Tunnel underpass ruins of adverse geological warning method and system

Abstract

The present application belongs to the technical field of geotechnical engineering, and specifically relates to a tunnel underpassing a site adverse geological early warning method and system. A multi-source detection network is arranged in a set region of tunnel construction, original data reflecting geological abnormal conditions are obtained and preprocessed, a multi-dimensional feature index set including wave speed, dielectric constant, rock mass integrity coefficient, etc. is obtained, a three-dimensional perspective model containing spatial distribution and physical properties of adverse geological bodies is generated through a data fusion algorithm, and a geological anomaly grade G is determined accordingly. The structure vulnerability grade V of a target historical site is obtained, and a disturbance influence coefficient S of the site by current construction operation is calculated based on construction parameters. The geological anomaly grade G, the site vulnerability grade V and the construction influence coefficient S are input into a preset risk coupling evaluation model, a comprehensive risk value R is dynamically calculated, and corresponding graded early warning is triggered according to the threshold interval of the R value.

Description

Technical Field

[0001] This invention belongs to the field of geotechnical engineering technology, and specifically relates to a method for ensuring the safety of historical sites. Background Technology

[0002] The statements in this section merely refer to the background art related to this invention and do not necessarily constitute prior art.

[0003] Tunnel engineering faces extremely stringent safety constraints when traversing historical and cultural heritage sites. These sites possess inherent vulnerabilities due to their age, material deterioration, unique construction techniques, and varying preservation conditions. Ground disturbances caused by construction activities (such as blasting and mechanical excavation) can easily induce irreversible damage to the sites, such as structural cracking, foundation settlement, and damage to artifacts caused by earthquakes. Current risk assessments for detected adverse geological features are largely based on pre-construction geological survey reports, which are static and isolated evaluations. These reports fail to consider the inherent vulnerabilities of the site structures due to their varying ages, materials, construction techniques, and preservation conditions, and also fail to incorporate the time-varying disturbances caused by tunnel excavation (such as blasting and mechanical excavation) and support operations. Consequently, early warnings are often triggered only after deformation or damage has become apparent, resulting in a delay. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method and system for early warning of adverse geological conditions when tunnels pass under archaeological sites. By constructing a quantitative risk assessment model that couples "geological anomaly (G) - site vulnerability (V) - construction dynamic disturbance (S)", it achieves dynamic early warning across the entire chain from data acquisition and fusion imaging to risk assessment, ensuring the safety of the archaeological site in real time during construction.

[0005] The first aspect of this invention discloses a method for early warning of adverse geological conditions when a tunnel passes under an archaeological site, comprising the following steps: A multi-source detection network is deployed in the designated area of ​​tunnel construction to obtain raw data reflecting geological anomalies. The raw data includes at least one of seismic waves, electromagnetic waves, and measurement-while-drilling data. The raw data is preprocessed and features are extracted to obtain a set of multi-dimensional feature indicators, including wave velocity, dielectric constant, and rock mass integrity coefficient. Based on a multi-dimensional feature index set, a three-dimensional stereoscopic perspective model containing the spatial distribution and physical properties of adverse geological bodies is generated through a data fusion algorithm, and the geological anomaly level G is determined accordingly. Obtain the structural vulnerability level V of the target historical site, and calculate the disturbance impact coefficient S of the current construction operation on the site based on the construction parameters; The geological anomaly level G, the site vulnerability level V, and the construction impact coefficient S are input into the preset risk coupling assessment model to dynamically calculate the comprehensive risk value R, and trigger corresponding graded early warnings based on the threshold range of the R value. The risk coupling assessment model uses a weighted scoring model to calculate the comprehensive risk value. During the calculation of the comprehensive risk value, the weight corresponding to the geological anomaly level G is higher than the weight corresponding to the site vulnerability level V, and the weight corresponding to the site vulnerability level V is equal to the weight corresponding to the construction impact coefficient S.

[0006] Furthermore, the multi-source detection network includes a controllable mechanical source refined seismic detection module, a fully polarized borehole radar transceiver antenna module, a measurement-while-drilling module, and an environmental background noise monitoring module; the controllable mechanical source refined seismic detection module is used to excite and receive seismic wave signals, and its excitation energy and frequency are adjustable, and the sources are synchronized through GPS.

[0007] Further preprocessing and feature extraction include: Wavelet threshold denoising and adaptive median filtering are performed on the original data; P-wave velocity, S-wave velocity and attenuation coefficient are extracted from the denoised seismic data; dielectric constant, conductivity and scattering coefficient are extracted from the radar data; rock drillability index and rock mass integrity coefficient are calculated from the measurement-while-drilling data. The Kriging interpolation algorithm is used to uniformly resample multi-dimensional feature indicators to a standard three-dimensional mesh.

[0008] Furthermore, the data fusion algorithm employs an image fusion model based on a deep convolutional neural network to perform pixel-level fusion of seismic impedance inversion results, radar dielectric constant distribution, and drilling lithology parameters to generate a three-dimensional perspective model.

[0009] Furthermore, obtaining the structural vulnerability level V of the target historical site includes the following steps: Based on a pre-stored site archive database, multiple evaluation factors for the target site are obtained. The evaluation factors include at least the site type, construction date, average material strength, preservation status rating, historical protection level, structural integrity coefficient, environmental sensitivity, archaeological value level, and existing reinforcement measures. The weights of each evaluation factor were determined using the analytic hierarchy process (AHP). The fuzzy comprehensive evaluation method is used to fuzzify each evaluation factor individually, and a comprehensive evaluation is performed based on the weights to output the structural vulnerability level V.

