Super-long tunnel geologic body state evaluation method and system fusing geophysics-geomechanics-drilling data

Abstract

The invention provides an ultra-long tunnel geologic body state evaluation method and system fusing geophysics-geomechanics-drilling data, and relates to the technical field of tunnel construction safety. The method comprises the following steps: acquiring geophysical exploration data in front of a tunnel working face, lithologic data disclosed by drilling and measurement while drilling data; preliminary spatial distribution of rock mass mechanical parameters is obtained through inversion of a rock physical model, correction and spatial interpolation are conducted on the preliminary spatial distribution through a machine learning model by means of the correlation between measurement while drilling data and rock mass strength, and a three-dimensional geomechanical model is constructed; constructing a finite element-discrete element coupling numerical model on the basis of a three-dimensional geomechanical model, and analyzing a catastrophe evolution mechanism of surrounding rock by simulating a tunneling unloading process to obtain a catastrophe high-risk area subjected to numerical simulation; and performing spatial superposition and coupling analysis on the abnormal area, the parameter weakening area and the catastrophe high-risk area to obtain a stability grading and disaster risk early warning result of the geologic body in front of the working face.

Description

Technical Field

[0001] This application relates to the field of tunnel construction safety technology, and in particular to a method and system for assessing the geological condition of ultra-long tunnels by integrating geophysical, geomechanical, and drilling data. Background Technology

[0002] Ultra-long tunnels (such as mountain railway tunnels and cross-basin water diversion tunnels) often traverse fault fracture zones, water-rich karst cavities, and other unfavorable geological formations during excavation. The unclear state of these formations is a major cause of construction disasters such as water and mud inrushes and collapses. Achieving accurate and advanced prediction of the geological conditions ahead of the working face is a core challenge for ensuring safe and efficient tunnel construction. Currently, several methods are mainly relied upon, all of which have limitations: geological analysis methods rely on experience, have short prediction distances, and are weak at identifying hidden hazards; geophysical exploration methods, such as seismic wave detection, have long detection distances but low resolution and significant interpretation ambiguity; ground-penetrating radar is sensitive to water bodies but attenuates rapidly in water-rich strata, and neither can directly obtain the mechanical parameters required for support design; drilling methods provide reliable results but are costly and time-consuming, only reflecting local information and unable to construct a comprehensive three-dimensional model of the working face; measurement-while-drilling technology has the potential for continuous exploration, but is currently mostly used for qualitative lithology identification and has not yet established a quantitative relationship with surrounding rock stability and disaster risk.

[0003] Furthermore, the aforementioned methods operate independently, with inconsistent data formats and scales, creating "information silos." Geophysical data reveals macroscopic anomalies but fails to clarify their mechanical properties and failure modes; drilling and measurement-while-drilling data reflect local intensity but are difficult to extrapolate effectively; and none of these methods can predict the failure mechanism and dynamic evolution of geological bodies before a disaster. Therefore, how to achieve accurate and quantitative evaluation of the geological state ahead of the tunnel face and dynamic early warning of disaster risks is an urgent technical problem to be solved. Summary of the Invention

[0004] In view of the above problems, this application provides a method and system for assessing the geological condition of ultra-long tunnels by integrating geophysical, geomechanical and drilling data. It can integrate macroscopic geophysical anomalies, mesoscopic drilling information and microscopic mechanical mechanisms to achieve accurate and quantitative evaluation of the geological condition of the geological body in front of the tunnel working face and dynamic early warning of disaster risks.

[0005] A first aspect of this application discloses a method for assessing the geological condition of ultra-long tunnels by integrating geophysical, geomechanical, and drilling data, the method comprising: Collect geophysical exploration data ahead of the tunnel working face, lithological data revealed by drilling, and measurement-while-drilling data; Based on the geophysical exploration data and the lithological data revealed by drilling, the preliminary spatial distribution of rock mass mechanical parameters is obtained by inversion through a rock physics model. By utilizing the correlation between the measurement-while-drilling data and the rock mass strength, the preliminary spatial distribution is corrected and spatially interpolated through a machine learning model to construct a three-dimensional geomechanical model. Based on the aforementioned three-dimensional geomechanical model, a finite element-discrete element coupled numerical model is constructed. By simulating the unloading process during tunnel excavation, the catastrophic evolution mechanism of the surrounding rock is analyzed, and the high-risk catastrophic zone in the numerical simulation is obtained. By spatially superimposing and coupling analysis of the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas, the stability classification and disaster risk warning results of the geological body in front of the working face are obtained.

[0006] Optionally, the geophysical exploration data includes tunnel seismic wave exploration data and ground-penetrating radar data; collecting geophysical exploration data in front of the tunnel working face includes: By arranging excitation holes and receiving sensor arrays on the rear sidewall of the tunnel excavation face, the tunnel seismic wave exploration data is collected. The tunnel seismic wave exploration data includes the P-wave and S-wave velocity structure, wave impedance profile and reflection interface distribution within the target range in front of the tunnel working face. For key anomaly areas revealed by the tunnel seismic wave exploration data, intensified detection was carried out using ground-penetrating radar to obtain the ground-penetrating radar detection data.

[0007] Optionally, the lithological data revealed by drilling includes core mechanical parameters and digital borehole camera data; the acquisition of lithological data revealed by drilling and measurement-while-drilling data includes: Advanced horizontal drilling is carried out, and the drilling measurement data is recorded throughout the drilling process. The drilling measurement data includes drilling speed, torque, thrust, and vibration parameters. After drilling is completed, images of the borehole wall are obtained using digital borehole cameras to obtain digital borehole camera data that characterizes the rock mass structure; Indoor rock mechanics tests were conducted on the drilled core samples to obtain the core mechanical parameters, which include at least one of uniaxial compressive strength, elastic modulus, Poisson's ratio, and cohesion.

[0008] Optionally, based on the geophysical exploration data and the lithological data revealed by drilling, a preliminary spatial distribution of rock mass mechanical parameters is obtained through rock physics model inversion, including: Using the core mechanical parameters and their corresponding geophysical parameters at the borehole locations, a rock physics model reflecting the relationship between the rock mass mechanical properties and the geophysical parameters is established. Substituting the geophysical exploration data into the rock physics model, the preliminary spatial distribution of the rock mass mechanical parameters is obtained by inversion.

[0009] Optionally, utilizing the correlation between the drilling measurement data and rock mass strength, a machine learning model is used to correct and spatially interpolate the preliminary spatial distribution, constructing a three-dimensional geomechanical model, including: Using the rock mechanics parameters at the borehole locations in the preliminary spatial distribution as label samples and the corresponding drilling measurement data as feature samples, a machine learning model is trained to establish a nonlinear mapping relationship between the drilling measurement parameters and the rock mechanics parameters. Based on the trained machine learning model, the preliminary spatial distribution is corrected, and spatial interpolation is performed on the areas not covered by the advance drilling to generate a three-dimensional data volume of rock mass mechanical parameters. The three-dimensional data volume of rock mass mechanical parameters is meshed and its attributes are assigned to obtain the three-dimensional geomechanical model.

[0010] Optionally, the three-dimensional data volume of rock mass mechanical parameters is meshed and its attributes are assigned to obtain the three-dimensional geomechanical model, including: Based on the tunnel design axis, a three-dimensional model space is defined. The three-dimensional model space covers at least a first distance in front of the tunnel face along the tunnel excavation direction, and covers a range of at least N times the tunnel diameter in both the horizontal and vertical directions. The three-dimensional model space is discretized by meshing, wherein a dense mesh is used in the region close to the tunnel excavation outline and a sparse mesh is used in the region far from the tunnel excavation outline. The three-dimensional data volume of rock mass mechanical parameters is used as the basic parameter field, and the geological interface obtained by interpreting the geophysical exploration data and the rock mass structural surface information identified by digital borehole camera data in the lithology data revealed by drilling are used as constraints for spatial distribution. Based on the fundamental parameter field and the constraints, rock mechanics parameter values ​​are assigned to each grid cell in the three-dimensional model space using a spatial interpolation algorithm. Based on the rock mass integrity coefficient and structural surface development assigned to the grid cells, the material constitutive model type corresponding to each grid cell in the numerical simulation is determined, thus obtaining the three-dimensional geomechanical model.

[0011] Optionally, based on the three-dimensional geomechanical model, a finite element-discrete element coupled numerical model is constructed, including: From the aforementioned three-dimensional geomechanical model, extract the target section model that includes key adverse geological bodies; Based on the rock mass integrity coefficient and structural surface development characteristics of each grid unit in the target section model, the region with a rock mass integrity coefficient greater than the target threshold and no obvious structural surface is classified as a continuous medium region suitable for finite element simulation, and the region with a rock mass integrity coefficient less than the target threshold and joint and fissure development is classified as a discontinuous medium region suitable for discrete element simulation. Based on the division of the continuous medium region and the discontinuous medium region, the finite element-discrete element coupled numerical model is constructed.

[0012] Optionally, by simulating the unloading process during tunnel excavation, the catastrophic evolution mechanism of the surrounding rock is analyzed to obtain the numerically simulated high-risk catastrophic zones, including: In the finite element-discrete element coupled numerical model, the stress path change of the surrounding rock caused by tunnel excavation unloading is dynamically simulated by gradually activating the model elements representing the tunnel face. The damage evolution law of the surrounding rock during the simulation process is analyzed. The damage evolution law characterizes the initiation, propagation and penetration process of microcracks in the rock mass under stress, as well as the formation and development law of macroscopic plastic zone or failure zone. Based on the analysis results of the damage evolution law, potential instability modes and key parts of the surrounding rock are identified, and areas where large-scale damage occurs or significant plastic failure zones are identified as high-risk disaster zones.

