a target address matching their own, the main controller is awakened. After parsing the command, the controller sets the resistance of the digital potentiometer to a predefined, lower level via the I2C bus, thereby activating the hardware comparator. From standard value Adjust to high sensitivity value ;

[0153]

[0154] This is the high-sensitivity threshold adjusted by the hardware comparator. When the probe of the downstream node enters the high-sensitivity mode, this threshold is used for signal comparison and judgment.

[0155] The standard threshold of the hardware comparator is a reference voltage value used to determine whether a signal is abnormal under normal conditions.

[0156] The adjustment coefficient indicates that the standard threshold is adjusted proportionally to obtain the high-sensitivity threshold.

[0157] Adjustment coefficient The basis for determination is The settings need to balance tracking sensitivity and anti-interference capability. This has been achieved through simulation and field testing. A value between 0.5 and 0.7 is more suitable, for example, taking... =0.5 means that the probe trigger threshold of the downstream node is reduced to half of the original value, and the sensitivity is doubled.

[0158] Example: If the main probe threshold =0.02V, If the threshold is 0.5, then the probe threshold of the activated downstream node is temporarily adjusted to... =0.01V.

[0159] In a further embodiment of the present invention, the array activation module is specifically configured to perform the following operations:

[0160] Based on the geotechnical engineering mechanics model, the physical address of the downstream node is set for each probe in advance, forming a potential stress transmission path;

[0161] When a master node probe generates a local warning trigger event, its control circuit bypasses the central server and sends a bypass activation command containing wake-up and trigger threshold reduction instructions to the probes of its downstream nodes through a dedicated communication line.

[0162] All probes of downstream nodes that receive this instruction are awakened from sleep mode, and the physical trigger threshold of their internal hardware comparators is temporarily lowered, transforming them into locally highly sensitive monitoring arrays with directional activation.

[0163] Specifically, during the initial deployment phase, based on the analysis of the mechanical properties of geotechnical engineering, the physical address of the downstream node is pre-set for each dual-core differential probe in the network.

[0164] The calculation is based on the engineering mechanics model of the potential stress transmission path inside the soil and rock mass. For example, in the bedding landslide area, the probe located at the top of the slope is set as the next probe along the sliding surface, forming a potential physical stress transmission chain. This physical address information is permanently stored in the control circuit memory of each probe.

[0165] When a probe acting as a master node in the threshold judgment module generates a high-level local early warning trigger event, the event signal is not only reported, but also triggers the dedicated function of the probe's internal control circuit, which then initiates the bypass communication protocol.

[0166] Its operating logic is to bypass the central server or data aggregation node of the entire monitoring system, and send a specific digital instruction packet to the physical address of the probe of the downstream node set in its memory through a pre-laid dedicated communication line such as shielded twisted pair or low power wireless radio frequency link. This instruction packet is the bypass activation instruction.

[0167] The instruction contains explicit command codes. Its core functions are: first, to forcibly wake up the probes of downstream nodes that are in low-power sleep mode; and second, to command the probes of downstream nodes to temporarily adjust the physical trigger threshold of their internal hardware comparators to a lower value. The control circuits of all probes of downstream nodes that receive this bypass activation instruction parse and execute the instruction.

[0168] The probe is awakened from deep sleep mode, and all functional circuits resume full-power operation. Then, the probe's internal control circuit, via a digital potentiometer or analog switch, temporarily adjusts the reference voltage input of its hardware comparator from the standard physical trigger threshold (e.g., 0.03 V) under normal monitoring conditions to a lower sensitivity threshold (e.g., 0.015 V). This lower sensitivity threshold of 0.015 V was determined based on experimental verification and engineering requirements, aiming to detect deformation trends earlier in potentially risky areas while balancing the false alarm rate.

[0169] Through the above process, the master node probe that was initially triggered, and all the downstream node probes that were awakened by its bypass activation command and had their trigger thresholds lowered, together form a monitoring sub-network with higher detection sensitivity in the physical space, targeting the potential stress transmission direction indicated by the master node. This sub-network is the local high-sensitivity monitoring array.

[0170] The activation of this array does not depend on the scheduling of the central server; it is a directed response triggered by direct hardware logic based on physical events between nodes.