[0010] Furthermore, based on the construction parameters, the disturbance impact coefficient S of the current construction operation on the site is calculated, as shown in the following formula: S= + + ; in, , , All are weighting coefficients; Q is the charge per stage, V is the blasting vibration velocity, D is the average daily advance, and Q0, V0, and D0 are preset benchmark parameters according to specifications and protection requirements.

[0011] Furthermore, in the triggered graded early warning, different levels of early warning instructions have corresponding construction response measures, which include at least one of the following: adjusting the monitoring frequency, limiting blasting parameters, controlling the tunneling advance, and stopping construction.

[0012] A second aspect of the present invention discloses an early warning system for adverse geological conditions when a tunnel passes under an archaeological site, comprising: The data acquisition module is configured to: deploy a multi-source detection network in a designated area of ​​tunnel construction to acquire raw data reflecting geological anomalies, including at least one of seismic waves, electromagnetic waves, and drilling measurement data. The data processing and feature extraction module is configured to preprocess and extract features from the raw data to obtain a set of multi-dimensional feature indicators, including wave velocity, dielectric constant, and rock mass integrity coefficient. The data fusion imaging module is configured to: generate a three-dimensional stereoscopic perspective model containing the spatial distribution and physical properties of adverse geological bodies based on a multi-dimensional feature index set and a data fusion algorithm, and determine the geological anomaly level G accordingly; The risk assessment and early warning module is configured to: obtain the structural vulnerability level V of the target historical site, and calculate the disturbance impact coefficient S of the current construction operation on the site based on the construction parameters; The risk assessment and early warning module is also configured to: input the geological anomaly level G, the site vulnerability level V and the construction impact coefficient S into the preset risk coupling assessment model, dynamically calculate the comprehensive risk value R, and trigger the corresponding graded early warning according to the threshold range of the R value; The risk coupling assessment model uses a weighted scoring model to calculate the comprehensive risk value. During the calculation of the comprehensive risk value, the weight corresponding to the geological anomaly level G is higher than the weight corresponding to the site vulnerability level V, and the weight corresponding to the site vulnerability level V is equal to the weight corresponding to the construction impact coefficient S.

[0013] A third aspect of the present invention discloses a computer program product comprising computer-readable instructions, which, when executed on an electronic device, enable the electronic device to implement the aforementioned method for early warning of adverse geological conditions during tunneling under archaeological sites.

[0014] A fourth aspect of the present invention discloses an electronic device, including at least one processor and a memory connected to the processor, the memory being used to store a computer program; the processor being used to execute the computer program, enabling the electronic device to implement the aforementioned method for early warning of adverse geological conditions when a tunnel passes under an archaeological site.

[0015] Compared with existing technologies, one or more of the above technical solutions have the following beneficial effects: 1. By constructing a quantitative risk assessment model that couples the geological anomaly (G), the site's inherent vulnerability (V), and construction dynamic disturbance (S), the previously isolated pre-construction geological assessment, site protection requirements, and real-time construction parameters are integrated. As construction progresses, the comprehensive risk value R is dynamically calculated and graded early warnings are triggered, transforming risk management from "post-event judgment" to "pre-event prediction and in-event control." This solves the problem of the static risk assessment model being disconnected from the dynamic construction process and the problem of delayed early warnings, providing a quantitative decision-making basis for ensuring the safety of the site during construction.

[0016] 2. By deeply fusing multi-source data such as seismic waves, electromagnetic waves, and drilling parameters, and extracting multi-dimensional feature indicators directly related to soil and rock mechanical properties and water-bearing state, a three-dimensional perspective model with physical properties was constructed. This model not only has high spatial resolution but can also distinguish the subtle differences in physical response between "natural adverse geology" and "architectural structure," greatly reducing the ambiguity of single methods.

[0017] 3. By storing and integrating multi-source detection data, three-dimensional fusion models, dynamic risk assessment results and early warning information in a unified three-dimensional visualization environment, decision-makers can intuitively and in real time grasp the spatial relationship between adverse geological bodies and sites and the dynamic evolution of risks. Attached Figure Description

[0018] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0019] Figure 1 A schematic diagram illustrating the adverse geological warning process for a tunnel passing under an archaeological site, provided for one or more embodiments of the present invention; Figure 2 A schematic diagram of an early warning system architecture for a tunnel passing under an archaeological site, provided for one or more embodiments of the present invention; Figure 3 A schematic diagram of the functional architecture of an early warning system for adverse geological conditions under a tunnel passing through an archaeological site, provided in one or more embodiments of the present invention; Figure 4 This is a schematic diagram of the architecture of an electronic device provided for one or more embodiments of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0022] Tunnel projects inevitably require traversing complex historical and cultural sites and adjacent sensitive facilities such as existing railways. From a construction perspective, the historical and cultural sites and adjacent existing railways traversed by the tunnel project are equivalent to unfavorable geological formations (such as cavities, faults, and weak interlayers) during construction. These unfavorable geological formations are highly susceptible to ground deformation, collapses, and water inrushes under construction disturbances, directly threatening the long-term safety of the site structures and the stable operation of the existing railways.

