Intelligent analysis method and system for geological structure of mining area based on three-dimensional visual modeling

Abstract

The present application relates to the technical field of three-dimensional geological visualization modeling, and discloses a mine geological structure intelligent analysis method and system based on three-dimensional visualization modeling, which comprises the following steps: acquiring continuous drilling process parameters along a drill hole, establishing a parameter fingerprint library to identify the causes of parameter changes, identifying effective physical beacon points capable of representing real geological discontinuity, and then generating a geological reality guide field in three-dimensional space using the beacon points as a source, and imposing constraints on the weight across potential geological boundaries when performing spatial interpolation. The present application converts long-neglected drilling process data into effective geometric constraints for the three-dimensional modeling kernel, changes the model construction from pure mathematical filling to a form closer to the engineering reality guided by the physical process, and ensures the reliability of the constraint source through an internal anti-fake mechanism, thereby improving the structural reality of the final geological model.

Description

Technical Field

[0001] This invention relates to an intelligent analysis method and system for geological structures in mining areas based on three-dimensional visualization modeling, belonging to the field of three-dimensional geological visualization modeling technology. Background Technology

[0002] In the field of 3D visualization modeling technology, especially for the analysis and construction of geological structures in mining areas, the technical approach that is widely followed and relied upon in the industry is to construct a continuous 3D geological entity model based on discrete geological sample point information revealed by engineering, such as borehole cores and tunnel logging, and to use spatial interpolation algorithms. This approach provides fundamental technical support for the visualization of underground resources and the estimation of reserves.

[0003] However, this widely accepted technical approach is based on an implicit technical compromise in principle. In order to achieve mathematical continuity, it cognitively treats the vast area between sample points as a pure information vacuum and completely discards the continuous process data carrying rock mechanics response information generated in the physical process of acquiring these sample points, such as the drilling rate and torque changes of the drilling rig. These physical quantities, which directly reflect the real-time interaction between the drill bit and the rock, are usually only archived as engineering logs in the existing modeling process, and their inherent geological information value is not utilized by the modeling engine.

[0004] Technological advancements in this field have long focused on finding more optimized spatial interpolation algorithms. From Kriging to emerging machine learning models, their essence is to find a mathematically smoother or more accurate filling function within the existing information vacuum framework. However, this has not addressed the root of the problem: the core of 3D modeling, in its design philosophy, is naturally geared towards static, discrete geometric and attribute data, lacking an intrinsic mechanism to receive, understand, and apply dynamic, continuous procedural physical data to impose constraints on the core geometric interpolation process that originate from physical reality. Specifically, existing technologies have the following shortcomings: 1. The core driving force of model construction is purely mathematical algorithms, and the form of the results lacks a direct intrinsic connection with the physical causes of geological structures. Especially in areas with complex tectonic changes, the model's realism depends on the density of sample points. 2. Continuous physical process data obtained at low cost during exploration is systematically ignored, and this data is not fully utilized. The model lacks a physical basis for verifying and constraining its geometric form. Therefore, the technical problem to be solved by this invention is how to enable the 3D modeling kernel to break through the limitation of only being able to process discrete static samples, and establish a system that can transform the continuous physical process data generated during exploration into an effective constraint on spatial interpolation algorithms, so that model construction can shift from pure mathematical filling to physical process-guided morphological generation that is closer to the actual geological reality. Summary of the Invention

[0005] This invention provides an intelligent analysis method and system for geological structures in mining areas based on three-dimensional visualization modeling. Its main purpose is to solve the problem that existing three-dimensional modeling methods cannot utilize continuous physical data from the exploration process, resulting in model construction relying only on discrete samples and mathematical filling, thus lacking physical reality constraints.

[0006] To achieve the above objectives, this invention provides an intelligent analysis method for geological structures in mining areas based on three-dimensional visualization modeling, comprising the following steps:

[0007] Step a: At the sample point location with known geological information, extract the local feature vector of the drilling process parameters corresponding to each sample point, and associate the local feature vector with the geological event type at that location to construct a parameter fingerprint database for identifying the cause of the signal.

[0008] Step b: Obtain the drilling rate and torque that change continuously along the borehole depth, calculate the local rate of change of drilling rate and torque within the determined sliding depth window, and mark the locations where the local rate of change exceeds the threshold as potential physical beacon points.

[0009] Step c: Extract the local feature vectors of drilling process parameters at each potential physical beacon point, and identify them according to the parameter fingerprint database established in step a. Only potential physical beacon points whose parameter features are associated with real geological discontinuities are used as valid physical beacon points.

[0010] Step d: Using all valid physical beacon points as the source, a scalar geological authenticity guide field is generated in the three-dimensional modeling space through distance transformation;

[0011] Step e: When performing three-dimensional spatial interpolation using discrete geological sample points, the weight of any sample point acting on the interpolated point is adjusted punitively based on the field value of the geological authenticity guiding field on the path connecting the sample point and the interpolated point.

[0012] Preferably, the penalized adjustment of the weights in step e is specifically as follows: the original weights are multiplied by a penalty factor to obtain the adjusted weights. The penalty factor is uniquely determined by the path integral of the geological authenticity guiding field value on the path connecting the sample point and the interpolated point, and the penalty factor is a monotonically decreasing function of the path integral value.