[0013] Optionally, the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas are spatially superimposed and coupled to obtain the stability classification and disaster risk warning results of the geological body ahead of the working face, including: In the three-dimensional visualization platform, the anomaly areas revealed by the geophysical exploration data, the rock mass mechanical parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas obtained by the numerical simulation are overlaid and displayed. Based on the superimposed spatial information, a stability grading standard is established; wherein the stability grading standard includes at least three levels: high risk, medium risk and low risk, and the determination of each level is based on the spatial coupling relationship and quantitative indicators among the anomaly area, the parameter weakening area and the high-risk disaster area. By calculating the spatial overlap between the anomaly zone, the parameter weakening zone, and the high-risk disaster zone, and comparing it with the quantitative indicators in the stability grading standard, the stability level of the geological body in front of the working face is determined. Based on the determined stability level, a disaster risk early warning report is generated and output, which includes the location of the disaster risk, the type of risk, the level of risk, and recommendations for prevention and control measures.

[0014] A second aspect of this application discloses a geological condition assessment system for ultra-long tunnels that integrates geophysical, geomechanical, and drilling data. The system includes: The sensing and data acquisition module is used to collect geophysical exploration data ahead of the tunnel working face, lithological data revealed by drilling, and measurement-while-drilling data. The core intelligent processing and modeling module is used to obtain the preliminary spatial distribution of rock mass mechanical parameters through rock physics model inversion based on the geophysical exploration data and the lithological data revealed by drilling. It then uses the correlation between the measurement-while-drilling data and the rock mass strength to correct and spatially interpolate the preliminary spatial distribution through a machine learning model to construct a three-dimensional geomechanical model. Based on the three-dimensional geomechanical model, it constructs a finite element-discrete element coupled numerical model and analyzes the catastrophic evolution mechanism of the surrounding rock by simulating the tunnel excavation unloading process to obtain the numerically simulated high-risk catastrophic zone. The application and decision support module is used to perform spatial overlay and coupling analysis on the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas to obtain the stability classification and disaster risk warning results of the geological body in front of the working face.

[0015] A third aspect of this application discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method for assessing the state of ultra-long tunnel geological bodies by integrating geophysical-geomechanical-drilling data as described in the first aspect of this application.

[0016] The embodiments of this application have the following advantages: In this embodiment, by collecting and correlating three types of heterogeneous data—geophysical exploration data, drill-revealed lithological data, and drilling-while-operated measurement data—a unified three-dimensional geomechanical model is constructed by integrating macroscopic geophysical anomalies, mesoscopic borehole-revealed information, and microscopic drilling-while-operated response phase structures. This breaks down information silos between related methods and improves the overall integrity and completeness of advanced geological prediction. Furthermore, a preliminary inversion is performed using a rock physics model, and nonlinear correction and spatial interpolation are performed using drilling-while-operated measurement data closely related to rock mass strength through a machine learning model. This process overcomes the inherent multiple-solution problem of geophysical inversion, generating a high-resolution, high-precision three-dimensional geomechanical model whose results can be directly used for support design and engineering analysis.

[0017] Based on the constructed three-dimensional geomechanical model, a coupled finite element-discrete element numerical model was built to simulate the unloading process during tunnel excavation. This model can quantitatively reproduce the entire catastrophic process of the surrounding rock, clarify the location of the disaster, and reveal the evolution mechanism of the disaster, providing a scientific basis for proactive disaster prevention and engineering decision-making. Finally, by spatially superimposing and coupling geophysical anomalies, weakened mechanical parameter zones, and high-risk disaster zones simulated in numerical simulations, a comprehensive assessment and early warning of the stability of the geological body ahead was achieved, from qualitative to quantitative and from static to dynamic, providing precise scientific decision support for the safe and intelligent construction of ultra-long tunnels. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the steps of a method for assessing the geological condition of an ultra-long tunnel that integrates geophysical, geomechanical, and drilling data, as provided in an embodiment of this application. Figure 2 This is a simulation effect diagram of the three-dimensional distribution of the strength of the geological body before tunnel excavation, based on the inversion of multi-source heterogeneous data fusion, provided in an embodiment of this application. Figure 3 This is a schematic diagram of a geological body condition assessment system for ultra-long tunnels that integrates geophysical, geomechanical, and drilling data, provided in an embodiment of this application. Figure 4 This is a detailed flowchart of the operation of an ultra-long tunnel geological body condition assessment system that integrates geophysical, geomechanical, and drilling data, provided in an embodiment of this application. Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] Reference Figure 1 As shown, Figure 1This is a flowchart illustrating the steps of a method for assessing the geological condition of ultra-long tunnels that integrates geophysical, geomechanical, and drilling data, as provided in an embodiment of this application. Figure 1 As shown, the method may include steps S110 to S140: Step S110: Collect geophysical exploration data, lithological data revealed by drilling, and measurement-while-drilling data in front of the tunnel working face.

[0022] By deploying multiple sensors and devices, three types of data can be collected simultaneously: geophysical exploration data, lithological data revealed by drilling, and measurement-while-drilling data, to comprehensively understand the geological conditions ahead of the tunnel face. Geophysical exploration data refers to data obtained through tunnel seismic wave exploration (TSP) and ground-penetrating radar (GPR). TSP utilizes the propagation characteristics of seismic waves in rock mass to detect wave velocity structures and discontinuities (such as faults) over a relatively large area (e.g., 150-200 meters). Ground-penetrating radar uses high-frequency electromagnetic waves to conduct detailed detection of key anomaly areas revealed by TSP, identifying small structures, water-bearing areas, etc.

[0023] The lithological data revealed by drilling can be obtained directly through advanced horizontal drilling, and may include core mechanical parameters (e.g., parameters such as uniaxial compressive strength and elastic modulus of rock obtained through laboratory tests) and digital borehole camera data (e.g., high-definition images obtained through borehole cameras for accurate identification of lithology, fractures, dissolution and other structural features).

[0024] Measurement while drilling (MWD) data are drilling parameters recorded in real time by sensors integrated on the drilling rig during advanced drilling operations. MWD data includes, but is not limited to, parameters such as drilling speed, torque, thrust, and vibration. These parameters reflect the dynamic response of the drill bit's interaction with the rock mass during drilling and are correlated with the rock mass's mechanical properties, such as strength and integrity.

[0025] In some embodiments, after collecting geophysical exploration data, drilling-revealed lithological data, and drilling-while-drilling measurement data, the data from different sources, formats, and spatiotemporal references can be uniformly registered and standardized to eliminate scale and coordinate system differences, and construct a unified data model that can comprehensively reflect the physical properties, lithology, and mechanical potential of the geological body ahead of the tunnel, laying the foundation for subsequent quantitative analysis.

[0026] Step S120: Based on the geophysical exploration data and the lithological data revealed by drilling, the preliminary spatial distribution of rock mass mechanical parameters is obtained by inversion through a rock physics model; using the correlation between the drilling measurement data and the rock mass strength, the preliminary spatial distribution is corrected and spatially interpolated through a machine learning model to construct a three-dimensional geomechanical model.

[0027] Among them, the rock physics model is used to describe the mechanical properties of rock masses, such as the empirical or theoretical relationship between strength and geophysical parameters (wave velocity, resistivity). First, based on the rock physics model, the preliminary spatial distribution of rock mass mechanical parameters (such as strength and integrity coefficient) is derived using geophysical survey data. This preliminary spatial distribution is affected by the ambiguity of geophysical methods and has limited accuracy.

[0028] Subsequently, to address the multiple solutions issue in geophysical inversion and improve model resolution, drilling-while-measuring data correlated with rock mass strength was introduced as a real-time verification and correction source. A nonlinear mapping relationship between drilling-while-measuring parameters and rock mass mechanical parameters (i.e., inverted mechanical parameters) was established using machine learning models (such as gradient boosting decision trees and random forests, among other supervised learning algorithms). This machine learning model was then used to correct the initial spatial distribution of rock mass mechanical parameters and to perform high-confidence spatial interpolation on areas not covered by the borehole, ultimately constructing a high-resolution, high-accuracy three-dimensional geomechanical model. This three-dimensional geomechanical model not only includes the detailed three-dimensional distribution of parameters such as rock mass strength and elastic modulus but also distinguishes between continuous and discontinuous media regions through structural surface information.

[0029] Thus, this step enables a quantitative mapping from geophysical response to rock mechanics parameters, and improves the accuracy and resolution of the three-dimensional geomechanical model.

[0030] Step S130: Based on the three-dimensional geomechanical model, a finite element-discrete element coupled numerical model is constructed, and the catastrophic evolution mechanism of the surrounding rock is analyzed by simulating the tunnel excavation unloading process, so as to obtain the catastrophic high-risk area in numerical simulation.

[0031] This step moves from static modeling to dynamic mechanism analysis. Based on the three-dimensional geomechanical model constructed in step S120, a finite element-discrete element coupled numerical model is built for key adverse geological sections (such as low-intensity zones and fracture zones). This finite element-discrete element coupled numerical model can simultaneously simulate the stress and strain of continuous rock masses (finite element method) and the opening and slippage of discontinuities such as joints and fractures (discrete element method).