[0023] Traditional geophysical methods (such as high-density electrical resistivity tomography and ground-based seismic reflection methods) often require significant excitation energy or the deployment of large-scale observation systems to achieve sufficient depth. The resulting vibrations, electromagnetic fields, or interventions in the site itself pose a potential threat to fragile archaeological sites. While lightweight and non-destructive methods (such as ground-penetrating radar) are site-friendly, their depth, resolution, and interference resistance are limited, especially in imaging complex underground structures where their reliability is insufficient.

[0024] Secondly, current risk assessments are mostly based on pre-construction geological survey reports, resulting in a static and isolated evaluation. They fail to consider the inherent vulnerability of the site's structure due to variations in its age, materials, construction techniques, and preservation status, nor do they incorporate the time-varying disturbances caused by construction activities such as tunneling (e.g., blasting, mechanical excavation) and support. For example, a small, low-risk fissure in an ordinary mountain might be a high-risk point for a rammed-earth site that is thousands of years old; a blast with conventional parameters could trigger irreversible damage under specific geological and site vulnerability conditions. The lack of a quantitative risk assessment model that can dynamically couple and analyze "geological anomalies," "site vulnerability," and "intensity of instantaneous construction disturbance" means that warnings are often triggered only after deformation or damage has occurred, which is too late.

[0025] Meanwhile, the geological conditions in the construction area are complex. Adverse geological features (such as small-scale cavities, seepage channels, and weak interlayers) often exhibit similar physical responses to the site's structural features (such as rammed earth cracks, burial chamber cavities, and loose foundations). Relying solely on single parameters such as seismic wave velocity or radar reflection waves makes it difficult to accurately distinguish between "natural adverse geological features" and "site structural features," easily leading to misjudgments. Existing technologies fail to extract key integrated indicators that directly characterize the mechanical properties (such as strength and integrity) and water content of soil and rock from the perspective of multi-physics field (elastic wave field, electromagnetic wave field) joint inversion. This results in risk source identification remaining at the level of vaguely delineating "abnormal areas" rather than accurately diagnosing the "essence of the problem." Furthermore, geological data, monitoring data, site archives, and construction parameters are scattered across different units and systems, with varying formats, making integration difficult. Decision-makers cannot intuitively see the spatial relationship between adverse geological features and key parts of the site, as well as the dynamic changes in risk levels during the construction process, within a unified, spatiotemporally synchronized three-dimensional scene. Inefficient information transmission severely impacts the ability to conduct multi-departmental collaborative consultations and rapid responses in emergency situations.

[0026] Therefore, this solution provides a method and system for early warning of adverse geological conditions when tunnels pass under historical sites. Under the premise of ensuring the absolute safety of the historical sites, it enables precise identification and dynamic risk warning of adverse geological bodies in the vicinity of tunnel construction. By constructing a quantitative dynamic risk assessment model that couples "geological anomalies (G) - site vulnerability (V) - construction dynamic disturbance (S)," and incorporating a "one-vote veto" safety redundancy design and an adaptive feedback closed loop of "assessment-early warning-control-reassessment," a full-chain, intelligent early warning and control system is formed, encompassing multi-source data collaborative acquisition, cross-physical field fusion imaging, dynamic risk coupling judgment, and intelligent control of construction parameters.

[0027] Example 1: The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site includes the following steps: A multi-source detection network was deployed in the area adjacent to the tunnel construction site and in the sensitive area of ​​the site to simultaneously acquire raw data reflecting the geological and structural anomalies of the site. The raw data included at least one of seismic waves, electromagnetic waves and drilling measurement data. The raw data is preprocessed and features are extracted to obtain a set of multi-dimensional feature indicators, including wave velocity, dielectric constant, and rock mass integrity coefficient. Based on a multi-dimensional feature index set, a three-dimensional perspective model is generated by a cross-physical field data fusion algorithm, which simultaneously includes the spatial distribution and physical attributes of adverse geological bodies and site structures. Based on preset distinction criteria and quantification thresholds, the geological anomaly level G is determined. The structural vulnerability level V of the target historical site is obtained. Level V is dynamically calculated based on site archive data and construction disturbance sensitivity model. Based on real-time construction parameters and dynamic safety thresholds, the disturbance impact coefficient S of the current construction operation on the site is calculated. The geological anomaly level G, the site vulnerability level V, and the construction impact coefficient S are input into the preset risk coupling assessment model to dynamically calculate the comprehensive risk value R. Based on the threshold range of the R value and the preset "one-vote veto" safety rules, corresponding graded early warning and construction control instructions are triggered. Among them, the risk coupling assessment model uses a weighted scoring model to calculate the comprehensive risk value R. The weight design follows the principle of "risk source dominance, protection and disturbance equal emphasis" and can be dynamically adjusted according to the vulnerability level V of the site.

