A Drilling Data-Driven 3D Geological Modeling Method and System for Surrounding Rock of Roadways

By uniformly mapping multi-hole drilling data and constructing a three-dimensional continuous field in the three-dimensional geological modeling of the tunnel, the spatial correlation problem of drilling data was solved, and the three-dimensional expression and modeling of surrounding rock properties were realized.

CN123118948APending Publication Date: 2026-11-17INFORMATION RES INST OF EMERGENCY MANAGEMENT DEPT
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611544586.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-08
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing three-dimensional geological modeling methods for tunnels fail to make full use of drilling data, making it difficult to achieve a three-dimensional continuous expression of properties such as rock strength, rock mass integrity, and surrounding rock quality. Furthermore, multi-hole drilling data lacks a unified spatial correlation, resulting in insufficient modeling precision.

Method used

By establishing a local three-dimensional coordinate system for the tunnel, the multi-hole drilling data is uniformly mapped to a three-dimensional space, a three-dimensional continuous field of drilling data characteristics is constructed, and the surrounding rock properties are identified by adaptively adjusting the three-dimensional spatial unit volume, thus constructing a three-dimensional geological model of the tunnel surrounding rock.

Benefits of technology

It enables quantitative assignment and visualization of properties such as rock strength, rock mass integrity, and surrounding rock quality in three-dimensional space, improving the modeling precision and engineering applicability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123118948A_ABST
    Figure CN123118948A_ABST
Patent Text Reader

Abstract

This invention discloses a three-dimensional geological modeling method and system for tunnel surrounding rock driven by drilling data, relating to the field of surrounding rock property perception and geological modeling technology. Addressing the problems of insufficient utilization of drilling data, lack of unified spatial representation of multi-well data, and difficulty in achieving fine three-dimensional modeling of surrounding rock properties in existing technologies, this invention acquires spatial information and drilling data from multiple boreholes and extracts features from the drilling data. It maps the features of each borehole to a three-dimensional space to obtain a multi-well drilling spatial dataset. A spatial estimation method is used to construct a three-dimensional continuous field of drilling data features. Based on the spatial variation characteristics of the continuous field, adaptive adjustment of unit cells is performed, constructing the drilling data feature vectors for each unit cell. These vectors are then input into a surrounding rock property identification model to obtain surrounding rock property identification results and construct a three-dimensional geological model. This achieves a three-dimensional spatial distribution expression of surrounding rock properties and can be widely applied to surrounding rock quality evaluation and disaster prevention during mining and tunnel engineering construction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of surrounding rock property perception and geological modeling technology, and in particular to a drilling data-driven three-dimensional geological modeling method and system for roadway surrounding rock. Background Technology

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

[0003] Three-dimensional geological modeling is a technology that combines geological information with spatial location information to digitally represent the spatial morphology, structural relationships, and attribute distribution of underground geological bodies. During mining operations, three-dimensional geological models of tunnels can intuitively reflect the spatial distribution of strata and adverse geological formations surrounding the tunnel, providing crucial information for identifying geological anomalies ahead of the tunnel, disaster prevention and control, optimizing construction plans, and ensuring safe and efficient mining. Drilling is an important means of obtaining geological information ahead of the tunnel face, and borehole data can provide reliable spatial constraints for three-dimensional geological modeling.

[0004] Existing 3D geological modeling methods for tunnels primarily rely on geological survey data, borehole data, geological profiles, and tunnel exposure information. These methods construct 3D models of stratigraphic interfaces, faults, and other adverse geological bodies through geological interpretation, spatial interpolation, implicit modeling, or mesh generation. However, these methods emphasize the representation of the geometric shape and spatial boundaries of geological bodies, while being relatively inadequate in the continuous 3D representation of properties such as rock strength, rock mass integrity, and surrounding rock quality. Furthermore, they fail to fully utilize drilling data acquired continuously during drilling, such as drill pressure, rotational speed, torque, and drilling rate.

[0005] In terms of drilling data analysis, existing methods typically focus on a single borehole or a local drilling section, using empirical indicators, statistical analysis, or machine learning to establish a mapping relationship between drilling parameters and surrounding rock properties, obtaining attribute identification results distributed along the borehole depth direction. However, because the borehole opening location, azimuth, dip angle, and drilling depth vary among different boreholes, the drilling data for each borehole are discrete in different spatial locations, lacking a unified coordinate system and spatial correlation. This makes it difficult to directly extend the one-dimensional borehole direction identification results to a three-dimensional spatial distribution of surrounding rock properties ahead of the tunnel face. Summary of the Invention

[0006] To address the shortcomings of existing 3D geological modeling methods for tunnels, such as insufficient utilization of drilling data, lack of unified spatial representation of multi-hole drilling data, and difficulty in achieving precise 3D modeling of surrounding rock properties like rock strength, rock mass integrity, and surrounding rock quality, this invention provides a drilling data-driven 3D geological modeling method and system for tunnel surrounding rocks. By uniformly mapping the features of drilling data from each borehole to a 3D space, a multi-hole drilling spatial dataset is constructed, solving the problem of spatial correlation and unified representation of multi-source discrete drilling data. Furthermore, by constructing a 3D continuous field of drilling data features and adaptively adjusting the 3D spatial unit cells, the problem of representing the spatial variability of surrounding rock properties and precise modeling of inter-bore regions is solved. Finally, by obtaining the surrounding rock property identification results of each unit cell through a surrounding rock characteristic identification model and constructing a 3D geological model, a 3D spatial distribution representation of surrounding rock properties in front of the tunnel face is achieved, effectively improving the precision and engineering applicability of tunnel surrounding rock geological modeling.

[0007] On the one hand, a drilling data-driven three-dimensional geological modeling method for roadway surrounding rock is provided, including: Obtain the drilling space information and drilling data of multiple boreholes in front of the tunnel face, preprocess the drilling data, and extract the drilling data features corresponding to different depth sections of each borehole. Based on the borehole spatial information, the drilling data features of each borehole are mapped to a three-dimensional space to obtain a multi-hole drilling spatial dataset. Based on the multi-hole drilling spatial dataset, a three-dimensional continuous field of drilling data features is constructed using a spatial estimation method. Based on the three-dimensional continuous field of the drilling data features, an initial three-dimensional spatial unit model is established, and the spatial gradient, local variance and feature difference of the internal and adjacent regions of the three-dimensional spatial unit are calculated. Different regions are identified accordingly, and the initial unit model is adaptively adjusted according to the region identification results to form the final unit division result. The drilling data feature vector corresponding to each unit is constructed. The feature vectors of the drilling data of each unit are input into the pre-constructed surrounding rock characteristic identification model to obtain the surrounding rock attribute identification results of each unit. Based on the surrounding rock attribute identification results, a three-dimensional geological model of the surrounding rock of the tunnel is constructed.