[0032] By dynamically simulating the unloading process during tunnel excavation (such as the unloading process during the gradual advancement of the tunnel face) in this finite element-discrete element coupled numerical model, the entire catastrophic evolution process, from stress redistribution in the surrounding rock, the initiation and propagation of microcracks, to the penetration of macroscopic plastic zones or block instability, can be quantitatively reproduced. Through simulation, the catastrophic evolution mechanism of surrounding rock instability under different geological conditions (such as shear slip, tensile failure, and block collapse) is revealed, and high-risk catastrophic zones exhibiting instability characteristics such as large-scale plastic zone penetration and abrupt displacement are identified in the simulation.

[0033] Step S140: Spatial superposition and coupling analysis is performed on the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas to obtain the stability classification and disaster risk warning results of the geological body in front of the working face.

[0034] Among them, the anomaly zones revealed by geophysical exploration data (such as strong reflection interfaces and water-bearing anomaly zones) reflect the macroscopic spatial location of significant physical property differences at interfaces or unfavorable geological bodies (such as faults, fracture zones, and water-rich areas) within the strata. The parameter weakening zones in the three-dimensional geomechanical model characterize the spatial range where key mechanical parameters such as rock mass strength and elastic modulus are significantly lower than design or background values, quantitatively revealing the attenuation of rock mass bearing capacity and the zoning of its quality. The high-risk disaster zones obtained from numerical simulations are key locations predicted by mechanical mechanism analysis, where microscopic damage accumulation, crack penetration, and ultimately macroscopic instability (such as collapses and large deformations) may occur under the unloading action of tunnel excavation.

[0035] In the 3D visualization platform, the three types of regions (anomaly zone, parameter weakening zone, and high-risk disaster zone) are spatially overlaid and coupled for analysis. By calculating the spatial overlap of these regions (e.g., area overlap ratio, spatial inclusion relationship, etc.) and combining the strength thresholds of various indicators (e.g., geophysical reflection amplitude, percentage of rock mass strength below design value, and plastic zone range in numerical simulation), a set of quantitative and qualitative stability grading standards is established. For example, the stability of the geological body ahead is divided into three levels: high risk (Level A), medium risk (Level B), and low risk (Level C). Based on this stability grading standard, the risk level is determined, and disaster risk warning results (disaster risk warning reports) are generated, directly serving construction decisions. The disaster risk warning results can clearly indicate the spatial location, scale, and inferred nature of the adverse geological body in the form of figures, text, and tables; at the same time, specific disaster risk types (e.g., water inrush, landslide), risk levels, and targeted prevention and control measures suggestions from the associated engineering experience database (e.g., adjusting tunneling parameters, strengthening support, etc.) are also provided.

[0036] In this way, by deeply integrating and spatially coupling the results of macroscopic detection, mesoscopic modeling, and microscopic simulation, a multi-dimensional, visualized, and operable dynamic early warning of the stability of the geological body in front of the tunnel working face is achieved. This successfully connects "information silos" from different sources and scales into a complete risk cognition chain, elevating the forecast conclusion from "possible anomalies" to a decision support level of "where, what nature, how much risk, and how to prevent and control it," greatly enhancing the accuracy, foresight, and intelligence of risk pre-control in ultra-long tunnel construction.

[0037] The technical solution of this application, by collecting and associating three types of heterogeneous data—geophysical exploration data, drilling-revealed lithological data, and drilling-while-operated measurement data—constructs a unified three-dimensional geomechanical model by integrating macroscopic geophysical anomalies, mesoscopic borehole-revealed information, and microscopic drilling-while-operated response phase structures. This breaks down information silos between related methods and improves the overall integrity and completeness of advanced geological prediction. Furthermore, a preliminary inversion is performed using a rock physics model, and nonlinear correction and spatial interpolation are performed using drilling-while-operated measurement data closely related to rock mass strength through a machine learning model. This process overcomes the inherent multi-solution problem of geophysical inversion, generating a high-resolution, high-precision three-dimensional geomechanical model whose results can be directly used for support design and engineering analysis. Based on the constructed three-dimensional geomechanical model, a finite element-discrete element coupled numerical model is built to simulate the tunnel excavation unloading process. This model can quantitatively reproduce the entire catastrophic process of the surrounding rock, clarify the location of the disaster, and reveal the evolution mechanism of the disaster, providing a scientific basis for proactive disaster prevention and engineering decision-making. Ultimately, by spatially superimposing and coupling geophysical anomalies, weakened mechanical parameters, and high-risk areas of numerical simulation disasters, a comprehensive assessment and early warning of the stability of the geological body ahead was achieved, from qualitative to quantitative and from static to dynamic, providing precise scientific decision support for the safe and intelligent construction of ultra-long tunnels.

[0038] In an optional embodiment, the geophysical exploration data includes tunnel seismic wave exploration data and ground-penetrating radar detection data; the step S110 above, "collecting geophysical exploration data in front of the tunnel working face," specifically includes sub-steps S110-1 to S110-2: Step S110-1: By arranging excitation holes and receiving sensor arrays on the rear sidewall of the tunnel excavation face, the tunnel seismic wave exploration data is collected. The tunnel seismic wave exploration data includes the P-wave and S-wave velocity structure, wave impedance profile and reflection interface distribution within the target range in front of the tunnel working face.

[0039] This step can perform a large-scale anomaly scan and interface identification of the geological conditions in front of the working face. The tunnel seismic wave exploration data can reflect the rock mass wave velocity structure and macroscopic structural interface information within the target range (usually 150-200 meters) in front of the tunnel working face.

[0040] Specifically, a series of excitation holes and receiving sensor arrays are arranged in a predetermined geometric pattern on the sidewall behind the current tunnel excavation face. The excitation holes are equipped with a seismic source (such as explosives or a controlled seismic source) to generate seismic waves; the receiving holes are equipped with high-sensitivity accelerometers (usually three-component, capable of recording vibrations in multiple directions) to receive seismic wave signals reflected or refracted from the rock mass ahead.

[0041] By exciting the seismic source and simultaneously recording the time series of seismic wavefields received by each sensor, the raw seismic record is obtained. The acquired raw seismic record is then processed (e.g., filtering, velocity analysis, migration imaging), ultimately interpreting the P-wave and S-wave velocity structure, wave impedance profile, and reflection interface distribution within the target area (e.g., 100 meters) in front of the tunnel working face. P-waves (longitudinal waves) and S-waves (transverse waves) are two basic types of seismic waves propagating in rock masses. Their velocities and ratios are key parameters for inverting rock mass mechanical properties and determining water content. Wave impedance is the product of rock density and wave velocity; abrupt transitions at these interfaces constitute reflection interfaces (e.g., faults, lithological interfaces).

[0042] Step S110-2: For key anomaly areas revealed by the tunnel seismic wave exploration data, use ground-penetrating radar for intensified detection to obtain the ground-penetrating radar detection data.

[0043] This step can perform intensive detection on the key anomaly areas revealed in step S110-1. Intensive detection refers to reducing the spacing between measurement points and conducting more intensive data collection to improve the ability to distinguish small structures and water-bearing features.

[0044] Specifically, ground-penetrating radar (GPR) transmits high-frequency electromagnetic pulses into the rock mass via an antenna and receives reflected waves from interfaces with different electrical properties. By analyzing radar images (radar profiles), based on the intensity of reflected waves, the continuity of the phase axis, and waveform characteristics (such as waveform disorder and strong reflection), it can further identify fine geological anomalies such as small structures (such as small faults and dense joint zones) and water-bearing areas. Its effective detection depth is typically shallow, approximately 30-50 meters.

[0045] By adopting the technical solution of this application embodiment, a complementary geophysical data acquisition system is constructed through a collaborative detection strategy of first performing macroscopic scanning of tunnel seismic wave exploration and then performing fine identification by ground-penetrating radar. This system enables comprehensive perception of multi-level and multi-parameter geological conditions in front of the tunnel working face, providing a richer and more reliable data foundation for subsequent data fusion and accurate inversion.

[0046] In an optional embodiment, the lithological data revealed by drilling includes core mechanical parameters and digital borehole camera data; the step S110 above, "collecting lithological data and measurement-while-drilling data revealed by drilling," specifically includes sub-steps S110-3 to S110-5: Step S110-3: Conduct advanced horizontal drilling and record the drilling-while-drilling measurement data throughout the drilling process. The drilling-while-drilling measurement data includes drilling speed, torque, thrust, and vibration parameters.

[0047] Specifically, a special drilling rig is used to perform horizontal drilling operations at the tunnel face to the interior of the rock mass ahead. During the drilling process, a series of drilling process parameters are recorded in real time and continuously by sensors integrated on the drill rod or drilling rig (forming a measurement-while-drilling system).

[0048] Among them, drilling speed is the drilling depth of the drill bit per unit time, which can reflect the drillability (hardness) of the rock mass; torque and thrust are the rotational torque and axial pressure required to drive the drill bit to break the rock, which are closely related to the rock mass strength and integrity; vibration parameters are the vibration signals generated during drilling, and their spectrum and amplitude characteristics can reflect the heterogeneity and structural surface development of the rock mass.

[0049] This step can obtain continuous, real-time measurement-while-drilling data, which contains information on the continuous changes in the mechanical properties of the rock mass along the borehole trajectory, providing valuable feature samples that are strictly corresponding to the spatial location for subsequent machine learning correction.

[0050] Step S110-4: After drilling is completed, use digital borehole cameras to acquire borehole wall images to obtain the digital borehole camera data characterizing the rock mass structure.