[0171] The stress transmission path is the spatial direction in which the internal failure of the rock and soil mass may extend, as predicted in advance by the geomechanical model. The downstream node is a probe node located one or more positions above or below the current main node on the preset stress transmission path. The preset downstream node is based on the principle of spatial proximity and geological trend, and is specifically set during deployment through the system configuration software. The three-dimensional geographic coordinates, engineering geological profile, identification of potential weak interlayers, and main joint direction of each probe are input into the configuration system.

[0172] For any probe Pi, search for all other probes within its spatial neighborhood, such as radius Rm, where R is set to 20–50m according to engineering scale.

[0173] Based on the main potential sliding direction determined by the geological profile, such as the dip direction of the slope and the tunnel excavation direction, the probe that is closest to the tunnel and located downstream in the projection direction is selected as the primary downstream node Pj of probe Pi.

[0174] Probes within a certain lateral range perpendicular to the main direction can be selected as secondary downstream nodes to cope with possible path bifurcation. The generated (Pi, Pj) mapping table is compiled and distributed to each probe and stored in its non-volatile memory, forming a stress transmission path mapping. This mapping is essentially a static network topology configuration based on engineering geological judgment, which is the preset knowledge basis for this module to achieve directional intelligent response.

[0175] A physical address is a unique identifier for each probe in the network, such as an IP address or MAC address. Its characteristic attribute is an identification code embedded in the device hardware, and its setting is based on network topology planning. A bypass activation command is a digital signal packet containing specific opcodes and data fields. Its characteristic attribute is structured data, and its function is to control the operating mode and parameters of probes at downstream nodes point-to-point. The command format and content are set to ensure reliable transmission and parsing in complex noise environments.

[0176] Sleep mode is a low-power operating state designed by the probe to save power. In this state, the main sensing and communication circuits of the probe are turned off or in a very low-power standby state, and only the basic circuit for listening to activation commands is retained.

[0177] A local high-sensitivity monitoring array is a temporary monitoring cluster consisting of a master node probe and the probes of its activated downstream nodes. It is a dynamic network with spatial directionality and higher sensitivity. Its high sensitivity comes from the physical trigger threshold of the probes of the downstream nodes being temporarily lowered.

[0178] For example, if the main node probe, such as the one numbered N01, generates a local early warning trigger event, and its downstream node is preset to be probe N02 located 20m ahead along the tunnel axis, the control circuit of N01 will then send a bypass activation command to the probe with the physical address N02 through a dedicated communication line. The command content is to wake up and set the physical trigger threshold to 0.015 V.

[0179] Upon receiving the command, probe N02, which was in sleep mode, activated all its functions and adjusted the physical trigger threshold of its internal hardware comparator from the normal 0.03 V to 0.015 V. At this time, probes N01 and N02 together formed a local high-sensitivity monitoring array. This array was arranged along the tunnel excavation direction, such as the direction of potential stress disturbance transmission, and the downstream node probe N02's sensitivity to deformation detection was doubled, ready to capture the deformation trend that continued to advance.

[0180] The data integration module drives all activated probes in the local high-sensitivity monitoring array to synchronously execute the data acquisition module and threshold judgment module. If the probes of downstream nodes are also triggered in a short period of time, the triggering time, geographical coordinates and differential signal amplitude of all probes in the array are collected and integrated to generate a chain response data cluster containing spatiotemporal evolution information.

[0181] In a further embodiment of the present invention, the data integration module is specifically configured to perform the following operations:

[0182] Each activated probe in the local high-sensitivity monitoring array independently runs its front-end differential circuit and hardware comparator to determine the deformation trend of its respective region;

[0183] The main control unit listens to and records the timestamp of each probe in the array that generates a local early warning trigger event, and simultaneously retrieves its pre-stored geographic coordinates and the amplitude of the differential response signal at the moment of triggering.

[0184] The timestamps, geographic coordinates, and differential response signal amplitude data of the triggered probes are structurally integrated to generate a chain response data cluster containing spatiotemporal evolution information.

[0185] Specifically, the system drives all activated probes within the local high-sensitivity monitoring array to synchronously perform monitoring and judgment, and integrates the data to generate a chain response data cluster describing the abnormal propagation process. The probes of all downstream nodes that have been awakened by the bypass activation command and have their trigger thresholds lowered, as well as the master node probe that initially triggered the array, together constitute the operating local high-sensitivity monitoring array.