[0008] Furthermore, the drilling data includes drilling parameters and petrographic parameters; the drilling parameters include drilling control parameters and drilling response parameters; the drilling control parameters include one or more of drilling pressure and rotational speed, and the drilling response parameters include one or more of torque and drilling speed; the petrographic parameters include one or more of mineral composition, mineral content, and drill cuttings size; the drilling data features include drilling parameter features and petrographic features; the drilling parameter features include one or more of raw drilling parameter features, statistical features, fluctuation features, parameter coupling features, and rock-breaking energy-derived features; the petrographic features include one or more of mineral composition, mineral content, and drill cuttings size.

[0009] Furthermore, when mapping the drilling data features of each borehole to three-dimensional space based on the borehole spatial information, a local three-dimensional coordinate system of the roadway is established with the roadway axis as the first coordinate axis, the roadway transverse direction as the second coordinate axis, and the roadway vertical direction as the third coordinate axis. The three-dimensional spatial coordinates of the drilling data sampling points of each borehole are calculated based on the borehole opening coordinates, azimuth angle, dip angle, drilling depth, and borehole trajectory, and a multi-hole drilling spatial dataset is constructed accordingly.

[0010] Furthermore, the spatial estimation method includes one or more of three-dimensional spatial interpolation methods or spatial inference methods.

[0011] Furthermore, constructing the three-dimensional continuous field of the drilling data features specifically includes: Analyze the distribution patterns, trends, and spatial variability of different drilling data characteristics in the axial, lateral, and vertical directions of the roadway, and determine the correlation distance, variability, main variability direction, and anisotropy ratio parameters in the three directions; Based on the three-dimensional spatial coordinates of the sampling points and the location to be estimated from the drilling data, the coordinate differences between the two in the axial, lateral and vertical directions of the tunnel are calculated. The coordinate differences in different directions are scaled according to the anisotropic scaling parameter to obtain the anisotropic distance between the sampling point and the location to be estimated. Based on the anisotropic distance and spatial correlation parameters, the estimated value of the drilling data characteristics of the inter-hole region is obtained by three-dimensional spatial interpolation or spatial inference methods. Based on the three-dimensional spatial coordinates of the location to be estimated, the estimated values ​​are spatially organized and correlated to form a three-dimensional continuous field corresponding to each drilling data feature.

[0012] Furthermore, the adaptive adjustment of the initial three-dimensional spatial unit model and the construction of the feature vector of the drilling data specifically include: Based on the spatial range of the modeling area in front of the tunnel face, an initial three-dimensional spatial unit model is established according to the preset initial unit size; Based on the three-dimensional continuous field of each drilling data feature, the drilling data feature values ​​inside and adjacent regions of each initial three-dimensional spatial unit are obtained, and the spatial gradient, local variance and feature difference are calculated respectively. The spatial gradient, local variance, and feature difference of each spatial unit are compared with the corresponding preset judgment thresholds to identify stable regions, gradual regions, abrupt regions, and local abnormal regions in the drilling data. Following the rules of merging stable regions and subdividing abrupt regions, the initial three-dimensional spatial unit cells are adaptively refined or merged to form the unit cell division results. After completing the adaptive partitioning, for each type of drilling data feature, the representative value of the drilling data corresponding to each spatial unit is calculated using one of the following: mean, weighted mean, median, truncated mean, or quantile. The representative values ​​of different types of drilling data corresponding to each spatial unit are combined in a preset order to form the feature vector of drilling data corresponding to each unit.

[0013] Furthermore, the surrounding rock characteristic identification model adopts one of artificial neural networks, support vector machines, random forests, or gradient boosting trees, taking the feature vector of drilling data as input and the corresponding measured value of surrounding rock attributes as output; the surrounding rock attributes include one or more of rock strength, rock mass integrity index, surrounding rock quality index, and surrounding rock grade.

[0014] On the other hand, a drilling data-driven 3D geological modeling system for roadway surrounding rock is provided, including: The data acquisition and preprocessing module is configured to: acquire the borehole spatial information and drilling data of multiple boreholes in front of the tunnel face, preprocess the drilling data, and extract the drilling data features corresponding to different depth sections of each borehole. The three-dimensional spatial positioning module is configured to: map the drilling data features of each borehole to three-dimensional space according to the borehole spatial information to obtain a multi-hole drilling spatial dataset; The continuous field construction module is configured to: construct a three-dimensional continuous field of drilling data features based on the multi-hole drilling spatial dataset using a spatial estimation method; The adaptive partitioning and feature construction module is configured to: establish an initial three-dimensional spatial unit model based on the three-dimensional continuous field of the drilling data features; calculate the spatial gradient, local variance and feature difference of the three-dimensional spatial unit and its adjacent regions; identify different regions accordingly; and adaptively adjust the initial unit model according to the region identification results to form the final unit partitioning result; and construct the drilling data feature vector corresponding to each unit. The surrounding rock attribute identification and modeling module is configured to: input the feature vector of the drilling data of each unit into the pre-constructed surrounding rock characteristic identification model, obtain the surrounding rock attribute identification results of each unit, and construct a three-dimensional geological model of the surrounding rock of the tunnel based on the surrounding rock attribute identification results.

[0015] In another aspect, a computer device is also provided, including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, wherein when the processor executes the program, it performs the method described in the first aspect.

[0016] In another aspect, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, performs the method described in the first aspect.

[0017] The above technical solution has the following advantages or beneficial effects: (1) By establishing a local three-dimensional coordinate system for the roadway, the multi-hole drilling data from different locations and directions are uniformly transformed into the same spatial framework, realizing the mapping and spatial association of one-dimensional drilling data to three-dimensional spatial discrete points, effectively solving the problem that multi-source drilling data are difficult to integrate due to differences in borehole position, azimuth angle and dip angle.

[0018] (2) By performing three-dimensional spatial interpolation or spatial inference on discrete drilling data, a three-dimensional continuous field of drilling data features is constructed, which overcomes the shortcomings of traditional methods that lack measured data for the region between boreholes and cannot be finely characterized. At the same time, different regions are identified based on the spatial gradient, local variance and feature difference of the continuous field, and the initial unit model is adaptively adjusted according to the region identification results, so that the unit is automatically refined in the region with drastic attribute changes and automatically merged in the region with stable attributes. Compared with the traditional spatial division method of fixed segment spacing and fixed block, the modeling accuracy of the region between boreholes and the boundary of surrounding rock changes is significantly improved.