[0051] Specifically, after drilling is completed, a dedicated digital camera probe is inserted into the borehole. During the lifting or lowering process, the borehole wall is continuously scanned 360 degrees to obtain digital borehole camera data. By interpreting this digital borehole camera data, rock mass structural features such as lithological change interfaces, the attitude (strike, dip angle) of fractures (joints), opening degree, filling materials, and dissolution cavities can be accurately identified and statistically analyzed.

[0052] This step compensates for the loss of structural information due to insufficient core recovery or core fragmentation, providing intuitive and quantitative data on the development of structural planes in the rock mass surrounding the borehole. This information is a key constraint for subsequent construction of numerical models (especially the discrete element part) that reflect the true structural characteristics of the rock mass.

[0053] Step S110-5: Conduct indoor rock mechanics tests on the drilled core samples to obtain the core mechanical parameters, which include at least one of uniaxial compressive strength, elastic modulus, Poisson's ratio, and cohesion.

[0054] Among them, rock core mechanical parameters are basic parameters that characterize the mechanical properties of rock materials themselves and are directly measured by experiments; uniaxial compressive strength is the ultimate stress when a rock sample fails under uniaxial pressure; elastic modulus and Poisson's ratio are parameters that describe the stress-strain relationship of rock in the elastic deformation stage; cohesion is the bonding strength between particles inside the rock and is an important component of the shear strength of the rock mass.

[0055] The core mechanical parameters obtained in this step are the baseline data for establishing the rock physics model, and also the basic input for calibrating and verifying the rock mass constitutive model in numerical simulation, ensuring the reliability and accuracy of the mechanical parameters of the entire evaluation chain.

[0056] By adopting the technical solution of this application embodiment, three types of key data with different properties are acquired simultaneously during advanced drilling operations: continuous dynamic measurement while drilling data, high-resolution digital borehole camera data, and core mechanical parameters. This lays a solid and diverse data foundation for subsequent steps such as correcting geophysical inversion, constructing a high-precision three-dimensional geomechanical model, and setting up a numerical model that conforms to reality. It is a prerequisite for achieving deep fusion of multi-source data.

[0057] In an optional embodiment, step S120 above, "based on the geophysical exploration data and the lithological data revealed by drilling, obtaining the preliminary spatial distribution of rock mass mechanical parameters through rock physics model inversion," specifically includes sub-steps S120-1 to S120-2: Step S120-1: Using the core mechanical parameters and their corresponding geophysical parameters at the borehole location, establish a rock physics model that reflects the correspondence between the rock mass mechanical properties and the geophysical parameters.

[0058] Among them, the core mechanical parameters are obtained through indoor tests at a specific depth in the borehole and represent the mechanical properties of the rock material at that point; the corresponding geophysical parameters refer to the corresponding parameters extracted from geophysical exploration data at the same depth in the borehole (for example, wave velocity parameters extracted from tunnel seismic wave exploration data, or resistivity parameters inferred from ground-penetrating radar data).

[0059] Based on the aforementioned paired core mechanical and geophysical parameters, a rock physics model reflecting the correspondence between rock mass mechanical properties and geophysical parameters can be established through mathematical statistical analysis or theoretical modeling and fitting. For example, rock physics models (or empirical formulas) such as "wave velocity-rock mass strength" and "wave velocity-integrity coefficient" can be established.

[0060] Understandably, rock physics models are specific to the work area and can fit measured data under specific geological conditions in that area. This provides a reliable conversion relationship for deriving mechanical parameters from geophysical data, forming the basis for inversion. Different rock physics models, such as wave velocity-intensity, wave velocity-integrity coefficient, and resistivity-porosity models, can be selected and established based on the geological characteristics of the work area.

[0061] Step S120-2: Substitute the geophysical exploration data into the rock physics model to obtain the preliminary spatial distribution of the rock mass mechanical parameters.

[0062] This step utilizes the established rock physics model to convert the large-scale, continuous geophysical exploration data obtained in step S110 into a preliminary spatial distribution of rock mass mechanical parameters, thereby achieving a quantitative estimation from physical properties to mechanical properties.

[0063] Specifically, the geophysical exploration data is substituted into the rock physics model, and based on the transformation relationships defined in the model, the geophysical parameter values ​​for each spatial location (3D grid cell or pixel) are calculated, thereby directly retrieving the corresponding rock mass mechanical parameter values ​​(such as strength, elastic modulus, etc.). This process generates a preliminary 3D data volume of rock mass mechanical parameters with the same spatial range as the geophysical data volume, i.e., the preliminary spatial distribution of rock mass mechanical parameters.

[0064] The technical solution of this application embodiment constructs a rock physics model that connects geophysical exploration data and rock mechanics parameters. It uses precise mechanical information at the borehole point to calibrate and interpret the continuous response on the geophysical surface, thereby overcoming the limitation that traditional geophysical anomalies can only be qualitatively interpreted and cannot be quantitatively converted into mechanical parameters. This provides an important initial field and spatial framework for subsequent machine learning correction and spatial interpolation using drilling measurement data.

[0065] In an optional embodiment, step S120 above, "using the correlation between the drilling measurement data and the rock mass strength, correcting and spatially interpolating the preliminary spatial distribution through a machine learning model to construct a three-dimensional geomechanical model," may include sub-steps S120-3 to S120-5: Step S120-3: Using the rock mechanics parameters at the borehole locations in the preliminary spatial distribution as label samples and the corresponding drilling measurement data as feature samples, train a machine learning model to establish a nonlinear mapping relationship between the drilling measurement parameters and the rock mechanics parameters.

[0066] First, from the preliminary spatial distribution obtained in step S120-1, rock mass mechanical parameter values ​​(e.g., uniaxial compressive strength, integrity coefficient, etc.) are extracted at the borehole locations. These parameter values ​​are derived from geophysical data inversion and serve as the model's prediction targets, i.e., label samples. Simultaneously, from the drilling measurement-while-drilling data collected in step S110, drilling measurement parameters recorded at the same borehole locations (e.g., drilling speed, rotational speed, torque, thrust, vibration parameters, etc.) are extracted. These parameters reflect the dynamic response during drilling and constitute the model's input features, i.e., feature samples.

[0067] Machine learning models can be algorithms capable of handling complex nonlinear relationships, such as random forests, gradient boosting decision trees, and support vector regression. Subsequently, the selected machine learning model is trained under supervision using a large number of paired labeled samples (mechanical parameters) and feature samples (drilling parameters), enabling the model to learn and establish a nonlinear mapping relationship between drilling measurement parameters (features) and rock mechanics parameters (labels).

[0068] Thus, by linking the measurement-while-drilling data with the preliminary estimates of rock mechanics parameters based on geophysical inversion through machine learning models, a foundation is laid for subsequently using denser measurement-while-drilling data to correct and predict the mechanical parameters of the entire space.

[0069] Step S120-4: Based on the trained machine learning model, the preliminary spatial distribution is corrected, and spatial interpolation is performed on the areas not covered by the advance drilling to generate a three-dimensional data volume of rock mass mechanical parameters.

[0070] For areas already covered by boreholes, the measurement-while-drilling data at that location can be input into a trained machine learning model to obtain a new rock mechanics parameter value based on the model's prediction. This predicted rock mechanics parameter value is then compared, fused, or directly replaced with the preliminary inversion value from step S120-2 to achieve local correction of the preliminary spatial distribution.

[0071] For areas not covered by advanced drilling, since there is no direct measurement-while-drilling (MWD) data, the preliminary inversion mechanical parameter values ​​obtained at that location, along with the MWD parameters estimated from neighboring boreholes based on spatial correlation (such as through Kriging interpolation), can be input into a trained machine learning model. Based on the learned global patterns, the model can predict the rock mass mechanical parameter values ​​for the uncovered area, thereby achieving high-resolution spatial interpolation for the entire region.

[0072] Thus, this step utilizes the high resolution, continuity, and strong correlation with rock mass strength of the drilling measurement data to verify and locally optimize the preliminary inversion results (preliminary spatial distribution) that depend on geophysics. This improves the confidence and spatial resolution of the mechanical parameter predictions for the un-drilled area, and ultimately merges them to generate a spatially continuous three-dimensional data volume of rock mass mechanical parameters that has been constrained and corrected by the drilling test data.

[0073] In some embodiments, after generating the three-dimensional data volume of rock mass mechanics parameters, a leave-one-out cross-validation method can be used to verify the confidence level of the three-dimensional data volume. Specifically, at the borehole location, the measurement-while-drilling data and core mechanical parameters at a certain depth are used as validation points. The model is reconstructed using all other data, and the core mechanical parameters at that point are predicted. The mean absolute error (MAE) and root mean square error (RMSE) between the predicted and measured values ​​are calculated. For example, if the mean absolute error of the predicted rock mass strength is reduced by more than 20% after correction using the measurement-while-drilling data, then the three-dimensional data volume of rock mass mechanics parameters is a high-confidence model.

[0074] In some embodiments, during the machine learning correction process, the confidence interval or standard deviation of the predicted mechanical parameters of each grid cell can also be output simultaneously and visualized in the three-dimensional model in the form of a semi-transparent color gamut or isosurface, so as to intuitively show the credibility of the three-dimensional data volume of rock mass mechanical parameters at different spatial locations.

[0075] Step S120-5: Mesh and assign attribute values ​​to the three-dimensional data volume of rock mass mechanical parameters to obtain the three-dimensional geomechanical model.