[0028] As a further implementation method, the process of generating a three-dimensional perspective model specifically includes: Data registration and gridding: Multi-source heterogeneous data are uniformly resampled to a standard three-dimensional grid (e.g., 1m×1m×1m) using the Kriging interpolation algorithm to establish a one-to-one spatial correspondence. Cross-physics joint feature extraction and correlation: Based on a rock physics model, establish the correlation between seismic wave velocity (Vp, Vs) and rock mass strength and integrity, and establish the dielectric constant (Vp, Vs). The correlation between drilling parameters (such as torque and vibration) and water content and porosity was established, and the correlation between drilling parameters and rock drillability and integrity was established. Using a coupled inversion framework of the Zoeppritz equation and the Maxwell equation, elastic and electrical parameters were simultaneously inverted, reducing the ambiguity of single inversions. Multi-source data pixel-level fusion: An image fusion model based on deep convolutional neural networks (DCNN) (such as the U-Net structure) is adopted to fuse seismic impedance inversion results, radar dielectric constant distribution, and drilling lithology parameters at the feature level and decision level. The network is trained with a large number of "site-non-site" sample pairs to enhance its ability to identify site features; 3D Reconstruction and Visualization: The improved Marching Cubes algorithm is used to extract isosurfaces from the fused physical property parameter volume to generate a high-resolution (e.g., 0.5m) 3D mesh model, and colors and transparency are assigned to distinguish different geological bodies and site structures.

[0029] As a further implementation method, the determination of geological anomaly level G includes specific criteria for distinguishing between "natural adverse geology" and "the structure of the site itself": Feature-based collaborative clustering: The fuzzy C-means clustering (FCM) algorithm is used to perform cluster analysis on multi-dimensional features of grid cells, such as wave velocity, dielectric constant, attenuation coefficient, and anisotropy parameters. Prior knowledge constraints: The known distribution range and structural type (e.g., rammed earth, brick and stone, burial chamber) of the site are input as constraints into the clustering process. For example, for a known rammed earth site area, its wave velocity is usually lower than that of intact bedrock but higher than that of loose soil layers, and its dielectric constant has a specific range; Quantization threshold determination: Set a comprehensive discrimination threshold. For example, when a unit simultaneously satisfies: V p <1800 m / s, If the value is between 6 and 9, and located within the prior area of ​​the site, it is determined to be a "fracture in the rammed earth of the site"; if V p <1500 m / s, If the anomaly is greater than 12 and is not located within the site area, it is classified as a "water-rich fractured zone". Based on factors such as the type, size, and distance from the site of the anomaly, G is divided into levels 1-5.

[0030] As a further implementation method, obtaining the structural vulnerability level V of the target historical site specifically includes the following steps: Construct a vulnerability assessment factor system for archaeological sites, with factors including at least: site type, structural form, construction date, average material strength, preservation status rating, historical protection level, structural integrity coefficient, environmental sensitivity, archaeological value level, existing reinforcement measures, and spatial relationship with the tunnel (distance and orientation).

[0031] Factor engineering quantification: Construction date: mapped to the material weathering coefficient K age This reduces the dynamic elastic modulus E. dynamic ' = K age E dynamic .

[0032] Structural form: mapped to a stress diffusion model (e.g., arched and strip foundations correspond to different stress influence angles α), used to determine the sensitive areas of construction disturbance.

[0033] Preservation status rating: mapped to the fracture development index F crack This is used to set the alarm threshold for wave velocity attenuation.

[0034] Historical preservation level / archaeological value level: As a weighting factor, it reflects its importance in the comprehensive evaluation.

[0035] Dynamic weight determination: A combination of the Analytic Hierarchy Process (AHP) and entropy weighting is used for weighting. AHP reflects the experts' judgment on the relative importance of factors, while entropy weighting objectively adjusts the weights based on the dispersion of each factor's data. In addition, a construction disturbance sensitivity matrix is ​​introduced to dynamically fine-tune the weights of each factor based on the tunnel construction method (blasting / shield tunneling), relative location, etc.

[0036] Fuzzy comprehensive evaluation: The fuzzy comprehensive evaluation method is used to fuzzify each evaluation factor individually and to make a comprehensive evaluation based on dynamic weights, outputting the structural vulnerability level V (levels 1-4, corresponding to low, medium, high, and extremely high, respectively).

[0037] As a further implementation method, the approach also includes mechanisms to address missing site data: Before construction begins, if there is no detailed record of the target site, the "similar site comparison assessment" process is initiated, and the data of the most similar known site is selected as the initial input based on the site type, region, and age.

[0038] During construction, a high-precision monitoring network (such as laser scanning and photogrammetry) is deployed facing the site to acquire information on the structure, size, and materials of newly revealed sites in real time, and trigger the "vulnerability assessment dynamic update" process to incrementally integrate new data into the assessment model and optimize the V value.

[0039] As a further implementation method, the disturbance impact coefficient S of the current construction operation on the site is calculated based on the construction parameters, and dynamic safety threshold setting and feedback control are realized: The formula for calculating the disturbance influence coefficient S is as follows: S= + + ; in, , , All are weighting coefficients; Q is the charge per stage, V is the blasting vibration velocity, D is the average daily advance, and Q0, V0, and D0 are preset benchmark parameters according to specifications and protection requirements.

[0040] Dynamic safety thresholds: Q0, V0, and D0 are not fixed values, but rather functions of the site's vulnerability level V. The system has a built-in "threshold inversion module" that dynamically inverts and updates the site's safe vibration threshold V0(V) and corresponding Q0(V) in the current state based on real-time vibration response monitoring data.

[0041] Feedback-based control loop: When an early warning is triggered, the system not only issues an alarm but also automatically generates "construction parameter adjustment suggestions" (such as reducing the single-stage chemical dosage to Q and adjusting the advance to D). After the construction party adjusts the parameters, the system re-collects data, calculates the new S value, and forms a closed loop of "monitoring-early warning-control-reassessment" until the risk is reduced to an acceptable level.