[0019] (3) By inputting the feature vector of the drilling data of each unit into the surrounding rock property identification model, the identification results of the surrounding rock properties of each unit are obtained and a three-dimensional geological model of the surrounding rock of the tunnel is constructed. This realizes the quantitative assignment and visualization of properties such as rock strength, rock mass integrity and surrounding rock quality in three-dimensional space, so that the modeling results can directly serve the evaluation of the surrounding rock quality of the tunnel and the optimization decision of the construction plan. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention; Figure 2 This is a flowchart of the three-dimensional continuous reconstruction of borehole drilling data in Embodiment 1 of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings. Those skilled in the art should understand that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0023] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] Example 1 In order to achieve unified spatial positioning, continuous reconstruction, and surrounding rock attribute assignment of multi-hole drilling data, realize the three-dimensional expression of attribute information such as rock strength, rock mass integrity, and surrounding rock quality, and improve the precision and engineering applicability of tunnel surrounding rock geological modeling, this embodiment provides a drilling data-driven three-dimensional geological modeling method for tunnel surrounding rock.

[0025] Figure 1 This is an overall flowchart of the method according to Embodiment 1 of the present invention. The method includes the following steps: S1: Obtain the borehole spatial information and drilling data of multiple boreholes in front of the tunnel face, preprocess the drilling data, and extract the drilling data features corresponding to different depth sections of each borehole. S2: Determine the three-dimensional spatial coordinates of each drilling data sampling point based on the borehole spatial information, map the drilling data features of each borehole to three-dimensional space, and obtain a multi-hole drilling spatial dataset. S3: Based on the multi-hole drilling spatial dataset, a three-dimensional continuous field of drilling data features is constructed using spatial estimation methods; S4: Based on the three-dimensional continuous field of drilling data features, establish an initial three-dimensional spatial unit model, calculate the spatial gradient, local variance and feature difference of the three-dimensional spatial unit and its adjacent regions, and identify stable regions, gradual regions, abrupt regions and local abnormal regions accordingly. Adaptively refine or merge the spatial units to form the three-dimensional spatial unit partitioning results, and construct the drilling data feature vector corresponding to each spatial unit. S5: Input the feature vector of each unit's drilling data into the pre-constructed surrounding rock characteristic identification model to obtain the surrounding rock attribute identification results of each unit, and assign the surrounding rock attribute identification results to the corresponding spatial unit to construct a three-dimensional geological model of the tunnel surrounding rock.

[0026] The following is a detailed explanation of each step.

[0027] In step S1, the borehole spatial information and drilling data of multiple boreholes ahead of the tunnel face are first acquired. The borehole spatial information includes one or more of the following: borehole opening location, azimuth angle, dip angle, drilling depth, and borehole trajectory. The drilling data includes drilling parameters and petrographic parameters. Drilling parameters are directly acquired by a drilling parameter monitoring system during drilling at a preset sampling frequency, including drilling control parameters and drilling response parameters. Drilling control parameters include one or more of drill pressure and rotational speed, and drilling response parameters include one or more of torque and drilling speed. In this embodiment, the preferred drilling control parameters are a combination of drill pressure and rotational speed, and the preferred drilling response parameters are a combination of torque and drilling speed. Petrographic parameters are obtained by collecting drill cuttings generated during drilling and performing image recognition and spectral detection on the drill cuttings, including one or more of the following: mineral composition, mineral content, and drill cutting particle size. In this embodiment, the preferred selection is a combination of mineral composition, mineral content, and drill cutting particle size.

[0028] After obtaining the above data, the drilling data is preprocessed. Preprocessing methods include one or more of outlier identification and removal, noise filtering, data smoothing, depth alignment, and normalization. After preprocessing, the continuous data is segmented according to a preset drilling depth interval or a sliding window, and drilling data features are extracted from the drilling data corresponding to different depth segments of each borehole. The preset drilling depth interval is determined based on the original sampling frequency of the drilling parameters, the drilling speed, the spatial modeling resolution, and the minimum scale of the geological anomaly to be identified. The window length of the sliding window is not less than the preset drilling depth interval, and a preset overlap ratio is set between adjacent sliding windows. As one implementation, this embodiment performs depth alignment of the drilling data according to a drilling depth interval of 0.1m, and uses a sliding window with a length of 0.5m and a sliding step size of 0.1m to segment the continuous data.

[0029] Drilling data features include drilling parameter features and petrographic features. Drilling parameter features include one or more of the following: raw drilling parameter features, statistical features, fluctuation features, parameter coupling features, and rock-breaking energy-derived features. In this embodiment, a combination of raw drilling parameter features, statistical features, fluctuation features, parameter coupling features, and rock-breaking energy-derived features is preferred. Raw drilling parameter features are drilling parameter values ​​such as drill pressure, rotational speed, torque, and drilling speed directly collected by the drilling parameter monitoring system at a preset sampling frequency, or representative values ​​of a section obtained after depth alignment of the drilling parameter values. Statistical features reflect the overall level and dispersion of drilling parameters within a preset drilling depth section, including one or more of the following: mean, variance, maximum value, minimum value, median, and range. In this embodiment, a combination of mean, variance, maximum value, minimum value, median, and range is preferred. The fluctuation characteristics reflect the local fluctuations and abrupt changes of drilling parameters along the drilling depth direction. They include one or more of the following: difference between adjacent sampling points, rate of change, slope of change, peak-valley difference, and fluctuation frequency. In this embodiment, a combination of difference between adjacent sampling points, rate of change, slope of change, peak-valley difference, and fluctuation frequency is preferred.

[0030] The parameter coupling feature is a feature obtained by two or more drilling parameters through ratio, difference, product, correlation analysis or combination operation, which reflects the relationship between drilling control parameters and drilling response parameters. It includes one or more of the following: the ratio of drilling pressure to drilling speed, the ratio of rotational speed to torque, the ratio of torque to drilling pressure, and the correlation coefficients of different drilling parameters in the corresponding drilling depth range. In this embodiment, the preferred choice is a combination of the ratio of drilling pressure to drilling speed, the ratio of rotational speed to torque, the ratio of torque to drilling pressure, and the correlation coefficients of different drilling parameters in the corresponding drilling depth range.