[0076] Among them, the three-dimensional geomechanical model is a geomechanical model with spatial coordinates, geometric shape and detailed mechanical property distribution. It is the accurate input and core foundation for subsequent numerical simulation of the disaster process (step S130) and coupling analysis (step S140).

[0077] Specifically, based on the requirements of simulation accuracy and computational efficiency, the three-dimensional spatial region in front of the tunnel to be analyzed (e.g., at least 100 meters in front of the tunnel face) can be divided into a series of small units (e.g., hexahedral or tetrahedral units) to form a computational grid. Subsequently, the parameter values ​​in the three-dimensional data volume of rock mass mechanics parameters generated in step S120-4 are mapped to the nodes or center of each grid unit through spatial interpolation algorithms (e.g., inverse distance weighted interpolation, Kriging interpolation, etc.). Each grid unit is then assigned a complete set of mechanical properties (e.g., density, elastic modulus, Poisson's ratio, compressive strength, tensile strength, cohesion, internal friction angle, etc.). These properties together constitute a three-dimensional digital model that can quantitatively describe the mechanical state of the rock mass ahead, i.e., a three-dimensional geomechanical model.

[0078] The technical solution adopted in this application introduces machine learning technology to deeply integrate drilling measurement data with preliminary geophysical inversion results. This enables intelligent correction of the preliminary model and high-confidence spatial interpolation of unexposed areas, effectively overcoming the ambiguity of geophysical inversion. The final constructed three-dimensional geomechanical model exhibits significantly improved accuracy and reliability compared to models based solely on geophysics, providing a high-quality, quantitative core carrier for the entire evaluation method.

[0079] In some embodiments, the step S120-5 above, "meshing and assigning attribute values ​​to the three-dimensional data volume of rock mass mechanical parameters to obtain the three-dimensional geomechanical model," may include steps S120-5-1 to S120-5-5: Step S120-5-1: Based on the tunnel design axis, define a three-dimensional model space. The three-dimensional model space covers at least a first distance in front of the tunnel face along the tunnel excavation direction, and covers a range of at least N times the tunnel diameter in both the horizontal and vertical directions.

[0080] The three-dimensional model space is a three-dimensional region encompassing all simulated objects (tunnel and surrounding rock mass). Its extent is determined based on the tunnel design axis, specifically: along the tunnel excavation direction (X-axis), it covers at least a first distance (e.g., 100 meters) in front of the tunnel face; and in the transverse (Y-axis) and vertical (Z-axis) directions perpendicular to the axis, its extent is at least N times (e.g., 3 to 5 times) the tunnel diameter. This extent setting ensures that the model fully includes the excavation influence zone while eliminating the interference of artificial boundaries on the simulation results (i.e., eliminating boundary effects).

[0081] Step S120-5-2: Discretize the three-dimensional model space into a mesh, wherein a dense mesh is used in the area close to the tunnel excavation outline, and a sparse mesh is used in the area far from the tunnel excavation outline.

[0082] Structured hexahedral meshes or unstructured tetrahedral meshes can be used to discretize the 3D model space, resulting in a meshed 3D model space. Furthermore, the mesh size can be dynamically adjusted according to accuracy requirements. In areas near the tunnel excavation outline where stress is concentrated and deformation is severe, smaller, denser meshes (e.g., 0.5 to 1.0 meters) are used to improve computational accuracy; while in far-field regions far from the tunnel, larger, sparser meshes (e.g., 2 to 5 meters) are used to save computational resources.

[0083] In this way, while ensuring the simulation accuracy of key areas, the computational scale and efficiency are effectively balanced, and the optimization of computational accuracy and efficiency is achieved.

[0084] Step S120-5-3: Use the three-dimensional data volume of the rock mass mechanical parameters as the basic parameter field, and use the geological interface obtained by interpreting the geophysical exploration data and the rock mass structural surface information identified by digital borehole camera data in the lithology data revealed by drilling as the spatial distribution constraints.

[0085] Among them, the geological interfaces interpreted from geophysical exploration data include: the main reflection interfaces interpreted from tunnel seismic wave exploration data, and the anomalous boundaries identified from ground-penetrating radar data.

[0086] Specifically, the three-dimensional data volume of rock mass mechanical parameters generated in step S120-4 provides a preliminary spatial distribution trend of parameters such as rock mass strength and elastic modulus, which serves as the basic parameter field. Geological interfaces interpreted from geophysical exploration data and rock mass structural surface information (such as the attitude and spacing of dominant joint sets) identified from digital borehole camera data are used as constraints on the spatial distribution to correct and guide the aforementioned preliminary distribution trend, making it more consistent with the key information of the actual geological structure. This provides a dual guarantee for the next step of attribute interpolation.

[0087] Step S120-5-4: Based on the basic parameter field and the constraint conditions, the rock mass mechanics parameter values ​​are assigned to each grid cell of the three-dimensional model space through a spatial interpolation algorithm.

[0088] Spatial interpolation algorithms refer to mathematical methods that estimate the values ​​of unknown locations based on the values ​​of known spatial points, such as Kriging interpolation or inverse distance weighted interpolation.

[0089] Using a selected spatial interpolation algorithm, with the fundamental parameter field as the global trend, the interpolation results are forced to exhibit the appropriate parameter abrupt changes or anisotropy characteristics at the geological interfaces and structural planes identified by the constraints, thereby assigning rock mechanics parameter values ​​to each grid cell. Ultimately, a smooth, continuous, and accurate 3D mesh model is generated that can characterize key geological structures (such as faults, fracture zones, lithological interfaces, and other geological boundaries).

[0090] Step S120-5-5: Based on the rock mass integrity coefficient and structural surface development assigned to the grid cells, determine the material constitutive model type corresponding to each grid cell in the numerical simulation to obtain the three-dimensional geomechanical model.

[0091] Among them, the constitutive model type of a material refers to the mathematical theoretical model used in numerical simulation to describe the stress-strain relationship of a material. For rock masses, it is often divided into continuous medium models (such as elastoplastic models, which are commonly used for finite element simulation) applicable to intact and continuous rock masses, and discontinuous medium models (such as block models, which are commonly used for discrete element simulation) applicable to fractured and jointed rock masses.

[0092] Specifically, based on the rock mass integrity coefficient and structural surface development assigned to each mesh element, the corresponding material constitutive model type in the numerical simulation is determined for each mesh element. This includes: mesh elements with a high rock mass integrity coefficient (greater than the integrity coefficient threshold) and no obvious structural surfaces are marked as continuous media suitable for finite element simulation. Conversely, regions with a low rock mass integrity coefficient (not greater than the integrity coefficient threshold) and well-developed joints and fractures are marked as discontinuous media suitable for discrete element simulation or requiring the embedding of specific structural surfaces. In subsequent numerical simulations, these can be transformed into discrete element blocks or embedded with fracture networks.

[0093] Thus, this step serves as a bridge connecting the static geological model and the dynamic disaster simulation, making the final constructed three-dimensional geomechanical model not only an attribute database, but also a preprocessed model with clearly defined mechanical behavior that directly supports subsequent finite element-discrete element coupled numerical simulations, thereby improving the relevance and efficiency of the numerical simulation.

[0094] The technical solution adopted in this application realizes intelligent allocation of computing resources through variable-density mesh partitioning. By introducing multi-source geological constraints for attribute interpolation, the geological realism and boundary characterization accuracy of the model are significantly improved. Finally, by predefining a constitutive model based on rock mass quality, a seamless transition is prepared for subsequent multi-scale disaster simulation. As the core data carrier of the entire evaluation method, the high-quality construction of this model is a fundamental prerequisite for ensuring the accuracy and reliability of subsequent analysis and early warning results.

[0095] In an optional embodiment, step S130 above, "constructing a finite element-discrete element coupled numerical model based on the three-dimensional geomechanical model," may include sub-steps S130-1 to S130-3: Step S130-1: Extract the target section model, which includes key adverse geological bodies, from the three-dimensional geomechanical model.

[0096] Among them, key adverse geological bodies refer to geological units identified in the three-dimensional geomechanical model that pose a significant potential threat to tunnel safety, such as delineated low-intensity zones, fracture zones, water-rich areas, or areas with strong geophysical anomalies. From the complete three-dimensional geomechanical model with a large coverage area, a local three-dimensional region (i.e., the target section model) containing the tunnel design outline and the key adverse geological bodies is selected; the length of the target section model along the tunnel axis must be sufficient to include the adverse geological bodies and cover their potential influence range.

[0097] Step S130-2: Based on the rock mass integrity coefficient and structural surface development characteristics of each grid unit in the target section model, the region with a rock mass integrity coefficient greater than the target threshold and no obvious structural surface is divided into a continuous medium region suitable for finite element simulation, and the region with a rock mass integrity coefficient less than the target threshold and joint and fissure development is divided into a discontinuous medium region suitable for discrete element simulation.

[0098] Among them, the rock mass integrity coefficient characterizes the degree of rock mass fragmentation, and the structural surface development characteristics are used to describe the density and number of fractures and joints. The target threshold is a pre-set discriminant value used to distinguish the dominant modes of rock mass mechanical behavior. For example, a rock mass integrity coefficient of 0.55 or 0.75 can be used as the target threshold.

[0099] Specifically, by traversing each grid cell of the target section model, automatic classification is performed based on its attributes: if the rock mass integrity coefficient of a grid cell is greater than the target threshold and there are no obvious penetrating structural surfaces, it is classified as a continuous medium region; the rock mass in this type of region is relatively intact, and its deformation and failure are controlled by the mechanical properties of the rock material itself, making it suitable for simulation using the finite element method based on continuous medium mechanics. If the rock mass integrity coefficient of a cell without obvious structural surfaces is less than the target threshold and joints and fissures are well-developed, it is classified as a discontinuous medium region; the mechanical behavior of the rock mass in this type of region is mainly controlled by structural surfaces, and sliding and rotation between blocks may occur, making it suitable for simulation using the discrete element method based on discontinuous medium mechanics.