[0042] As a further implementation method, the calculation and dynamic adjustment of weights in the weighted scoring model of the risk coupling assessment model are as follows: Overall risk value R = ω g (V)×G+ω v (V)×V+ωS ×S.

[0043] Weight design principles and dynamic adjustment: Basic principle: Set ω g >ω v ≈ω S ω g >ω v ≈ω S This reflects the engineering understanding that "risk sources (adverse geological conditions) are the root cause of disasters, while the vulnerability of the site and construction disturbances are the conditions."

[0044] Dynamic adjustment mechanism: weight ω g and ω V It is a function of V. As the vulnerability level V of the site increases (≥3), ω is gradually increased. V The weighting reflects a preference for extra protection of highly vulnerable sites. s Relatively stable, reflecting the intensity of the construction disturbance itself.

[0045] "One-vote veto" rule: To ensure "absolute security," the system presets several conditions that directly trigger the highest level (red) warning, for example: a) The 3D model identified a cavity or a severely water-rich area directly beneath the foundation of the site (overlapping projection area); b) The minimum clearance between the key load-bearing structure of the site and the tunnel excavation face is less than the critical safety distance dynamically calculated based on the V value; c) Real-time monitoring shows that the vibration velocity or deformation rate of the site structure exceeds a certain high percentage (e.g., 80%) of its material limit.

[0046] If any condition is met, regardless of the R value, a red alert will be triggered immediately and the strictest control measures will be implemented.

[0047] As a further implementation method, the approach also includes a reverse optimization process when new relics are discovered during construction: When unrecorded sites or archaeological information are discovered during construction, the following procedures shall be initiated: 1. Obtain geometric and physical characteristic data of new relics through rapid on-site detection (such as ground-penetrating radar and 3D scanning).

[0048] 2. Call the "Similar Site Analogy Assessment" module to generate its preliminary vulnerability assessment V_initial.

[0049] 3. Integrate the spatial location and attribute information of the new relics into the existing three-dimensional perspective model.

[0050] 4. Based on the updated model, reassess the risks of the entire construction area and make global optimization and adjustment to the early warning thresholds and construction plans for subsequent construction sections.

[0051] 5. The system records the discovery and processing process to enrich the sample library and optimize future evaluation models.

[0052] This solution achieves precise early warning under the premise of absolute safety of the site: through the “GVS” ternary dynamic coupling model, cross-physical field fusion differentiation technology, “one-vote veto” safety rules and adaptive threshold adjustment, risk management is upgraded from passive response to accurate pre-event prediction and active in-event control, providing quantifiable technical support for the protection of core cultural relics.

[0053] This solution breaks through the contradiction between "detection and safety" and the dilemma of "identification confusion": it adopts site-friendly detection methods such as controllable low-energy seismic sources and borehole radar, and effectively distinguishes between natural adverse geology and site structure through joint inversion and deep learning fusion, thus significantly reducing the misjudgment rate.

[0054] This solution constructs a dynamic closed loop of "monitoring-assessment-early warning-control": the system can not only dynamically assess risks, but also generate suggestions for adjusting construction parameters based on early warning results, and optimize its own model through feedback data, thus forming an intelligent control capability that adapts to changes in the construction process.

[0055] This solution is capable of handling uncertainties: through the "analogy with similar sites" and "dynamic update" mechanisms, it effectively solves the problems of missing site data or newly discovered relics during construction, thereby enhancing the system's engineering practicality and robustness.

[0056] This solution provides an intuitive and efficient collaborative decision-making platform: all information is integrated into a unified 3D visualization platform, enabling all parties to intuitively and in real time grasp the risk situation, greatly improving the efficiency of multi-departmental collaborative consultation and emergency response.

[0057] Example 2: An early warning system for adverse geological conditions when the tunnel passes under the archaeological site, such as... Figure 1 As shown, it includes a data acquisition module 100, a data processing and feature extraction module 200, a data fusion imaging module 300, a risk assessment and early warning module 400, and a data storage and visualization module 500.

[0058] like Figure 2 As shown, the data acquisition module 100 is used to simultaneously acquire geophysical data such as seismic waves and electromagnetic waves through a multi-source detection network deployed in the tunnel construction area. In actual operation, the data acquisition module 100 includes various detection devices.

[0059] For example, the controllable mechanical seismic source detection module 101 adopts a hydraulic servo control system, which can accurately control the excitation energy (0.5-5kJ adjustable) and excitation frequency (10-200Hz) of the seismic source, and realize the synchronous excitation of multiple seismic sources through GPS synchronization technology.

[0060] For example, the fully polarized drilling radar antenna module 102 adopts a broadband dipole antenna design with an operating frequency range of 50-500MHz. It achieves fully polarized data acquisition through a polarization switching circuit and can simultaneously acquire electromagnetic wave data in four polarization modes: VV, VH, HV, and HH.

[0061] For example, the measurement-while-drilling module 103 is integrated behind the drill bit and includes a triaxial accelerometer, a magnetometer, and a gamma ray sensor. It has a sampling frequency of up to 100Hz and can transmit drill bit attitude, drilling speed, and lithological characteristic parameters in real time.

[0062] For example, the environmental noise monitoring module 104 uses a low-frequency velocity sensor (0.1-100Hz) and an electromagnetic field probe to establish a background interference database, providing a benchmark for effective signal extraction.