[0031] The rock-breaking energy derivative characteristics are one or more of the rock-breaking energy per unit footage and mechanical specific energy calculated based on parameters such as drilling pressure, torque, rotational speed, drilling speed and drill bit size, combined with the drilling rock-breaking energy relationship. In this embodiment, the preferred choice is a combination of rock-breaking energy per unit footage and mechanical specific energy.

[0032] Lithofacies features are characteristics selected from lithofacies parameters or formed through statistical processing, used to characterize the rock composition and microstructure within the corresponding drilling depth section. These features include one or more of mineral composition, mineral content, drill cuttings size, drill cuttings shape, heterogeneity, and image color features. In this embodiment, a combination of mineral composition, mineral content, drill cuttings size, drill cuttings shape, heterogeneity, and image color features is preferred.

[0033] Through the aforementioned preprocessing and feature extraction, the original drilling signals are transformed into quantitative indicators that reflect changes in surrounding rock properties, providing a standardized data foundation for subsequent spatial positioning and modeling. The extracted drilling data features encompass information on drilling control parameters, drilling response parameters, and lithofacies parameters across multiple dimensions, including original values, statistical distributions, fluctuations, parameter coupling, rock-breaking energy, and lithofacies composition. These quantitative indicators, which characterize surrounding rock properties from different perspectives, can provide sufficient input information for subsequent identification of surrounding rock attributes.

[0034] In step S2, after completing the data acquisition and preprocessing in step S1, the three-dimensional spatial positioning of the drilling data features of each borehole is performed.

[0035] Specifically, a local three-dimensional coordinate system for the roadway is established, with the roadway axis as the first coordinate axis, the roadway transverse direction as the second coordinate axis, and the roadway vertical direction as the third coordinate axis. Based on the borehole opening coordinates, azimuth, dip angle, drilling depth, and borehole trajectory, the three-dimensional spatial coordinates of each borehole's drilling data sampling point are calculated, transforming the one-dimensional drilling data distributed along the borehole depth direction into discrete drilling data points with three-dimensional spatial positions.

[0036] In one specific implementation of this embodiment, the borehole data comes from advance drilling operations in the roadway, and all are straight holes. Let the coordinates of the borehole opening be... The borehole azimuth angle is The angle of inclination is Where the azimuth angle is the angle between the projection direction of the borehole onto the horizontal plane and the first coordinate axis, and the inclination angle is the angle between the borehole axis and the horizontal plane, then the coordinates of the sampling point at a drilling depth s from the borehole opening are... Calculate using the following formula: ; ; ; By calculating the three-dimensional spatial coordinates of each sampling point one by one using the above formula, the one-dimensional drilling data distributed along the borehole depth direction is converted into discrete drilling data points with three-dimensional spatial positions.

[0037] Then, based on the three-dimensional spatial coordinates, drilling data characteristics, borehole number, and drilling depth of each discrete drilling data point, a multi-hole drilling spatial dataset is constructed. This dataset unifies the drilling data from different boreholes into the same spatial coordinate system, providing a spatial data foundation for the estimation of drilling data characteristics and the construction of a three-dimensional continuous field in the inter-hole region in subsequent steps.

[0038] In step S3, to address the issues of discrete distribution of multi-hole drilling data and lack of measured data in inter-hole regions, this embodiment performs three-dimensional continuous reconstruction of the multi-hole drilling data features, such as... Figure 2 As shown. Based on the multi-hole drilling spatial dataset constructed in step S2, the spatial distribution, correlation, and anisotropy of the drilling data features are analyzed. Three-dimensional spatial interpolation or spatial inference methods are used to estimate the drilling data feature values ​​in the inter-hole region, and a three-dimensional continuous field of drilling data features and its spatial uncertainty results are constructed.

[0039] Specifically, step S3 includes the following sub-steps: S3-1, Spatial Distribution and Variation Feature Analysis. Based on the three-dimensional spatial coordinates of each sampling point in the multi-hole drilling spatial dataset, the sampling points are divided along the tunnel axis, transverse direction, and vertical direction, and grouped according to a preset spatial distance interval. For each type of drilling data feature, the mean, standard deviation, and coefficient of variation of the feature values ​​of each group of sampling points are statistically analyzed to determine their distribution in different directions and spatial locations. Moving average or linear fitting methods are used to analyze the changing trends of the drilling data features along different directions. Directional variability function or spatial correlation analysis methods are used to analyze the degree of spatial variation of the drilling data features in different directions, providing a basis for determining subsequent spatial correlation parameters and anisotropy ratio parameters.

[0040] S3-2, Determination of Spatial Correlation Parameters and Anisotropy Ratio Parameters. Based on the directional variability function or spatial correlation analysis results obtained in step S3-1, the correlation distances and variability degrees of each drilling data feature in the axial, lateral, and vertical directions of the tunnel are determined. Specifically, the distance corresponding to the directional variability function stabilizing with increasing sampling point spacing is determined as the correlation distance, and the variability degree is determined based on the stable value and variation amplitude of the directional variability function. By comparing the correlation distances and variability degrees in different directions, the direction with the larger correlation distance and stronger spatial continuity is determined as the main variability direction. Using the correlation distance in the main variability direction as a benchmark, the anisotropy ratio parameter is determined based on the ratio between the correlation distances in other directions and the correlation distance in the main variability direction, providing a parameter basis for subsequent anisotropic distance calculations and drilling data feature estimation in the inter-hole area.

[0041] S3-3, Spatial Relationship and Anisotropic Distance Calculation. Based on the three-dimensional spatial coordinates of the sampling point and the location to be estimated from the drilling data, calculate the coordinate differences between them in the axial, lateral, and vertical directions of the tunnel. When the principal anisotropic direction is inconsistent with the local three-dimensional coordinate axes of the tunnel, transform the coordinate differences to a coordinate system based on the principal anisotropic direction. According to the correlation distances or anisotropic scaling parameters determined in step S3-2, scale the coordinate differences in different directions respectively, and synthesize the scaled coordinate differences in each direction to obtain the anisotropic distance between the sampling point and the location to be estimated. Directions with larger correlation distances correspond to smaller distance scaling, and directions with smaller correlation distances correspond to larger distance scaling. The smaller the anisotropic distance, the stronger the spatial correlation between the sampling point and the location to be estimated.