[0186] Each probe in the array, including the master node and downstream nodes, independently and synchronously executes its complete internal data acquisition module and threshold judgment module operation process.

[0187] Each probe continuously acquires the real-time front-end differential response signal through its dual-core differential structure and uses its hardware comparator, such as a partially lowered threshold, to determine whether the signal exceeds its respective current physical trigger threshold.

[0188] The judgment process of each probe is parallel and independent. Once its hardware comparator is triggered, it will generate a local early warning trigger event belonging to that probe. The main control unit starts to monitor the network status of the entire local high-sensitivity monitoring array in real time. The main control unit can be a central server or a pre-designated master node. The event listening and recording process runs within the main control unit.

[0189] When a local early warning trigger event is detected by any probe in the array, the process takes action and records the timestamp of the event with microsecond precision. At the same time as recording the timestamp, the main control unit retrieves the absolute geographic coordinates of the trigger probe in three-dimensional space from its pre-stored system deployment geographic information database based on the unique identifier of the trigger probe, i.e., its physical address.

[0190] The main control unit also obtains the real-time voltage amplitude of the front-end differential response signal of the probe at the moment when the local early warning trigger event is generated through the data acquisition link, that is, the differential response signal amplitude. The above three data items, timestamp, geographic coordinates and differential response signal amplitude, are temporarily cached as a set of associated data.

[0191] The main control unit continues to perform the above-mentioned listening, recording and retrieval operations until the preset array monitoring window period ends, or until no new triggering events occur within a certain period of time.

[0192] During this period, the main control unit will collect multiple sets of related data from all triggered probes, arrange and combine them in chronological order, and integrate these discrete data points into a whole dataset by defining unified data structure fields such as probe ID, time, longitude, latitude, elevation, and signal amplitude.

[0193] The dataset not only contains information from multiple points, but also, through the temporal order and spatial distribution of the data at each point, the spatiotemporal evolution process of the abnormal event from the initial main node and possibly spreading to downstream nodes along a specific path, the dataset is defined as a chain response data cluster.

[0194] The main control unit is the core processing module responsible for coordinating, monitoring, and recording events and data within the array. Its function is to synchronously collect data from multiple sources and build correlations.

[0195] A timestamp is a high-precision time point data that records when an event occurs. It has an absolute time value with a unified reference such as UTC time. Its precision setting is based on the need to distinguish the order of events at the millisecond or even microsecond level. It is usually provided by a clock source from a GPS embedded in the system or a high-stability crystal oscillator.

[0196] Geographic coordinates are three-dimensional data describing the probe's location in the real world, including longitude, latitude, and elevation. These values ​​are determined by surveying instruments and pre-stored in a database during probe deployment.

[0197] The amplitude of the differential response signal is the instantaneous voltage value of the differential response signal at the front end of the probe at the moment of triggering. Its physical meaning represents the relative intensity of the agile medium deformation at the trigger point.

[0198] Structured integration refers to the process of arranging and encapsulating multiple sets of data with the same fields according to a predetermined format. Its function is to generate a machine-readable data set that contains spatiotemporal relationships. The chain response data cluster is the final data set output after structured integration. Its data structure is set to support spatial vector analysis and disaster level matching in subsequent steps.

[0199] For example, the local high-sensitivity monitoring array composed of probes N01 and N02 starts working. Probe N01 is continuously triggered, and probe N02 is also triggered after 5 seconds due to increased sensitivity after activation.

[0200] The main control unit listens to and records: it records when probe N01 is triggered at time T0 and retrieves its pre-stored coordinates. The signal amplitude of 0.05 V at the moment of triggering is obtained; the triggering of probe N02 at time T0+5s is recorded, and its pre-stored coordinates are retrieved. And obtain the signal amplitude of 0.018 V at the moment of triggering.

[0201] After the monitoring window ends, the main control unit integrates the two sets of data in chronological order to generate a chain response data cluster. This data cluster clearly shows that the anomaly first occurred at coordinates […]. At point N01, the signal is relatively strong at 0.05 V; approximately 5 seconds later, an anomaly appears downstream at coordinates [missing information]. At point N02, the signal is weak at 0.018 V. This data cluster describes the potential process of anomaly propagation from N01 to N02 in a spatiotemporal dimension.