[0063] like Figure 2 As shown, the data processing and feature extraction module 200 is used to preprocess the raw data transmitted from the data acquisition module 100 and extract multi-dimensional feature indicators characterizing geological properties. Optionally, in one embodiment of this application, the data processing and feature extraction module 200 includes: a data preprocessing unit 201, a feature extraction unit 202, and a resampling unit 203.

[0064] The data preprocessing unit 201 performs denoising processing on the original signal. For seismic data, it adopts a wavelet threshold denoising algorithm, and for radar data, it adopts adaptive median filtering. At the same time, it performs time delay correction and amplitude recovery to improve the data signal-to-noise ratio.

[0065] Feature extraction unit 202 extracts multiple feature indices from the preprocessed data: for seismic data, it calculates the P-wave velocity V. p Shear wave velocity V s Speed ​​ratio V p / V s Attenuation coefficient Q value, etc.; for radar data, extract the dielectric constant ε. r Parameters include electrical conductivity σ, scattering coefficient, and anisotropy parameters; for drilling data, rock drillability index and rock mass integrity coefficient are calculated.

[0066] The resampling unit 203 uses the Kriging interpolation algorithm to unify multi-source data into a standard grid of 1m×1m×1m, and achieves resolution normalization through bicubic interpolation to eliminate scale differences caused by different detection principles.

[0067] like Figure 2 As shown, the data fusion imaging module 300 is used to determine the contribution weight of each feature index in the fusion and execute a multi-source data fusion imaging algorithm to generate a three-dimensional stereoscopic perspective model. Optionally, in one embodiment of this application, the data fusion imaging module 300 includes: a feature reconstruction unit 301, a fusion imaging unit 302, and a model optimization unit 303.

[0068] The feature reconstruction unit 301 uses a combination of principal component analysis (PCA) and autoencoder to reduce the extracted 25-dimensional original features to 8-dimensional principal features, retaining more than 95% of the original information while enhancing the discriminative power of the features.

[0069] The fusion imaging unit 302 establishes a fusion model based on a deep convolutional neural network (DCNN). The network structure includes 12 convolutional layers and 4 pooling layers, and skip connections are used to prevent gradient vanishing. This unit performs pixel-level fusion of multi-source imaging results, such as seismic impedance inversion results, radar dielectric constant distribution, and drilling lithology parameters, to generate a three-dimensional stereoscopic perspective model with a resolution of 0.5m×0.5m×0.5m.

[0070] The model optimization unit 303 adopts a method that combines an improved genetic algorithm (IGA) with least squares inversion. Using borehole verification data as constraints, it continuously corrects the parameters of the fusion model through iterative optimization, thereby improving the consistency between the model and the measured data to over 90%.

[0071] like Figure 2 As shown, the risk assessment and early warning module 400 is used to identify unfavorable geological bodies based on a three-dimensional perspective model and to determine the risk level by combining the vulnerability of the site and construction dynamics. Optionally, in one embodiment of this application, the risk assessment and early warning module 400 includes: a geological anomaly identification unit 401, a vulnerability query unit 402, a construction dynamics assessment unit 403, a comprehensive risk determination unit 404, and an early warning response unit 405.

[0072] The geological anomaly identification unit 401 employs the fuzzy C-means clustering (FCM) algorithm to classify geological bodies into five categories based on characteristic parameters such as wave velocity and dielectric constant: intact rock mass, fractured zone, fracture zone, cavity, and water-rich zone. An anomaly identification threshold is set: when the wave velocity is below 2000 m / s and the dielectric constant is above 8, it is identified as a high-risk, unfavorable geological body.

[0073] Vulnerability query unit 402 is used to determine the vulnerability level V (levels 1-4) of a site based on a pre-stored site archive database.

[0074] This unit executes an automated rating process based on the fuzzy comprehensive evaluation method, including data input, weight loading, single-factor fuzzification, comprehensive evaluation, and grade determination.

[0075] The following describes the specific process of automated rating.

[0076] Data input: Read the specific values ​​of 10 evaluation factors for the target site from the database, including: site type, structural form, construction date, average material strength, preservation status rating, historical protection level, structural integrity coefficient, environmental sensitivity, archaeological value level, and existing reinforcement measures.

[0077] Weight loading: Calls the factor weight vector A pre-defined by the analytic hierarchy process (AHP).

[0078] Single-factor fuzzification: For each factor, based on its type, call the corresponding membership function (quantitative factor) or query the membership table (qualitative factor) to obtain its membership vector R to the four vulnerability levels (low, medium, relatively high, and high). i .

[0079] Comprehensive evaluation: The membership vectors of the 10 factors are used to form a fuzzy relation matrix R. The comprehensive fuzzy evaluation vector B is obtained by performing the fuzzy synthesis operation B = AR.

[0080] Vulnerability Level Determination: Based on the principle of maximum membership or the weighted average method, the final vulnerability level V is determined by vector B and output to the comprehensive risk assessment unit.

[0081] Construction dynamic evaluation unit 403 calculates the construction influence coefficient S based on real-time acquired drilling and blasting parameters (single-stage charge Q, blasting vibration velocity V, and average daily advance D). The baseline parameters (Q0, V0, D0) are pre-set according to the "Blasting Safety Regulations," the site protection assessment report, and the construction organization design. To more scientifically characterize the impact of blasting vibration, this unit uses a simplified model based on the Sadovsky formula to calculate the S value: S= + + ; Among them, the weighting coefficient ( , , The intensity of construction disturbance can be initially determined using the expert-analytic hierarchy process and can be dynamically fine-tuned in the early stages of construction based on the correlation of monitoring data.