[0042] As one implementation method, let the coordinate differences between the sampling point and the position to be estimated in the coordinate system of the anisotropic principal directions be respectively. , , The relevant distances in the three directions are respectively , , Then the anisotropic distance It can be represented as: ; The relevant distance is used to scale the coordinate difference in the corresponding direction.

[0043] S3-4, Drilling data feature estimation for inter-hole regions. Based on the anisotropic distance and spatial correlation parameters between the sampling point and the location to be estimated, three-dimensional spatial interpolation or spatial inference methods are used to obtain the estimated drilling data feature values ​​for areas between boreholes lacking measured data. In this embodiment, three-dimensional spatial interpolation is preferred.

[0044] When using a three-dimensional spatial interpolation method, the influence weight of each sampling point on the estimated location is calculated based on anisotropic distance and spatial correlation. Sampling points with smaller anisotropic distances and stronger spatial correlations have greater influence weights. Then, the feature values ​​of the drilling data of each sampling point in the neighborhood of the estimated location are weighted and combined according to their influence weights to obtain the estimated drilling data feature value corresponding to the estimated location. When using Gaussian process regression or machine learning-based spatial inference methods, the estimated drilling data feature value of the estimated location is directly output by the corresponding spatial inference model.

[0045] S3-5, Construction of a 3D Continuous Field for Drilling Data Features. Based on the 3D spatial coordinates of the locations to be estimated, the estimated values ​​of the drilling data features obtained in step S3-4 are spatially organized and correlated to form a 3D continuous field corresponding to each drilling data feature in front of the tunnel face. Specifically, based on the 3D spatial coordinates of each location to be estimated, a corresponding 3D grid index is set, and the estimated values ​​of the drilling data features obtained in step S3-4 are filled into the corresponding 3D grid nodes, ensuring that each 3D grid node corresponds to its 3D spatial coordinates and the drilling data feature value. For grid nodes with measured drilling data feature values, the measured values ​​are retained first; for grid nodes lacking measured data between boreholes, the estimated values ​​obtained in step S3-4 are used. The grid nodes are organized according to their spatial arrangement in the tunnel axis, transverse direction, and vertical direction to form a 3D continuous field covering the modeling area in front of the tunnel face. When multiple drilling data features are included, the same 3D grid nodes and grid indices are used to construct the 3D continuous fields corresponding to each type of drilling data feature.

[0046] The final output is a three-dimensional continuous field of drilling data features and its spatial uncertainty results, providing foundational data for the adaptive partitioning of three-dimensional spatial units and the construction of drilling data feature vectors in step S4. This step introduces an anisotropic scaling parameter, enabling the interpolation process to adaptively adjust weight allocation based on the variability of the geological body in different directions. It also allows the spatial estimation process to adjust the influence range of each sampling point based on the correlation distance and variability of the drilling data features in different directions, thereby better characterizing the spatial variation differences of the drilling data features in different directions and providing a continuous data foundation for the subsequent adaptive partitioning of three-dimensional spatial units.

[0047] In step S4, based on the three-dimensional continuous field of drilling data features obtained in step S3, the modeling area in front of the tunnel face is adaptively divided into three-dimensional spatial units, and the drilling data feature vector corresponding to each three-dimensional spatial unit is constructed.

[0048] Specifically, step S4 includes the following sub-steps: S4-1, Initial three-dimensional spatial unit model establishment.

[0049] Based on the spatial range of the modeling area in front of the tunnel face in the axial, lateral and vertical directions of the tunnel, the modeling area is divided into multiple regular cube units according to the preset initial unit size.

[0050] The initial element size is determined based on the borehole spacing, the depth sampling interval of the drilling data, the size of the modeling area, and the target spatial resolution. When the borehole spacing is small or the required modeling resolution is high, a smaller initial element size can be set; when the borehole spacing is large or the modeling area is large, the initial element size can be appropriately increased.

[0051] Based on the arrangement of each initial spatial unit in the axial, lateral, and vertical directions of the roadway, a unique spatial index is set for each initial spatial unit, and its center point coordinates, side length, and spatial range are recorded, thereby establishing an initial three-dimensional spatial unit model covering the modeling area in front of the roadway face.

[0052] S4-2, Calculation of the variation characteristics of spatial unit cells during drilling data Based on the three-dimensional continuous field of each drilling data feature obtained in step S3, the drilling data feature values ​​of each initial spatial unit and its adjacent regions are acquired, and the spatial gradient, local variance, and feature difference are calculated respectively. Specifically: the spatial gradient is determined based on the variation of the drilling data feature values ​​within the spatial unit along the tunnel axis, transverse direction, and vertical direction, and is used to characterize the rate of change of the drilling data features in space; the local variance is determined based on the dispersion of the drilling data feature values ​​within the current spatial unit and its adjacent spatial units relative to the local average value, and is used to characterize the degree of fluctuation of the drilling data features within a local region; the feature difference is determined based on the difference in the representative values ​​of the drilling data features between the current spatial unit and its adjacent spatial units, and is used to characterize the degree of abrupt change in the drilling data features between adjacent spatial units.

[0053] When multiple drilling data features exist, the spatial gradient, local variance, and feature difference corresponding to each drilling data feature are calculated separately, and the different indicators are normalized to eliminate the differences in the units and numerical ranges of different drilling data features.

[0054] S4-3, Identification of Spatial Variation Regions in Drilling Data. The spatial gradient, local variance, and feature difference of each spatial unit are compared with the corresponding preset judgment thresholds to identify stable regions, gradual change regions, abrupt change regions, and local abnormal regions in the drilling data.

[0055] Specifically, when the spatial gradient, local variance, and feature difference are all less than the corresponding first judgment threshold, the corresponding spatial unit and its adjacent regions are identified as stable regions. When the spatial gradient, local variance, or feature difference is between the first and second judgment thresholds, and the drilling data features show continuous spatial variation, the corresponding region is identified as a gradual change region. When the spatial gradient or feature difference between adjacent spatial units is greater than the corresponding second judgment threshold, and high values ​​are continuously distributed among adjacent spatial units, the corresponding region is identified as a sudden change region. When the local variance or feature difference is significantly higher than that of adjacent regions, but the high values ​​are concentrated in only one or a few adjacent spatial units, the corresponding region is identified as a local anomaly region. The first and second judgment thresholds can be determined based on the statistical distribution of the spatial gradient, local variance, and feature difference within the modeling region, or based on historical modeling data, field measurement results, or preset quantiles. In this embodiment, preferably, the first and second judgment thresholds are determined based on the statistical distribution of the spatial gradient, local variance, and feature difference within the modeling region, wherein the first judgment threshold is less than the second judgment threshold.