[0202] The vector analysis module analyzes the chain response data clusters and, by comparing the triggering timing and spatial position relationship of each probe in the array, identifies and delineates the vector path from the master node to the final triggered node, thereby constructing a spatial failure vector chain that characterizes the direction and speed of internal failure propagation in the soil and rock mass.

[0203] In a further embodiment of the present invention, the vector analysis module is specifically configured to perform the following operations:

[0204] Extract the geographic coordinates of all triggered probes from the chain response data cluster and map them in three-dimensional space;

[0205] Based on the order in which each probe is triggered, the coordinates of adjacent triggered probes are connected to form a vector path;

[0206] The propagation speed is obtained by calculating the total length of the vector path and the trigger time difference between the first and last nodes. A spatial destruction vector chain is constructed. The total length of the vector path is obtained by calculating and accumulating the straight-line distance between adjacent coordinate points. The propagation speed is obtained by dividing the total length by the trigger time difference between the first and last nodes.

[0207] Specifically, by analyzing the spatiotemporal information contained in the chain response data cluster, a spatial failure vector chain characterizing the internal failure and propagation process of the soil and rock mass is constructed. From the structured data of the chain response data cluster, all records of probes that generated local early warning trigger events are extracted. For each record, the geographic coordinate field it contains is read. This field usually includes three values: longitude, latitude, and elevation.

[0208] These coordinate values ​​are input into the 3D spatial coordinate system transformation and mapping module. This module converts the latitude, longitude, and elevation data of each probe into points X, Y, and Z in a 3D rectangular coordinate system based on a predetermined coordinate system, such as the WGS-84 coordinate system. Then, in a virtual or graphical 3D engineering scene model, the map is marked on its corresponding real spatial location, completing the spatial mapping.

[0209] Read the timestamp field from the same record item, sort all triggered probe points in strict order from earliest to latest timestamp, and determine the order in which events are triggered.

[0210] Based on this chronological order, the coordinates of the first triggered probe (such as the main node) and the coordinates of the second triggered probe are connected by a directional line segment, with the arrow of the line segment pointing from the first triggered point to the second triggered point.

[0211] If a third or fourth point is triggered, the second point is connected to the third point, and so on. The basic logic of this connection process is that within a continuous monitoring time window, adjacent triggered events are most likely to have a spatial correlation.

[0212] By combining all the line segments connected in sequence, a broken line with a time arrow is formed, pointing from the initial trigger point to the final trigger point. This broken line is the vector path representing the direction of damage spread, and the system calculates two key quantitative parameters of this vector path.

[0213] First, the total path length L is obtained by summing the straight-line distances between all adjacent coordinate points. For example, when calculating points... Time The distance.

[0214] Suppose there are two points in space and The straight-line distance between two points The calculation formula is:

[0215]

[0216] It is the straight-line distance in three-dimensional space between point A and point B, in meters.

[0217] These are the three-dimensional coordinates of point A, in meters (m). They are usually obtained by transforming the pre-stored geographic coordinates of the point into a new coordinate system.

[0218] These are the three-dimensional coordinates of point B, with units and source as described above. same.

[0219] If the path consists of a series of points, let the distances between adjacent points be... , , ..., Then the total length of the path The calculation formula is:

[0220]

[0221] It is the total length of the vector path, in mm, and is the sum of the straight-line distances between all adjacent points.

[0222] It is the first in the vector path The straight-line distance between adjacent points, in meters.

[0223] It is the number of distances between adjacent points, that is, the number of adjacent point pairs in the path.

[0224] Then put all Add the values ​​to calculate the trigger time difference between the first node (e.g., the first trigger point) and the last node (e.g., the last trigger point) in the vector path. .

[0225] Average speed of destruction expansion It can be done through the formula The time difference was calculated. The unit is seconds (s), and the velocity is... The unit is m / s.

[0226] This vector path with the time arrow, its total length L, and its spread speed... The node sequence information involved is encapsulated and constructed into a data object that can intuitively reflect the spatiotemporal characteristics of the physical spread of the disaster. This object is defined as the spatial destruction vector chain.