[0082] The comprehensive risk assessment unit 404 establishes a three-dimensional risk decision-making model. This model uses the geological anomaly level G, the site vulnerability level V, and the construction impact coefficient S as inputs. Preliminary risk screening is performed using a pre-set expert rule base (If-Then rules), and extremely high-risk combinations (such as G≥4 and V≥4) are directly classified as red alerts. For most cases, a weighted scoring model is used to calculate the comprehensive risk value R. R= ; Weights ( , , Based on the principle of "risk source dominance, with equal emphasis on protection and disturbance," the analytic hierarchy process was used to determine the requirements. > ≈ And the sum is 1 (for example, a possible set of values ​​is 0.5, 0.25, 0.25). Finally, the risk level is determined based on the range of R values ​​(e.g., R<2.0 is blue, 2.0≤R<3.0 is yellow, 3.0≤R<4.0 is orange, R≥4.0 is red), and the corresponding warning is triggered.

[0083] The early warning response unit 405 triggers corresponding early warnings based on the final risk level: blue (R<2.5, low risk, normal construction), yellow (2.5≤R<4.0, medium risk, increase monitoring frequency to once / 4 hours), orange (4.0≤R<5.0, higher risk, adjust blasting parameters and limit advance), and red (R≥5.0, high risk, immediately stop construction and activate the emergency plan). Early warning information is pushed in real time through multiple channels such as the monitoring platform, SMS, and audible and visual alarms.

[0084] The data storage and visualization module 500 stores all raw data, feature data, fused imaging results, and risk assessment records, establishing a complete tunnel geological database. This module utilizes a WebGL-based 3D visualization platform to enable multi-angle rotation, cross-sectional analysis, and transparent display of the stereoscopic perspective model. It supports dynamic marking of risk areas and comparative analysis of historical data, providing intuitive technical support for construction decisions.

[0085] like Figure 3 As shown, the method for early warning of adverse geological conditions when a tunnel passes under an archaeological site includes the following steps: Step S301: Geophysical data is acquired synchronously through a multi-source detection network. A controllable source is used to generate seismic waves, a fully polarimetric radar is used to acquire electromagnetic responses, lithological parameters are obtained through drilling measurements, and background noise is recorded by environmental monitoring to establish a multi-source data acquisition system. Step S302 involves preprocessing the vibration data and extracting multi-dimensional feature indicators. Data quality is improved through wavelet denoising and median filtering, and feature parameters such as wave velocity and dielectric constant are extracted. Kriging interpolation is then used to achieve data gridding and unification. Step S303: A 3D perspective model is generated using a deep learning fusion algorithm. Feature reconstruction is performed using PCA and an autoencoder, pixel-level fusion of multi-source data is achieved based on DCNN, and model parameters are optimized using a genetic algorithm. Step S304: Determine the risk level by combining the site's vulnerability and construction dynamics. Fuzzy clustering is used to identify unfavorable geological features, the site's vulnerability level is queried, the construction impact coefficient is calculated, and the final risk level is determined through a risk matrix. Step S305: Trigger tiered early warning and output results. Based on the risk level, activate four levels of early warning: blue, yellow, orange, and red. Push early warning information through multiple channels to guide construction decisions.

[0086] According to the system and method proposed in the embodiments of this application, through multi-source data fusion imaging and stereoscopic perspective technology, the system enables precise detection and risk assessment of adverse geological conditions when tunnels pass under archaeological sites, improves the accuracy of geological anomaly identification to over 85%, and shortens the risk warning response time to less than 5 minutes, significantly improving construction safety and archaeological site protection.

[0087] Example 3: Figure 4 This is a schematic diagram of the structure of the electronic device provided in this embodiment. The electronic device may include: Memory 1, processor 2, and computer program stored on memory 1 and capable of running on processor 2.

[0088] When processor 2 executes the program, it implements the multi-index fusion alarm level calculation and risk level assessment method for vibration of historical sites provided in the above embodiments.

[0089] Furthermore, electronic devices also include: Communication interface 3 is used for communication between memory 1 and processor 2.

[0090] Memory 1 is used to store computer programs that can run on processor 2.

[0091] Memory 1 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0092] If memory 1, processor 2, and communication interface 3 are implemented independently, then communication interface 3, memory 1, and processor 2 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized into address buses, data buses, control buses, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0093] Optionally, in a specific implementation, if the memory 1, processor 2, and communication interface 3 are integrated on a single chip, then the memory 1, processor 2, and communication interface 3 can communicate with each other through an internal interface.

[0094] Processor 2 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0095] Example 4: A computer program product includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the aforementioned method for early warning of adverse geological conditions when a tunnel passes under an archaeological site.