[0056] S4-4, Adaptive Refinement and Merging of Spatial Units. Based on the identification results of spatially varying regions in the drilling data, the initial three-dimensional spatial units are adaptively refined or merged.

[0057] For stable regions of drilling data, multiple spatial cell blocks that are spatially adjacent, of the same size, and whose difference in representative values ​​of drilling data features is less than a preset merging threshold are merged into a larger spatial cell block to reduce the number of cells in the stable region.

[0058] For areas of abrupt changes in drilling data, the corresponding spatial unit is divided equally along the tunnel axis, transverse direction, and vertical direction to form multiple smaller sub-units, thereby improving the spatial resolution of areas with changing surrounding rock properties.

[0059] For local anomaly regions in the drilling data, the spatial cell containing the anomaly location is subdivided to preserve local anomaly characteristics; for gradually changing regions in the drilling data, the initial cell size can be maintained, or a first-level subdivision can be performed according to the modeling accuracy requirements.

[0060] The spatial gradient, local variance, and feature difference are recalculated for the subdivided sub-units. If the calculation results still meet the subdivision conditions, the subdivision continues until the corresponding indicators are lower than the preset subdivision threshold, or the unit reaches the preset minimum size or maximum subdivision level. In this embodiment, the second judgment threshold is used as the preset subdivision threshold, and its value is determined in the same way as in S4-3, that is, based on the statistical distribution of the spatial gradient, local variance, and feature difference within the modeling region.

[0061] Through the above processing, larger spatial units are formed in areas where the characteristics of the drilling data change little, and smaller spatial units are formed in areas where the characteristics of the drilling data change drastically or where there are local anomalies, thereby obtaining adaptive partitioning results of the three-dimensional spatial units of the roadway.

[0062] S4-5, Calculation of representative values ​​of spatial unit data during drilling.

[0063] After completing the adaptive partitioning of the three-dimensional spatial units, based on the spatial range of each three-dimensional spatial unit, the three-dimensional mesh nodes falling within the corresponding spatial unit and their corresponding drilling data feature values ​​are obtained from the three-dimensional continuous field of drilling data features constructed in step S3. Specifically, the three-dimensional mesh nodes corresponding to the measured drilling data sampling points use the measured drilling data feature values, while the three-dimensional mesh nodes lacking measured data between boreholes use the estimated drilling data feature values ​​obtained in steps S3-4. For spatial units that do not contain three-dimensional mesh nodes within their spatial range, the drilling data feature value corresponding to the center position of that spatial unit is used.

[0064] For each type of drilling data feature, based on the drilling data feature values ​​of each 3D grid node within the corresponding spatial unit, a representative value of the drilling data for that spatial unit is calculated using one of the following methods: average, weighted average, median, truncated average, or quantile. Specifically, when the distribution of drilling data feature values ​​within the spatial unit is relatively stable, the average or weighted average is used; when there are a few outliers or the data fluctuates significantly, the median or truncated average is used; and when it is necessary to characterize different distribution levels of drilling data feature values, a preset quantile is used.

[0065] When using a weighted average, the weights are determined based on the distance between each 3D mesh node and the center point of the spatial unit cell, with 3D mesh nodes closer to the center point of the spatial unit cell having greater weights. For cases involving multiple drilling data features, representative values ​​for each type of drilling data feature are calculated separately, providing a basis for constructing the unit cell drilling data feature vector in steps S4-6.

[0066] S4-6, Construction of feature vectors for unit-based drilling data.

[0067] The representative values ​​of different types of drilling data corresponding to each spatial unit are combined according to the feature types and arrangement order predetermined in the training stage of the surrounding rock characteristic identification model to form the unit drilling data feature vector corresponding to each spatial unit.

[0068] For example, the following can be combined in the order of drilling pressure characteristics, rotational speed characteristics, torque characteristics, drilling speed characteristics, drilling energy characteristics, drilling parameter coupling characteristics, and lithofacies characteristics to obtain: The feature vector of the unit while drilling data = [representative value of drilling pressure, representative value of rotation speed, representative value of torque, representative value of drilling speed, representative value of drilling energy, representative value of drilling parameter coupling feature, representative value of petrographic feature].

[0069] When each type of drilling data contains multiple features such as mean, standard deviation, and coefficient of variation, they are arranged sequentially according to a pre-established list of feature names.

[0070] Each three-dimensional spatial unit corresponds to a unit-based drilling data feature vector. The feature type, arrangement order, and normalization method of the unit-based drilling data feature vector are consistent with the input features used in the training phase of the surrounding rock characteristic identification model, and are used as the input to the surrounding rock characteristic identification model in step S5.

[0071] Finally, step S4 outputs the three-dimensional spatial unit division results of the roadway and the feature vector of the unit while drilling data corresponding to each spatial unit, providing input data for the identification and assignment of the surrounding rock attributes of the spatial unit in step S5.

[0072] In step S5, the surrounding rock properties of the spatial unit are identified and assigned values.

[0073] Specifically, the feature vectors of the drilling data of each three-dimensional spatial unit constructed in step S4 are input into the pre-constructed surrounding rock characteristic identification model to obtain the surrounding rock attribute identification results corresponding to the three-dimensional spatial unit.

[0074] A surrounding rock characteristic identification model is used to establish the mapping relationship between drilling data features and surrounding rock attributes. It can employ artificial neural networks, support vector machines, random forests, gradient boosting trees, or other regression or classification models. In this embodiment, an artificial neural network is preferably selected. The specific model type is chosen based on the type of surrounding rock attribute to be identified and the characteristics of the training samples. For continuous surrounding rock attributes such as rock strength, rock mass integrity indicators, and surrounding rock quality indicators, a regression model is used; for discrete surrounding rock attributes such as surrounding rock grade, a classification model is used.

[0075] The training samples for the surrounding rock property identification model are derived from historical or experimental boreholes where surrounding rock properties have been measured. During the construction of the training samples, the drilling section corresponding to the measured location of the surrounding rock properties is used as the training sample section. Following the same feature statistics method as in step S4, representative values ​​corresponding to different types of drilling data within this training sample section are calculated and combined according to the same feature types and arrangement order as in step S4 to form the drilling data feature vector corresponding to the training sample. The measured value of the surrounding rock properties corresponding to the training sample section serves as the output label for that training sample. Specifically, the measured rock strength value is obtained through core uniaxial compression tests, point load tests, or other rock mechanics tests; the rock mass integrity index is obtained through borehole imaging, core integrity, fracture statistics, or acoustic testing; and the surrounding rock quality index and surrounding rock grade are determined based on rock strength, rock mass integrity, and relevant engineering geological indicators.