[0100] Step S130-3: Based on the division of the continuous medium region and the discontinuous medium region, construct the finite element-discrete element coupled numerical model.

[0101] In a numerical simulation software platform, the model can be initialized separately for the divided continuous medium region and the discontinuous medium region to construct a finite element-discrete element coupled numerical model. Specifically, for the continuous medium region, a finite element mesh is generated and assigned the corresponding constitutive model and parameters; for the discontinuous medium region, it is discretized into a series of mutually contactable and separable blocks or particles (discrete element model), and the contact mechanical parameters between the blocks are defined.

[0102] By setting up special coupling interfaces or transition elements, a mechanical interaction mechanism (such as the transmission of force and displacement) between the finite element region and the discrete element region is established, ensuring that the two physically connected regions can respond collaboratively in numerical calculations. In this way, the constructed finite element-discrete element coupled numerical model can seamlessly simulate the complete chain-like catastrophe process from stress concentration and plastic yielding of intact rock mass to crack propagation and block slip instability in fractured rock mass in a single simulation, achieving integrated numerical reproduction of the multi-scale failure mechanism of tunnel surrounding rock.

[0103] The technical solution of this application embodiment constructs a numerical model for high-fidelity simulation of tunnel disaster processes. First, target interception improves computational feasibility. Then, region division is performed based on the properties of the high-precision geomechanical model to ensure the adaptability of the simulation method to the real structure of the rock mass. Finally, by constructing a coupled model, the inherent limitations of single continuous or discontinuous methods in simulating complex rock masses are overcome, providing key mechanical process basis for accurate early warning.

[0104] In an optional embodiment, step S130 above, "analyzing the catastrophic evolution mechanism of the surrounding rock by simulating the tunnel excavation unloading process, and obtaining the numerically simulated high-risk catastrophic zone," may include sub-steps S130-4 to S130-6: Step S130-4: In the finite element-discrete element coupled numerical model, the stress path change of the surrounding rock caused by tunnel excavation unloading is dynamically simulated by gradually activating the model elements representing the tunnel face.

[0105] Specifically, in the finite element-discrete element coupled numerical model, the tunnel chamber (i.e., the rock mass to be excavated) in its initial state is typically modeled as solid elements. The excavation simulation dynamically and progressively reproduces the tunnel excavation process by gradually activating model elements representing the tunnel face and a certain surrounding area for the next cycle of advance. Furthermore, as the elements representing the rock mass are gradually removed, the loads originally borne by these elements are transferred to the remaining surrounding elements, causing a continuous adjustment and redistribution of the magnitude and direction of stress in the surrounding rock. This process is called stress path change.

[0106] Step S130-5: Analyze the damage evolution law of the surrounding rock during the simulation process. The damage evolution law characterizes the initiation, propagation and penetration process of microcracks in the rock mass under stress, as well as the formation and development law of macroscopic plastic zone or failure zone.

[0107] Specifically, during the simulation, by monitoring the plastic strain, damage variables, or failure of contact forces between discrete element blocks, the initiation location, propagation direction, and interconnected network formation process of microcracks can be visually tracked. Simultaneously, by statistically analyzing and delineating macroscopic plastic or failure zones formed due to element yielding or block system instability, their spatial expansion and connectivity trends as excavation progresses can be observed.

[0108] In this way, the invisible internal damage process of the rock mass is visualized and quantified, including the intermediate process between construction disturbance (stress path change) and the final failure mode (such as collapse), revealing the complete process of how disasters start from local micro-damage, gradually accumulate, and finally erupt.

[0109] Step S130-6: Based on the analysis results of the damage evolution law, identify the potential instability modes and key parts of the surrounding rock, and determine the areas where large-scale damage occurs or significant plastic failure zones are formed as the high-risk disaster areas.

[0110] Specifically, the analysis of damage evolution patterns includes the final form of damage evolution (such as the formation of shear slip surfaces, top collapse zones, or sidewall wedges). By analyzing the final form of damage evolution, potential instability modes can be determined (whether it is large compression deformation, block collapse, or rockburst). At the same time, the key parts controlling this instability mode can be identified, such as the first crack to penetrate, the deepest area of ​​the plastic zone, or the point of greatest stress concentration.

[0111] Among them, the large-scale damage penetration zone refers to the area where the microcrack network is completely connected, resulting in the complete loss of the integrity of the rock mass; the significant plastic failure zone refers to the area where the plastic strain exceeds the critical value and the area has threatened the tunnel temporary support or design outline. The large-scale damage penetration zone or significant plastic failure zone appearing in the simulation is identified in the three-dimensional model and defined as the high-risk disaster zone predicted in this numerical simulation.

[0112] By employing the technical solution of this application, a coherent analysis process involving dynamic driving force input, multi-scale damage process observation, and quantitative identification of risk areas is achieved, enabling high-fidelity numerical reproduction and mechanism diagnosis of the entire process of surrounding rock disaster under tunnel unloading. The identified high-risk disaster areas are based on the results of physical and mechanical law deduction, greatly enhancing the scientific nature and foresight of risk warning, and providing an irreplaceable decision-making basis for taking precise and pre-controllable prevention and control measures.

[0113] In an optional embodiment, step S140 above, which involves "spatially overlaying and coupling the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas to obtain the stability classification and disaster risk warning results of the geological body ahead of the working face," may include sub-steps S140-1 to S130-4: Step S140-1: In the three-dimensional visualization platform, the anomaly areas revealed by the geophysical exploration data, the rock mass mechanical parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas obtained by the numerical simulation are overlaid and displayed.

[0114] Among them, the 3D visualization platform is a software system that integrates a 3D graphics engine and geological information management functions, which can uniformly render and display data volumes with spatial coordinates.

[0115] In this 3D visualization platform, different data layers, such as the anomaly zone from step S110 (e.g., TSP reflection surface, GPR strong reflection zone), the rock mass mechanical parameter weakening zone from step S120 (e.g., 3D cloud map with strength below the threshold), and the numerical simulation high-risk disaster zone from step S130 (e.g., predicted plastic penetration zone), are synchronously and precisely spatially overlaid and displayed in the same 3D scene. This allows for an intuitive and global examination of the spatial interrelationships between macroscopic anomalies, mesoscopic parameter deterioration, and microscopic instability mechanism predictions, laying a visual cognitive foundation for subsequent logical analysis.

[0116] Step S140-2: Based on the superimposed spatial information, establish a stability grading standard; wherein the stability grading standard includes at least three levels: high risk, medium risk and low risk, and the determination basis for each level is the spatial coupling relationship and quantitative indicators among the abnormal area, the parameter weakening area and the high-risk disaster area.

[0117] The stability grading standard integrates three major categories of indicators: the intensity of geophysical anomalies, the degree of weakening of geomechanical parameters, and the scope of catastrophic events in numerical simulations. It clearly divides stability into three levels: high risk, medium risk, and low risk. The determination of each level is based on the spatial coupling relationship (i.e., "triple coupling," "double / partial coupling," or "no coupling") between the anomaly area, the parameter weakening area, and the high-risk catastrophic area, and quantitative indicators (i.e., specific quantitative thresholds combining various indicators).

[0118] For example, the criteria for determining high risk are based on the triple coupling of "strong anomaly - extremely weak zone - large penetration," meaning that the three types of indicators highly overlap and corroborate each other spatially, indicating the existence of a defective geological body with severe properties and prone to disaster. Each level of indicator must simultaneously meet the following criteria: 1) Geophysical indicators: TSP detection shows a reflected wave amplitude ratio greater than 2.0 and continuous phase axes; GPR detection shows strong reflection and waveform disorder, inferred to be water-filled or severely fractured. 2) Geomechanical indicators: Rock mass strength is less than 30% of the design strength; rock mass integrity coefficient is less than 0.35; and the parameter weakening zone is spatially continuous. 3) Numerical simulation indicators: Numerical simulation results show that the plastic zone or damaged zone completely penetrates to the tunnel excavation outline, forming a potential slip wedge or large-area collapse zone, and micro-crack propagation is active.

[0119] The criteria for determining medium risk are based on a dual or partial coupling of "partial anomaly - weakened zone - local development," indicating the presence of clear adverse factors, but without the extreme condition of triple coupling, resulting in reduced stability. The indicators at each level are as follows: 1) Geophysical indicators: TSP or GPR detection shows a medium-intensity anomaly, but the anomaly boundary is blurred or the scale is small. 2) Geomechanical indicators: Rock mass strength is 30% to 60% of the design strength; rock mass integrity coefficient is between 0.35 and 0.55; the parameter weakened zone shows a localized patchy distribution. 3) Numerical simulation indicators: Numerical simulation results show that plastic or damaged zones are locally developed but not fully connected; stress concentration exists, and microcracks show a tendency to propagate.

[0120] The criteria for determining low risk are "no or weak anomalies - normal parameters - stable," meaning all indicators show a good surrounding rock condition or only isolated, weak, and irrelevant anomalies. The indicators at each level are as follows: 1) Geophysical indicators: No significant reflection anomalies detected by TSP or GPR, or only sporadic, isolated, and weak anomalies. 2) Geomechanical indicators: Rock mass strength not less than 60% of the design strength; rock mass integrity coefficient not less than 0.55; uniform distribution of mechanical parameters. 3) Numerical simulation indicators: Numerical simulation results show that the surrounding rock remains in an elastic state after stress redistribution, the plastic zone is basically undeveloped or extremely small, and there is no systematic crack propagation.