[0227] The distance between adjacent points is calculated using a formula, the specific algorithm formula is as follows:

[0228]

[0229] In the distance formula, the units for coordinates X, Y, and Z are meters.

[0230] It is the first Trigger probe number and the number The three-dimensional Euclidean distance between the trigger probes is expressed in meters.

[0231] in It is an index based on the order of the trigger timestamps. 0 corresponds to the first trigger probe.

[0232] , , They are the first The three-dimensional coordinates of the trigger probe are in meters (m). The values ​​are derived from the pre-stored geographic coordinates of the probe after coordinate system transformation.

[0233] , , They are the first The three-dimensional coordinates, units, and source of the trigger probe are as follows: Consistent.

[0234] The average spread rate is calculated using the following formula:

[0235]

[0236] In the velocity formula, the total length The unit is meters (m), representing the time difference. The unit is seconds (s), and the calculation result is... The unit is m / s;

[0237] The coordinates are the three-dimensional coordinates of the previous probe, in meters. The values ​​are derived from the pre-stored geographic coordinates of the probe after coordinate system transformation.

[0238] The three-dimensional coordinates of the next probe are from the same source;

[0239] d is the straight-line distance between adjacent points in three-dimensional space;

[0240] L is the total length of the vector path, which is the sum of the distances d between all adjacent points;

[0241] The difference between the trigger timestamps of the first and last nodes, in seconds;

[0242] : The average velocity of the disruption spreading along the vector path, in m / s.

[0243] Distance calculation is a fundamental operation in three-dimensional Euclidean geometry. In velocity calculation, the total path length L is related to the time difference. The settings are all based on measured data from the chain response data cluster, reflecting the real spatiotemporal evolution relationship.

[0244] Time difference: It is the difference between the timestamps of the first and last triggering probes. All probes in the system achieve microsecond-level time synchronization through built-in GNSS modules or network synchronization protocols, ensuring... The accuracy.

[0245] Among them, mapping in three-dimensional space refers to the process of converting and visually locating the latitude, longitude and elevation data of the probe into a three-dimensional rectangular coordinate system model. Its function is to establish an intuitive correspondence between data and physical space.

[0246] The chronological order refers to the sequence of events arranged according to timestamps. A vector path with a time arrow is a directed broken line formed by connecting spatial coordinate points in chronological order. It is set based on the principle of continuity and proximity in the transmission of rock and soil failure in the medium. The total length is the cumulative spatial span of the vector path in three-dimensional space, and the unit is meters.

[0247] The trigger time difference between the first and last nodes is the time interval between the first and last trigger events, measured in seconds.

[0248] The rate of damage propagation is a physical quantity that describes how quickly an anomaly spreads along a vector path. It is measured in m / s and is calculated based on the ratio of spatial distance to time interval.

[0249] The spatial damage vector chain is a composite data object that integrates information such as vector path, length, and velocity. Its characteristic attribute is a quantitative and vectorized description of the damage propagation process inside the rock and soil mass, and it serves as the input for disaster level matching.

[0250] For example, the chain response data cluster includes probe N01, such as coordinates ( ), time T0 and N02, such as coordinates ( The data is at time T0+5 s. Map the two points to the 3D model, according to the time sequence; for example, if T0 is earlier than T0+5 s, map the points (...). ) and point ( Connect the arrows, with the arrow pointing from N01 to N02, to form a vector path.

[0251] Formula for calculating vector path length:

[0252]

[0253] Assume the calculation result is 15 Calculate the time difference =5 This would disrupt the expansion rate. .

[0254] The final constructed spatial destruction vector chain includes the vector path N01->N02 with a length of 15. Speed ​​3 / This vector chain indicates that the damage spread from N01 to N02 over a distance of 15. The average spread rate is 3 per second. .

[0255] The decision output module, based on the length and spread speed of the spatial damage vector chain and the signal amplitude in the chain response data cluster, matches a preset disaster level rule base and directly outputs structured early warning decision instructions containing disaster type, impact range and urgency level without the need for complex model reasoning.

[0256] In a further embodiment of the present invention, the decision output module is specifically configured to perform the following operations:

[0257] The total length of the spatial destruction vector chain and the destruction propagation speed are used as input parameters to query the corresponding basic disaster type and risk level in the disaster level rule base;

[0258] Calculate the average amplitude of the signal from each probe in the cascade response data cluster;

[0259] The risk level is dynamically adjusted based on the average value; the higher the value, the higher the urgency level.