[0096] The above are merely preferred embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for early warning of adverse geological conditions when a tunnel passes under an archaeological site, characterized in that: Includes the following steps: A multi-source detection network is deployed in the designated area of ​​tunnel construction to obtain raw data reflecting geological anomalies. The raw data includes at least one of seismic waves, electromagnetic waves, and drilling measurement data. The raw data is preprocessed and features are extracted to obtain a set of multi-dimensional feature indicators, including wave velocity, dielectric constant, and rock mass integrity coefficient. Based on a multi-dimensional feature index set, a three-dimensional stereoscopic perspective model containing the spatial distribution and physical properties of adverse geological bodies is generated through a data fusion algorithm, and the geological anomaly level G is determined accordingly. Obtain the structural vulnerability level V of the target historical site, and calculate the disturbance impact coefficient S of the current construction operation on the site based on the construction parameters; The geological anomaly level G, the site vulnerability level V, and the construction impact coefficient S are input into the preset risk coupling assessment model to dynamically calculate the comprehensive risk value R, and trigger corresponding graded early warnings based on the threshold range of the R value. The risk coupling assessment model uses a weighted scoring model to calculate the comprehensive risk value. During the calculation of the comprehensive risk value, the weight corresponding to the geological anomaly level G is higher than the weight corresponding to the site vulnerability level V, and the weight corresponding to the site vulnerability level V is equal to the weight corresponding to the construction impact coefficient S.

2. The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site as described in claim 1, characterized in that, The multi-source detection network includes a controllable mechanical source fine-grained seismic detection module, a fully polarized borehole radar transceiver antenna module, a measurement-while-drilling module, and an environmental background noise monitoring module. The controllable mechanical source fine-grained seismic detection module is used to excite and receive seismic wave signals. Its excitation energy and frequency are adjustable, and the sources are synchronized through GPS.

3. The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site as described in claim 1, characterized in that, Preprocessing and feature extraction include: Wavelet threshold denoising and adaptive median filtering are performed on the original data; P-wave velocity, S-wave velocity and attenuation coefficient are extracted from the denoised seismic data; dielectric constant, conductivity and scattering coefficient are extracted from the radar data; rock drillability index and rock mass integrity coefficient are calculated from the measurement-while-drilling data. The Kriging interpolation algorithm is used to uniformly resample multi-dimensional feature indicators to a standard three-dimensional mesh.

4. The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site as described in claim 1, characterized in that, The data fusion algorithm adopts an image fusion model based on deep convolutional neural networks to perform pixel-level fusion of seismic wave impedance inversion results, radar dielectric constant distribution and drilling lithology parameters to generate a three-dimensional perspective model.

5. The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site as described in claim 1, characterized in that, Obtaining the structural vulnerability level V of a target historical site includes the following steps: Based on a pre-stored site archive database, multiple evaluation factors for the target site are obtained. The evaluation factors include at least the site type, construction date, average material strength, preservation status rating, historical protection level, structural integrity coefficient, environmental sensitivity, archaeological value level, and existing reinforcement measures. The weights of each evaluation factor were determined using the analytic hierarchy process (AHP). The fuzzy comprehensive evaluation method is used to fuzzify each evaluation factor individually, and a comprehensive evaluation is performed based on the weights to output the structural vulnerability level V.

6. The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site as described in claim 1, characterized in that, The disturbance impact coefficient S of the current construction operation on the site is calculated based on the construction parameters, as shown in the following formula: S= + + ； in, , , All are weighting coefficients; Q is the charge per stage, V is the blasting vibration velocity, D is the average daily advance, and Q0, V0, and D0 are preset benchmark parameters according to specifications and protection requirements.

7. The method for early warning of adverse geological conditions when a tunnel passes under an archaeological site as described in claim 1, characterized in that, In the triggered graded early warning, different levels of early warning instructions have corresponding construction response measures, which include at least one of the following: adjusting the monitoring frequency, limiting blasting parameters, controlling the tunneling advance, and stopping construction.

8. An early warning system for adverse geological conditions when a tunnel passes under an archaeological site, characterized in that: include: The data acquisition module is configured to: deploy a multi-source detection network in a designated area of ​​tunnel construction to acquire raw data reflecting geological anomalies, including at least one of seismic waves, electromagnetic waves, and drilling measurement data. The data processing and feature extraction module is configured to preprocess and extract features from the raw data to obtain a set of multi-dimensional feature indicators, including wave velocity, dielectric constant, and rock mass integrity coefficient. The data fusion imaging module is configured to: generate a three-dimensional stereoscopic perspective model containing the spatial distribution and physical properties of adverse geological bodies based on a multi-dimensional feature index set and a data fusion algorithm, and determine the geological anomaly level G accordingly; The risk assessment and early warning module is configured to: obtain the structural vulnerability level V of the target historical site, and calculate the disturbance impact coefficient S of the current construction operation on the site based on the construction parameters; The risk assessment and early warning module is also configured to: input the geological anomaly level G, the site vulnerability level V and the construction impact coefficient S into the preset risk coupling assessment model, dynamically calculate the comprehensive risk value R, and trigger the corresponding graded early warning according to the threshold range of the R value; The risk coupling assessment model uses a weighted scoring model to calculate the comprehensive risk value. During the calculation of the comprehensive risk value, the weight corresponding to the geological anomaly level G is higher than the weight corresponding to the site vulnerability level V, and the weight corresponding to the site vulnerability level V is equal to the weight corresponding to the construction impact coefficient S.

9. A computer program product, characterized in that, Includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to perform the steps in the adverse geological warning method for tunnels passing under archaeological sites as described in any one of claims 1-7.

10. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, the memory being used to store computer programs; the processor is used to execute the computer programs, enabling the electronic device to perform the steps in the adverse geological warning method for tunnels passing under archaeological sites as described in any one of claims 1-7.

Images