[0121] In this way, complex engineering experience, theoretical knowledge, and precise quantitative data are solidified into a standardized and calculable judgment logic. This transforms risk assessment from qualitative judgment relying on personal experience to a quantitative-qualitative combined analysis based on clear rules, improving the objectivity, consistency, and repeatability of the judgment.

[0122] Step S140-3: By calculating the spatial overlap between the abnormal area, the parameter weakening area and the high-risk disaster area, and comparing it with the quantitative indicators in the stability grading standard, the stability level of the geological body in front of the working face is determined.

[0123] Specifically, spatial geometric calculations are used to quantitatively analyze the spatial overlap between three types of regions (for example, defining an overlap area exceeding 70% as "highly coupled"). The calculated spatial overlap results are compared with the quantitative indicator thresholds specified in each level of the standards in step S140-2; simultaneously, it is checked whether the quantitative indicators of each region meet the requirements of the corresponding level. By combining these conditions, a preliminary stability level determination is given according to preset logical rules (such as "AND" and "OR" relationships). Then, engineers can interactively verify the basis of the automatic determination in a 3D visualization platform to finally confirm the preliminary stability level determination, ensuring human-machine collaboration and reliable results.

[0124] Step S140-4: Based on the determined stability level, generate and output a disaster risk early warning report containing the location of the disaster risk, the risk type, the risk level, and recommendations for prevention and control measures.

[0125] Specifically, based on the determined stability level, a report template is automatically invoked to generate a structured disaster risk early warning report. This report, presented in a graphic format, includes at least: the location of the disaster risk (clearly marked on a 3D schematic diagram and plan / section views), the risk type (the inferred disaster type, such as water inrush, landslide, rock burst, etc.), the risk level (a clear assessment of high, medium, or low risk), and recommended prevention and control measures (based on the risk level and type, targeted construction suggestions are automatically linked from a pre-set measures library, such as strengthening support parameters, adjusting tunneling techniques, and implementing pre-reinforcement).

[0126] In this way, the complex analysis results are ultimately distilled into a concise, clear, and highly actionable decision support document. The report is specific and targeted, and can be directly used to dynamically adjust tunneling parameters and optimize support design, achieving seamless integration between advanced geological forecasting and construction decision-making.

[0127] The technical solution adopted in this application proposes a stability grading method based on the spatial coupling of three information sources: macroscopic geophysical anomalies, weakened mesoscopic mechanical parameters, and microscopic catastrophe mechanism simulation, and sets specific quantitative criteria. This fundamentally changes the shortcomings of traditional forecasts, which are fragmented and rely on comprehensive judgment, achieving deep fusion and automated analysis of multi-source information. The resulting early warning report provides a comprehensive decision-making basis with spatial accuracy, mechanistic scientific basis, and targeted measures, thereby greatly improving the timeliness, accuracy, and feasibility of early warning for risks in ultra-long tunnel construction, achieving a leap from passive response to proactive and precise pre-control.

[0128] This application also provides a system for assessing the geological condition of ultra-long tunnels that integrates geophysical, geomechanical, and drilling data, referring to... Figure 3 As shown, Figure 3 This is a schematic diagram of a geological condition assessment system for ultra-long tunnels that integrates geophysical, geomechanical, and drilling data, provided in an embodiment of this application. The system includes: The sensing and data acquisition module 310 is used to acquire geophysical exploration data, lithological data revealed by drilling, and measurement-while-drilling data in front of the tunnel working face. The core intelligent processing and modeling module 320 is used to obtain the preliminary spatial distribution of rock mass mechanical parameters through rock physics model inversion based on the geophysical exploration data and the lithological data revealed by drilling; to correct and spatially interpolate the preliminary spatial distribution using the correlation between the measurement-while-drilling data and the rock mass strength, and to construct a three-dimensional geomechanical model; and, based on the three-dimensional geomechanical model, to construct a finite element-discrete element coupled numerical model, and to analyze the catastrophic evolution mechanism of the surrounding rock by simulating the tunnel excavation unloading process, and to obtain the numerically simulated high-risk catastrophic zone. The application and decision support module 330 is used to perform spatial superposition and coupling analysis on the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas to obtain the stability classification and disaster risk warning results of the geological body in front of the working face.

[0129] In this embodiment, the sensing and data acquisition module 310 is used for multi-parameter, synchronized, and real-time sensing of the geological conditions ahead of the tunnel working face, providing source data assurance for the entire system. For example... Figure 4 As shown, this module is an integrated front-end sensor array, specifically including: geophysical exploration equipment, drilling and borehole exploration equipment, and core testing equipment. The geophysical exploration equipment includes the source and sensor array of a tunnel seismic wave (TSP) detection system, a ground-penetrating radar (GPR) main unit, and an antenna, used to acquire macroscopic geophysical exploration data. The drilling and borehole exploration equipment is an advanced drilling rig integrating measurement-while-drilling (MWD) sensors and a digital borehole camera system; the advanced drilling rig acquires parameters such as drilling speed and torque in real time during drilling (i.e., MWD data); the digital borehole camera system acquires digital borehole camera data characterizing the rock mass structure after drilling. The core testing equipment is a rock mechanics testing machine located in the rear laboratory, used to perform mechanical testing on the rock cores and obtain core mechanical parameters.

[0130] The core intelligent processing and modeling module 320 is deployed in the back-end data center and is responsible for in-depth data processing and core model construction. This module typically consists of a central server, a high-performance computing cluster, and specialized software, and specifically performs two core tasks: three-dimensional geomechanical modeling and numerical simulation of disaster processes.

[0131] The three-dimensional geomechanical modeling includes: receiving multi-source data from the sensing and data acquisition module 310; establishing a rock physics model (e.g., wave velocity-intensity relationship) reflecting the correspondence between rock mass mechanical properties and geophysical parameters through a geomechanical parameter inversion and correction engine; obtaining a preliminary distribution of mechanical parameters (i.e., the preliminary spatial distribution of rock mass mechanical parameters) through rock physics model inversion; subsequently, using the rock mass mechanical parameters at the borehole location in the preliminary spatial distribution as label samples and the corresponding drilling-while-drilling measurement data as feature samples, training a machine learning model to establish a nonlinear mapping relationship between drilling-while-drilling measurement parameters and rock mass mechanical parameters; correcting the preliminary spatial distribution based on the trained machine learning model and performing spatial interpolation on areas not covered by advanced drilling to generate a three-dimensional data volume of rock mass mechanical parameters; and meshing and assigning attribute values ​​to the three-dimensional data volume of rock mass mechanical parameters to obtain the three-dimensional geomechanical model.

[0132] Numerical simulation of the disaster process includes: constructing a finite element-discrete element coupled numerical model that reflects the real structure of the rock mass in a finite element-discrete element coupled numerical simulation software using a multi-scale disaster process engine; dynamically simulating the stress path changes of the surrounding rock caused by tunnel excavation unloading by gradually activating model elements representing the tunnel face in the finite element-discrete element coupled numerical model; analyzing the damage evolution law of the surrounding rock during the simulation process; identifying potential instability modes and key parts of the surrounding rock based on the analysis results of the damage evolution law; and identifying areas where large-scale damage penetration or significant plastic failure zones occur as high-risk disaster zones.

[0133] The application and decision support module 330 serves as the system's decision command center, outputting evaluation conclusions in an intuitive format. This module is a graphical workstation equipped with a 3D visualization platform. Specifically, the 3D visualization platform receives and integrates the 3D geomechanical model and numerical simulation results (such as geomechanical cloud maps and high-risk disaster zones) from the core intelligent processing and modeling module 320, as well as the anomaly zones revealed by the perception and data acquisition module 310. The platform incorporates a stability grading standard and a spatial coupling analysis algorithm, capable of calculating the spatial overlap between anomaly zones, parameter weakening zones, and high-risk disaster zones, and comparing them with the quantitative indicators in the stability grading standard to determine the stability level (high risk, medium risk, low risk) of the geological body ahead of the work area. Finally, based on the determined stability level, it generates and outputs a disaster risk early warning report containing the disaster risk location, risk type, risk level, and prevention and control measure recommendations. This report includes the risk location, type, level, and specific prevention and control measure recommendations, and can be linked to a pre-set engineering measure library.

[0134] Understandably, the sensing and data acquisition module 310, the core intelligent processing and modeling module 320, and the application and decision support module 330 are interconnected through network communication links (such as industrial Ethernet, fiber optics, or wireless networks) to form a closed-loop information flow from on-site sensing, cloud-based intelligent processing to front-end decision support, ensuring the continuity and timeliness of the evaluation process.

[0135] The technical solution adopted in this application integrates dispersed data acquisition, complex modeling and simulation, and efficient decision output into a single system framework, realizing integrated operations of exploration, analysis, and early warning, and greatly improving work efficiency. By solidifying the processes and standards of data processing, model building, and risk analysis through modular software, errors and subjectivity caused by manual operation are avoided, ensuring the scientific, objective, and repeatable nature of the assessment results. Finally, the results are output in the form of an intuitive 3D visualization platform and clear early warning reports, enabling on-site engineers to quickly understand and apply cutting-edge integrated assessment technology. This truly realizes the efficient transformation of scientific research results into engineering productivity, providing powerful systematic tool support for the safe and intelligent construction of ultra-long tunnels.