[0260] By integrating and revising the risk level, disaster type, and spatial coverage of the vector path, a structured early warning decision instruction is generated that includes the disaster type, scope of impact, and urgency level.

[0261] Specifically, two key quantization parameters of the obtained spatial destruction vector chain are read and calculated: the total length of the vector path. Simultaneously, along with the average velocity V of the damage propagation, the differential response signal amplitudes recorded by all triggered probes are extracted from the generated chain response data cluster, and the arithmetic mean of these amplitudes is calculated to obtain the average amplitude of the differential signal. .

[0262] Vector chain length speed of spread and the average amplitude of the signal As a set of input parameters, they are fed into a pre-defined disaster level rule base for querying and matching.

[0263] The disaster severity rule base is a structured database or configuration file that predefines multiple decision rules. Each rule consists of a conditional statement and a conclusion instruction. For example, rule one is the determination of the length of the vector chain. Greater than 10 m, and the spread rate More than 2 cm per day is approximately If this occurs, the basic warning level will be set at Level II for landslide risk.

[0264] Input parameters , The system compares each rule's condition with the conditions in the rule base. When all conditions are met, the corresponding rule is matched to determine the basic disaster type and warning level.

[0265] Then, the third input parameter, the average amplitude of the differential signal, is combined. The established basic early warning levels will be dynamically revised.

[0266] The correction logic is also based on the preset amplitude correction sub-rules in the rule base. For example, if the basic level is level two landslide risk and the average signal amplitude is... If the threshold of 0.03 V is exceeded, the emergency warning level will be raised by one level. The higher the average amplitude of the signal, the greater the relative intensity of deformation of the agile medium at the monitoring point, and the higher the corresponding emergency level of the disaster.

[0267] After the above matching and correction, the system will finally determine the disaster type, warning level, and urgency, and combine this with the impact range indicated by the spatial damage vector chain, that is, the spatial coordinate range covered by the vector path, for example, from the starting coordinates. coordinates to the destination It integrates and generates complete and machine-readable text instructions.

[0268] The instruction is a structured early warning decision instruction, which specifically includes the qualitative and quantitative scope of the disaster and disposal recommendations. For example, a level-two regional landslide risk, with the coordinates of the affected area (…). )to( It is recommended to reinforce it.

[0269] Through its communication interface, this structured early warning decision instruction is broadcast to the connected emergency response system, monitoring center, or relevant personnel terminals, completing the entire process from trend monitoring to decision output.

[0270] The disaster level rule base is a pre-compiled knowledge base containing a series of condition-conclusion pairs. Its characteristic attribute is a structured data set. Its function is to map quantitative monitoring parameters to qualitative disaster descriptions and levels. The rules are set based on statistical analysis and solidification of expert experience from more than 1,000 historical geotechnical engineering disaster cases.

[0271] The input parameters refer to the vector chain length L, spread rate V, and average signal amplitude used to query the rule base. Dynamic correction refers to the process of adjusting the basic warning level determined by length and velocity based on the average signal amplitude. Its logical characteristic is a linear or piecewise function relationship, and its setting is based on the physical principle that the signal amplitude is positively correlated with the degree of deformation of the soil and rock.

[0272] The detailed explanation of the disaster level rule base is as follows:

[0273] The disaster level rule base is a structured database or configuration file built upon professional knowledge in the field of geological hazards and numerous practical case studies. It predefines multiple judgment rules to determine the severity of a disaster based on the input vector chain length L, propagation velocity V, and average signal amplitude. The level of a disaster is determined by parameters such as these.

[0274] Each rule consists of a conditional statement and a conclusion instruction. The conditional statement takes specific conditions as input parameters, while the conclusion instruction corresponds to the disaster level determined when the conditions are met. The disaster level rule base is designed to provide a standardized and repeatable judgment process for disaster early warning, enabling the assessment of disaster risk under different circumstances. The vector chain length... speed of spread and the average amplitude of the signal The parameter thresholds are set based on the following criteria:

[0275] Vector chain length Greater than 10 m: Analysis of a large number of landslide cases revealed that when the vector chain length exceeds 10 m, the probability of landslides increases significantly. Longer vector chains usually indicate a larger area of ​​unstable geological structure and potential landslide risk. This threshold is determined based on a comprehensive analysis of actual observation data and geomechanical models, and can screen out areas with landslide risk.