[0076] The training samples are divided into a training set and a validation set. The training set is used to train the surrounding rock characteristic identification model, and the validation set is used to verify the model's recognition performance, thus obtaining a surrounding rock characteristic identification model that meets the preset accuracy requirements. The feature vectors of the drilling data used in the training phase of the surrounding rock characteristic identification model have the same feature type, arrangement order, and normalization method as the feature vectors of the unit volume drilling data constructed in step S4.

[0077] The feature vectors of the drilling data corresponding to each three-dimensional spatial unit are sequentially input into the trained surrounding rock characteristic recognition model to obtain one or more of the following: rock strength, rock mass integrity index, surrounding rock quality index, and surrounding rock grade for each three-dimensional spatial unit. In this embodiment, it is preferable to obtain the above four surrounding rock attributes simultaneously to comprehensively characterize the quality of the roadway surrounding rock.

[0078] After obtaining the surrounding rock attribute identification results, the three-dimensional spatial unit division results formed in step S4 are used as the spatial geometric carrier. Based on the unique spatial index of each three-dimensional spatial unit, the surrounding rock attribute identification results are assigned to the corresponding three-dimensional spatial unit, so that each three-dimensional spatial unit has the corresponding surrounding rock attributes, thereby constructing a three-dimensional geological model of the tunnel surrounding rock.

[0079] The three-dimensional geological model of the tunnel surrounding rock includes the spatial geometric information of each three-dimensional spatial unit and its corresponding surrounding rock properties. The spatial geometric information includes one or more of the following: unique spatial index of the three-dimensional spatial unit, center point coordinates, spatial dimensions, and spatial range. The surrounding rock properties include one or more of the following: rock strength, rock mass integrity index, surrounding rock quality index, and surrounding rock grade. In this embodiment, both the spatial geometric information and the surrounding rock properties are preferably output using all types.

[0080] After completing the identification of surrounding rock attributes and the construction of the 3D geological model, the model is visualized. Based on the constructed 3D geological model of the tunnel surrounding rock, the attributes of different 3D spatial units are color-mapped and graded, forming a 3D spatial distribution result of rock strength, rock mass integrity, surrounding rock quality, or surrounding rock grade in front of the tunnel face. The 3D geological model of the tunnel surrounding rock is output in one or more forms, such as a 3D block model, a 3D mesh model, a 3D isosurface, a longitudinal profile, a transverse profile, or an attribute cloud map. In this embodiment, a combination of a 3D block model and an attribute cloud map is preferred. The 3D block model is used to display the overall spatial distribution of surrounding rock attributes, while the attribute cloud map is used to intuitively reflect the continuous changes in each attribute value. The combination of the two provides construction personnel with intuitive 3D spatial information. Through visualization, construction personnel can intuitively obtain the spatial distribution of the surrounding rock quality in front of the tunnel face, providing an intuitive basis for support scheme optimization and disaster prevention decisions.

[0081] In summary, this embodiment provides a method for three-dimensional geological modeling of tunnel surrounding rock driven by drilling data. This method first preprocesses and extracts features from drilling data from multiple boreholes, and then maps the drilling data features to a unified three-dimensional coordinate system of the tunnel based on borehole spatial information. Next, considering the spatial correlation and variation differences of the drilling data features in different directions, a three-dimensional continuous field of drilling data features is constructed using a spatial estimation method. Then, based on the spatial gradient, local variance, and feature difference of the continuous field, the three-dimensional spatial units are adaptively refined or merged, and a corresponding drilling data feature vector is constructed for each unit. Finally, a surrounding rock property identification model is used to obtain and assign corresponding surrounding rock properties to each unit, thereby constructing a three-dimensional geological model of the tunnel surrounding rock, realizing the three-dimensional spatial assignment and visualization of surrounding rock properties such as rock strength, rock mass integrity, surrounding rock quality, and surrounding rock grade. This method can fully utilize multi-hole drilling data obtained during tunnel construction, improve the precision of spatial modeling of surrounding rock properties, and provide three-dimensional spatial information for tunnel surrounding rock quality evaluation, construction scheme optimization, and disaster prevention.

[0082] Example 2 This embodiment provides a drilling data-driven 3D geological modeling system for roadway surrounding rock, including: The data acquisition and preprocessing module is configured to: acquire the borehole spatial information and drilling data of multiple boreholes in front of the tunnel face, preprocess the drilling data, and extract the drilling data features corresponding to different depth sections of each borehole. The three-dimensional spatial positioning module is configured to: map the features of the drilling data of each borehole to three-dimensional space based on the borehole spatial information to obtain a multi-hole drilling spatial dataset; The continuous field construction module is configured to: construct a three-dimensional continuous field of drilling data features based on the multi-hole drilling spatial dataset and using spatial estimation methods. The adaptive partitioning and feature construction module is configured to: establish an initial three-dimensional spatial unit model based on the three-dimensional continuous field of the drilling data features; calculate the spatial gradient, local variance and feature difference of the three-dimensional spatial unit and its adjacent regions; identify different regions accordingly; and adaptively adjust the initial unit model according to the region identification results to form the final unit partitioning result and construct the drilling data feature vector corresponding to each unit. The surrounding rock attribute identification and modeling module is configured to: input the feature vector of each unit's drilling data into the pre-constructed surrounding rock characteristic identification model, obtain the surrounding rock attribute identification results of each unit, and construct a three-dimensional geological model of the surrounding rock of the tunnel based on the surrounding rock attribute identification results.

[0083] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation process is the same, so it will not be repeated here.

[0084] Example 3 This embodiment also provides a computer device, including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor. When the processor executes the program, it completes the method described in Embodiment 1.

[0085] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0086] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.

[0087] In the implementation process, each step of the above method can be completed by the integrated logic circuits in the processor hardware or by software instructions.

[0088] The method in Embodiment 1 can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, a detailed description is not provided here.

[0089] Those skilled in the art will recognize that the units and algorithm steps described in connection with the various examples of this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0090] Example 4 This embodiment also provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the method described in Embodiment 1.