[0136] This application also provides an electronic device, see embodiments thereof. Figure 5 , Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of this application. For example... Figure 5 As shown, the electronic device 500 includes a memory 510 and a processor 520. The memory 510 and the processor 520 are connected via a bus for communication. The memory 510 stores a computer program that can run on the processor 520, thereby implementing the steps of the method for assessing the geological state of ultra-long tunnels by integrating geophysical, geomechanical, and drilling data as described in the embodiments of this application.

[0137] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0138] This application describes embodiments of methods and systems according to embodiments of this application with reference to flowchart illustrations and / or block diagrams. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0139] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0140] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0141] The above provides a detailed description of the method and system for assessing the geological condition of ultra-long tunnels that integrates geophysical, geomechanical, and drilling data. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for assessing the geological condition of ultra-long tunnels by integrating geophysical, geomechanical, and drilling data, characterized in that... include: Collect geophysical exploration data ahead of the tunnel working face, lithological data revealed by drilling, and measurement-while-drilling data; Based on the geophysical exploration data and the lithological data revealed by drilling, the preliminary spatial distribution of rock mass mechanical parameters is obtained by inversion through a rock physics model. By utilizing the correlation between the measurement-while-drilling data and the rock mass strength, the preliminary spatial distribution is corrected and spatially interpolated through a machine learning model to construct a three-dimensional geomechanical model. Based on the aforementioned three-dimensional geomechanical model, a finite element-discrete element coupled numerical model is constructed. By simulating the unloading process during tunnel excavation, the catastrophic evolution mechanism of the surrounding rock is analyzed, and the high-risk catastrophic zone in the numerical simulation is obtained. By spatially superimposing and coupling analysis of the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas, the stability classification and disaster risk warning results of the geological body in front of the working face are obtained.

2. The method according to claim 1, characterized in that, The geophysical exploration data includes tunnel seismic wave exploration data and ground-penetrating radar detection data; geophysical exploration data collected in front of the tunnel working face includes: By arranging excitation holes and receiving sensor arrays on the rear sidewall of the tunnel excavation face, the tunnel seismic wave exploration data is collected. The tunnel seismic wave exploration data includes the P-wave and S-wave velocity structure, wave impedance profile and reflection interface distribution within the target range in front of the tunnel working face. For key anomaly areas revealed by the tunnel seismic wave exploration data, intensified detection was carried out using ground-penetrating radar to obtain the ground-penetrating radar detection data.

3. The method according to claim 1, characterized in that, The lithological data revealed by drilling includes core mechanical parameters and digital borehole camera data; the lithological data revealed by drilling and the measurement-while-drilling data collected include: Advanced horizontal drilling is carried out, and the drilling measurement data is recorded throughout the drilling process. The drilling measurement data includes drilling speed, torque, thrust, and vibration parameters. After drilling is completed, images of the borehole wall are obtained using digital borehole cameras to obtain digital borehole camera data that characterizes the rock mass structure; Indoor rock mechanics tests were conducted on the drilled core samples to obtain the core mechanical parameters, which include at least one of uniaxial compressive strength, elastic modulus, Poisson's ratio, and cohesion.

4. The method according to claim 3, characterized in that, Based on the geophysical exploration data and the lithological data revealed by drilling, the preliminary spatial distribution of rock mass mechanical parameters is obtained through rock physics model inversion, including: Using the core mechanical parameters and their corresponding geophysical parameters at the borehole locations, a rock physics model reflecting the relationship between the rock mass mechanical properties and the geophysical parameters is established. Substituting the geophysical exploration data into the rock physics model, the preliminary spatial distribution of the rock mass mechanical parameters is obtained by inversion.

5. The method according to claim 1, characterized in that, By utilizing the correlation between the drilling measurement data and rock mass strength, a three-dimensional geomechanical model is constructed by correcting and spatially interpolating the preliminary spatial distribution using a machine learning model, including: Using the rock mechanics parameters at the borehole locations in the preliminary spatial distribution as label samples and the corresponding drilling measurement data as feature samples, a machine learning model is trained to establish a nonlinear mapping relationship between the drilling measurement parameters and the rock mechanics parameters. Based on the trained machine learning model, the preliminary spatial distribution is corrected, and spatial interpolation is performed on the areas not covered by the advance drilling to generate a three-dimensional data volume of rock mass mechanical parameters. The three-dimensional data volume of rock mass mechanical parameters is meshed and its attributes are assigned to obtain the three-dimensional geomechanical model.

6. The method according to claim 5, characterized in that, The three-dimensional data volume of rock mass mechanical parameters is meshed and its attributes are assigned to obtain the three-dimensional geomechanical model, including: Based on the tunnel design axis, a three-dimensional model space is defined. The three-dimensional model space covers at least a first distance in front of the tunnel face along the tunnel excavation direction, and covers a range of at least N times the tunnel diameter in both the horizontal and vertical directions. The three-dimensional model space is discretized by meshing, wherein a dense mesh is used in the region close to the tunnel excavation outline and a sparse mesh is used in the region far from the tunnel excavation outline. The three-dimensional data volume of rock mass mechanical parameters is used as the basic parameter field, and the geological interface obtained by interpreting the geophysical exploration data and the rock mass structural surface information identified by digital borehole camera data in the lithology data revealed by drilling are used as constraints for spatial distribution. Based on the fundamental parameter field and the constraints, rock mechanics parameter values ​​are assigned to each grid cell in the three-dimensional model space using a spatial interpolation algorithm. Based on the rock mass integrity coefficient and structural surface development assigned to the grid cells, the material constitutive model type corresponding to each grid cell in the numerical simulation is determined, thus obtaining the three-dimensional geomechanical model.

7. The method according to claim 1, characterized in that, Based on the aforementioned three-dimensional geomechanical model, a coupled finite element-discrete element numerical model is constructed, including: From the aforementioned three-dimensional geomechanical model, extract the target section model that includes key adverse geological bodies; Based on the rock mass integrity coefficient and structural surface development characteristics of each grid unit in the target section model, the region with a rock mass integrity coefficient greater than the target threshold and no obvious structural surface is classified as a continuous medium region suitable for finite element simulation, and the region with a rock mass integrity coefficient less than the target threshold and joint and fissure development is classified as a discontinuous medium region suitable for discrete element simulation. Based on the division of the continuous medium region and the discontinuous medium region, the finite element-discrete element coupled numerical model is constructed.

8. The method according to claim 1, characterized in that, By simulating the unloading process during tunnel excavation, the catastrophic evolution mechanism of the surrounding rock was analyzed, and the high-risk catastrophic zones identified through numerical simulation were obtained, including: In the finite element-discrete element coupled numerical model, the stress path change of the surrounding rock caused by tunnel excavation unloading is dynamically simulated by gradually activating the model elements representing the tunnel face. The damage evolution law of the surrounding rock during the simulation process is analyzed. The damage evolution law characterizes the initiation, propagation and penetration process of microcracks in the rock mass under stress, as well as the formation and development law of macroscopic plastic zone or failure zone. Based on the analysis results of the damage evolution law, potential instability modes and key parts of the surrounding rock are identified, and areas where large-scale damage occurs or significant plastic failure zones are identified as high-risk disaster zones.

9. The method according to claim 1, characterized in that, By spatially overlaying and coupling analysis of the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas, the stability classification and disaster risk warning results of the geological body ahead of the working face are obtained, including: In the three-dimensional visualization platform, the anomaly areas revealed by the geophysical exploration data, the rock mass mechanical parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas obtained by the numerical simulation are overlaid and displayed. Based on the superimposed spatial information, a stability grading standard is established; wherein the stability grading standard includes at least three levels: high risk, medium risk and low risk, and the determination of each level is based on the spatial coupling relationship and quantitative indicators among the anomaly area, the parameter weakening area and the high-risk disaster area. By calculating the spatial overlap between the anomaly zone, the parameter weakening zone, and the high-risk disaster zone, and comparing it with the quantitative indicators in the stability grading standard, the stability level of the geological body in front of the working face is determined. Based on the determined stability level, a disaster risk early warning report is generated and output, which includes the location of the disaster risk, the type of risk, the level of risk, and recommendations for prevention and control measures.

10. A geological condition assessment system for ultra-long tunnels integrating geophysical, geomechanical, and drilling data, characterized in that... include: The sensing and data acquisition module is used to collect geophysical exploration data ahead of the tunnel working face, lithological data revealed by drilling, and measurement-while-drilling data. The core intelligent processing and modeling module is used to obtain the preliminary spatial distribution of rock mass mechanical parameters through rock physics model inversion based on the geophysical exploration data and the lithological data revealed by drilling. It then uses the correlation between the measurement-while-drilling data and the rock mass strength to correct and spatially interpolate the preliminary spatial distribution through a machine learning model to construct a three-dimensional geomechanical model. Based on the three-dimensional geomechanical model, it constructs a finite element-discrete element coupled numerical model and analyzes the catastrophic evolution mechanism of the surrounding rock by simulating the tunnel excavation unloading process to obtain the numerically simulated high-risk catastrophic zone. The application and decision support module is used to perform spatial overlay and coupling analysis on the anomaly areas revealed by the geophysical exploration data, the parameter weakening areas in the three-dimensional geomechanical model, and the high-risk disaster areas to obtain the stability classification and disaster risk warning results of the geological body in front of the working face.

Images