[0276] Spread speed More than 2 cm per day is approximately Studies have shown that the spread rate of landslides is closely related to geological conditions, slope structure, and external triggering factors. When the spread rate exceeds 2 cm per day, it indicates that the landslide speed exceeds the common speed range during the slow and stable deformation stage of landslides, which may pose a threat to the surrounding environment and life and property. This threshold is derived from long-term monitoring data and numerical simulation, which can capture the stage of accelerated landslide development and provide a time window for early warning and response measures.

[0277] The disaster level rule base is constructed and updated in the following ways:

[0278] Construction process: The rule base is built based on the experience of geological experts, historical disaster data and related scientific research. A large amount of landslide case data is collected and organized, including geological parameters before the landslide, monitoring data and the final disaster level. Through data analysis and model building, the relationship between different parameters and disaster levels is determined, and then corresponding judgment rules are formulated.

[0279] With the continuous accumulation of new research findings and actual monitoring data, the rule base needs to be updated and optimized regularly. When new geological phenomena or disaster patterns are discovered, geological experts will evaluate and adjust existing rules or add new rules so that the rule base can always reflect the latest characteristics and risk levels of geological disasters.

[0280] Structured early warning decision instructions are standardized data packets that are the final output containing all key decision information. They have fixed field formats such as type, level, scope, and recommendations, and their function is to drive emergency response actions.

[0281] For example, the spatially disruptive vector chain of the output, such as the length =15 m, spread rate =3 m / s and the output chain response data cluster are used to calculate the average amplitude of the differential signal. .

[0282] Will =15 m =3 m / s Input disaster level rule base, one rule in the rule base is if If the velocity is greater than 10 m and V > 0.001 m / s, then the disaster type is landslide, the basic risk level is level two, and the input parameters meet this condition, so the match is successful. =0.034 V is corrected.

[0283] The rule in the rule base has been revised to apply when the base level is level 2 risk and... When the value is >0.03 V, the urgency level is high, and the final generated structured early warning decision instruction is a level-two regional landslide risk, with the affected area coordinates ( )to( The situation is highly urgent, and reinforcement is recommended. This instruction was broadcast to the emergency center.

[0284] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A geotechnical engineering monitoring trend prediction system based on big data, characterized in that, Includes the following modules: The data acquisition module deploys a dual-core differential probe in the monitoring area and calculates the physical difference between the agile core monitoring value and the reference core monitoring value in real time through an analog subtraction circuit to generate a front-end differential response signal. The threshold judgment module compares the front-end differential response signal with the preset physical trigger threshold, and generates a local warning trigger event when the physical trigger threshold is exceeded. The array activation module sends activation commands to the probes located at downstream nodes along a preset stress transmission path based on local early warning trigger events, so as to form a local high-sensitivity monitoring array. The data integration module collects and integrates the triggering time, geographic coordinates, and differential signal amplitude of the triggered probes in the local high-sensitivity monitoring array to generate a chain response data cluster. The vector analysis module constructs a spatial destruction vector chain based on the triggering timing and spatial location relationship of probes in the chain response data cluster. The decision output module matches the disaster level rule base with the length and spread speed of the spatial damage vector chain and the signal amplitude in the chain response data cluster, and outputs structured early warning decision instructions.

2. The geotechnical engineering monitoring trend prediction system based on big data according to claim 1, characterized in that, The data acquisition module is specifically configured to perform the following operations: The reference core is anchored in the deep, stable rock and soil within the monitoring area to collect background environmental baseline change data, including temperature and vibration. The agile core is encapsulated in a prefabricated agile medium, and the agile medium is deployed in the shallow stress concentration area of ​​the monitoring area to deploy the dual-core differential probe. The agile medium is composed of brittle cement mortar or conductive gel, and its physical strength is configured to be lower than that of the surrounding natural rock and soil. Under the stress of the precursor to the disaster, it will preferentially produce micro-fractures or micro-deformations compared to the surrounding rock and soil.