[0091] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data, characterized in that, include: Obtain the drilling space information and drilling data of multiple boreholes in front of the tunnel face, preprocess the drilling data, and extract the drilling data features corresponding to different depth sections of each borehole. Based on the borehole spatial information, the drilling data features of each borehole are mapped to a three-dimensional space to obtain a multi-hole drilling spatial dataset. Based on the multi-hole drilling spatial dataset, a three-dimensional continuous field of drilling data features is constructed using a spatial estimation method. Based on the three-dimensional continuous field of the drilling data features, an initial three-dimensional spatial unit model is established, and the spatial gradient, local variance and feature difference of the internal and adjacent regions of the three-dimensional spatial unit are calculated. Different regions are identified accordingly, and the initial unit model is adaptively adjusted according to the region identification results to form the final unit division result. The drilling data feature vector corresponding to each unit is constructed. The feature vectors of the drilling data of each unit are input into the pre-constructed surrounding rock characteristic identification model to obtain the surrounding rock attribute identification results of each unit. Based on the surrounding rock attribute identification results, a three-dimensional geological model of the surrounding rock of the tunnel is constructed.

2. The method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data according to claim 1, characterized in that, The drilling data includes drilling parameters and petrographic parameters; the drilling parameters include drilling control parameters and drilling response parameters; the drilling control parameters include one or more of drilling pressure and rotational speed, and the drilling response parameters include one or more of torque and drilling speed; the petrographic parameters include one or more of mineral composition, mineral content, and drill cuttings size; the drilling data features include drilling parameter features and petrographic features; the drilling parameter features include one or more of raw drilling parameter features, statistical features, fluctuation features, parameter coupling features, and rock-breaking energy derived features; the petrographic features include one or more of mineral composition, mineral content, and drill cuttings size.

3. The method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data according to claim 1, characterized in that, When mapping the drilling data features of each borehole to three-dimensional space based on the borehole spatial information, a local three-dimensional coordinate system of the roadway is established with the roadway axis as the first coordinate axis, the roadway transverse direction as the second coordinate axis, and the roadway vertical direction as the third coordinate axis. The three-dimensional spatial coordinates of the drilling data sampling points of each borehole are calculated based on the borehole opening coordinates, azimuth angle, dip angle, drilling depth, and borehole trajectory, and a multi-hole drilling spatial dataset is constructed accordingly.

4. The method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data according to claim 1, characterized in that, The spatial estimation method includes one or more of three-dimensional spatial interpolation methods or spatial inference methods.

5. The method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data according to claim 1, characterized in that, Constructing the three-dimensional continuous field of the drilling data features specifically includes: Analyze the distribution patterns, trends, and spatial variability of different drilling data characteristics in the axial, lateral, and vertical directions of the roadway, and determine the correlation distance, variability, main variability direction, and anisotropy ratio parameters in the three directions; Based on the three-dimensional spatial coordinates of the sampling points and the location to be estimated from the drilling data, the coordinate differences between the two in the axial, lateral and vertical directions of the tunnel are calculated. The coordinate differences in different directions are scaled according to the anisotropic scaling parameter to obtain the anisotropic distance between the sampling point and the location to be estimated. Based on the anisotropic distance and spatial correlation parameters, the estimated value of the drilling data characteristics of the inter-hole region is obtained by three-dimensional spatial interpolation or spatial inference methods. Based on the three-dimensional spatial coordinates of the location to be estimated, the estimated values ​​are spatially organized and correlated to form a three-dimensional continuous field corresponding to each drilling data feature.

6. The method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data according to claim 1, characterized in that, The adaptive adjustment of the initial three-dimensional spatial unit model and the construction of the feature vector of the drilling data specifically include: Based on the spatial range of the modeling area in front of the tunnel face, an initial three-dimensional spatial unit model is established according to the preset initial unit size; Based on the three-dimensional continuous field of each drilling data feature, the drilling data feature values ​​inside and adjacent regions of each initial three-dimensional spatial unit are obtained, and the spatial gradient, local variance and feature difference are calculated respectively. The spatial gradient, local variance, and feature difference of each spatial unit are compared with the corresponding preset judgment thresholds to identify stable regions, gradual regions, abrupt regions, and local abnormal regions in the drilling data. Following the rules of merging stable regions and subdividing abrupt regions, the initial three-dimensional spatial unit cells are adaptively refined or merged to form the unit cell division results. After completing the adaptive partitioning, for each type of drilling data feature, the representative value of the drilling data corresponding to each spatial unit is calculated using one of the following: mean, weighted mean, median, truncated mean, or quantile. The representative values ​​of different types of drilling data corresponding to each spatial unit are combined in a preset order to form the feature vector of drilling data corresponding to each unit.

7. The method for three-dimensional geological modeling of roadway surrounding rock driven by drilling data according to claim 1, characterized in that, The surrounding rock characteristic identification model adopts one of artificial neural networks, support vector machines, random forests, or gradient boosting trees, taking the feature vectors of drilling data as input and the corresponding measured values ​​of surrounding rock attributes as output; the surrounding rock attributes include one or more of rock strength, rock mass integrity index, surrounding rock quality index, and surrounding rock grade.

8. A drilling data-driven three-dimensional geological modeling system for roadway surrounding rock, characterized in that, include: The data acquisition and preprocessing module is configured to: acquire the borehole spatial information and drilling data of multiple boreholes in front of the tunnel face, preprocess the drilling data, and extract the drilling data features corresponding to different depth sections of each borehole. The three-dimensional spatial positioning module is configured to: map the drilling data features of each borehole to three-dimensional space according to the borehole spatial information to obtain a multi-hole drilling spatial dataset; The continuous field construction module is configured to: construct a three-dimensional continuous field of drilling data features based on the multi-hole drilling spatial dataset using a spatial estimation method; The adaptive partitioning and feature construction module is configured to: establish an initial three-dimensional spatial unit model based on the three-dimensional continuous field of the drilling data features; calculate the spatial gradient, local variance and feature difference of the three-dimensional spatial unit and its adjacent regions; identify different regions accordingly; and adaptively adjust the initial unit model according to the region identification results to form the final unit partitioning result; and construct the drilling data feature vector corresponding to each unit. The surrounding rock attribute identification and modeling module is configured to: input the feature vector of the drilling data of each unit into the pre-constructed surrounding rock characteristic identification model, obtain the surrounding rock attribute identification results of each unit, and construct a three-dimensional geological model of the surrounding rock of the tunnel based on the surrounding rock attribute identification results.

9. A computer device comprising a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the drilling data-driven three-dimensional geological modeling method for roadway surrounding rock as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the drilling data-driven three-dimensional geological modeling method for roadway surrounding rock as described in any one of claims 1-7.