3. The geotechnical engineering monitoring trend prediction system based on big data according to claim 2, characterized in that, The data acquisition module is also configured to perform the following operations: Connect the outputs of the reference core and the agile core to the two inputs of the analog subtraction circuit, respectively. An analog subtraction circuit is used to perform continuous subtraction of analog voltage signals at two input terminals without analog-to-digital conversion. The output is an analog voltage waveform that has been canceled out by common environmental noise interference, generating a real-time front-end differential response signal.

4. The geotechnical engineering monitoring trend prediction system based on big data according to claim 1, characterized in that, The threshold determination module is specifically configured to perform the following operations: The front-end differential response signal is input to the hardware comparator circuit, whose reference voltage is set to a preset physical trigger threshold. The hardware comparator continuously compares the real-time voltage of the front-end differential response signal with the physical trigger threshold; When the voltage of the front-end differential response signal exceeds the physical trigger threshold, the hardware comparator generates a high-level local warning trigger event.

5. The geotechnical engineering monitoring trend prediction system based on big data according to claim 4, characterized in that, The array activation module is specifically configured to perform the following operations: The physical addresses of downstream nodes are pre-stored in the probe control circuit to construct a potential stress transfer path mapping; When a local warning trigger event occurs, the control circuit bypasses the central server and sends a digital instruction packet containing a wake-up command to the probe of the downstream node corresponding to the physical address via a dedicated communication line as a bypass activation command.

6. The geotechnical engineering monitoring trend prediction system based on big data according to claim 1, characterized in that, The array activation module is also configured to perform the following operations: Based on the geotechnical engineering mechanics model, the physical address of the downstream node is set for each probe in advance, forming a potential stress transmission path; When a master node probe generates a local warning trigger event, its control circuit bypasses the central server and sends a bypass activation command containing wake-up and trigger threshold reduction instructions to the probes of its downstream nodes through a dedicated communication line. Upon receiving this instruction, the probes of the downstream nodes are awakened from sleep mode, and the physical trigger threshold of their internal hardware comparators is temporarily lowered, transforming them into a locally highly sensitive monitoring array with directional activation.

7. The geotechnical engineering monitoring trend prediction system based on big data according to claim 1, characterized in that, The data integration module is specifically configured to perform the following operations: Each activated probe in the local high-sensitivity monitoring array independently runs its front-end differential circuit and hardware comparator to determine the deformation trend of its respective region; The main control unit listens to and records the timestamp of each probe in the array that generates a local early warning trigger event, and simultaneously retrieves its pre-stored geographic coordinates and the amplitude of the differential response signal at the moment of triggering. The timestamps, geographic coordinates, and differential response signal amplitude data of the triggered probes are structurally integrated to generate a chain response data cluster containing spatiotemporal evolution information.

8. The geotechnical engineering monitoring trend prediction system based on big data according to claim 7, characterized in that, The vector analysis module is specifically configured to perform the following operations: Extract the geographic coordinates of the triggered probes from the chain response data cluster and map them in three-dimensional space; Based on the order in which each probe is triggered, the coordinates of adjacent triggered probes are connected to form a vector path; The total length of the vector path and the trigger time difference between the first and last nodes are calculated to obtain the propagation speed and construct a spatial destruction vector chain.

9. A geotechnical engineering monitoring trend prediction system based on big data according to claim 8, characterized in that, The vector analysis module is also configured to perform the following operations: The total length of the vector path is obtained by calculating and summing the straight-line distances between adjacent coordinate points. The propagation speed is then obtained by dividing the total length by the trigger time difference between the first and last nodes.

10. A geotechnical engineering monitoring trend prediction system based on big data according to claim 9, characterized in that, The decision output module is specifically configured to perform the following operations: The total length of the spatial destruction vector chain and the destruction propagation speed are used as input parameters to query the corresponding basic disaster type and risk level in the disaster level rule base; Calculate the average amplitude of the signal from each probe in the cascade response data cluster; The risk level is dynamically adjusted based on the average value; the higher the value, the higher the urgency level. By integrating and revising the risk level, disaster type, and spatial coverage of the vector path, a structured early warning decision instruction is generated that includes the disaster type, scope of impact, and urgency level.

Images