[0013] Preferably, the threshold used to identify potential physical beacon points in step b is spatially adaptively adjusted; the spatial adaptive adjustment includes: calculating the overall variance of drilling process parameters in each borehole as a regional geological contrast index, performing spatial interpolation based on the regional geological contrast index of all boreholes to generate a threshold modulation field, and multiplying a global reference threshold with the field value of the threshold modulation field at the corresponding position to obtain a dynamic threshold used for judgment at that position.

[0014] Preferably, the local feature vector in steps a and c is a multi-dimensional vector containing the mean, variance, and normalized gradient of the drilling process parameters within the depth window at the corresponding location. The identification of the parameter fingerprint database is achieved by calculating the vector distance between the local feature vector to be identified and the local feature vectors already stored in the database, and classifying them into the geological event type with the smallest distance.

[0015] Preferably, the field value of any point in the geological authenticity guidance field in step d is defined as the Euclidean distance from that point to the nearest effective physical beacon point in space.

[0016] Preferably, the method further includes performing a borehole trajectory deviation correction step between step c and step d. The correction step includes: identifying, based on the parameter fingerprint database, a reference beacon point corresponding to one or more laterally distributed stable geological marker layers among all valid physical beacon points; fitting the theoretical reference surface of the geological marker layer based on the design coordinates of the reference beacon point; calculating the deviation between the design depth of each reference beacon point and the corresponding depth of the theoretical reference surface; interpolating to generate a depth deviation field; and correcting the spatial coordinates of the valid physical beacon points based on the depth deviation field.

[0017] Preferably, the method further includes the following steps: extracting drilling process parameters within each borehole segment defined by adjacent valid physical beacon points identified in step c, calculating statistical characteristic values ​​of drilling process parameters within each borehole segment, and performing three-dimensional spatial interpolation based on the statistical characteristic values. Within the geological structure model, independently of the generation of the geological authenticity guiding field, an additional continuous attribute field characterizing the differences in physical properties within the rock mass is generated.

[0018] Preferably, the method further includes performing the following steps in parallel: calculating an index E characterizing drilling efficiency in real time based on the drilling rate and torque, wherein... ROP is the drilling rate, and T is the torque. The time series of the drilling efficiency index E is low-pass filtered to extract its long-term trend. When the negative slope of the long-term trend exceeds a wear judgment threshold determined based on historical data statistics, the drill bit is judged to have entered the wear state, and the drilling rate and torque obtained in step b are compensated and corrected according to the state.

[0019] Preferably, after step c, a step of identifying tectonic hotspot regions is performed. The identification step includes: based on the three-dimensional spatial location set of all valid physical beacon points, using a kernel density estimation algorithm, generating a beacon point spatial density field in the three-dimensional modeling space, and identifying regions in the beacon point spatial density field with a field value higher than a certain density threshold as tectonic hotspot regions; the punitive adjustment in step e is based entirely on the geological authenticity guidance field, and does not depend on the lithology or grade attribute values ​​of the geological sample points themselves.

[0020] A smart analysis system for geological structures in mining areas based on 3D visualization modeling includes:

[0021] The parameter fingerprint database construction module is used to extract the local feature vectors of drilling process parameters corresponding to each sample point with known geological information, and associate the local feature vectors with the geological event type at that location to construct a parameter fingerprint database for identifying the cause of signals.

[0022] The potential beacon point identification module is used to obtain the drilling rate and torque that change continuously along the borehole depth, calculate the local rate of change of drilling rate and torque within a defined sliding depth window, and mark the locations where the local rate of change exceeds the threshold as potential physical beacon points.

[0023] The beacon point identification module is used to extract the local feature vectors of drilling process parameters at each potential physical beacon point, and to identify them based on the parameter fingerprint database. Only potential physical beacon points whose parameter features are associated with real geological discontinuities are used as valid physical beacon points.

[0024] The guiding field generation module is used to generate a scalar geologically accurate guiding field in the three-dimensional modeling space by distance transformation, using all valid physical beacon points as the source.

[0025] The weight adjustment module is used to perform penalized adjustments on the weight of any sample point acting on the interpolated point when performing three-dimensional spatial interpolation using discrete geological sample points, based on the field value of the geological authenticity guidance field on the path connecting the sample point and the interpolated point.

[0026] Compared with the prior art, the beneficial effects of the present invention are:

[0027] 1. This invention establishes a novel working method for geological modeling. Instead of directly interpolating discrete geological sample points, it identifies physical beacon points characterizing geological discontinuities by utilizing the local variation features of drilling process parameters continuously acquired along the borehole path before interpolation. These beacon points are then used as the source to generate a scalar geological realism guiding field in three-dimensional space. During subsequent execution of any conventional spatial interpolation algorithm, this guiding field serves as a fundamental spatial constraint. Based on the guiding field values ​​between the interpolation point and the sample points, the interpolation weights are dynamically adjusted. Thus, the boundary of the geological model is no longer a purely mathematical fitting result determined by the spatial location of the sample points, but rather incorporates physical process information during the construction process, reflecting a boundary that reflects the discontinuity of rock mass mechanical properties.

[0028] 2. This invention further establishes an inherent verification mechanism for the source of the guiding field. After identifying potential physical beacon points, it does not immediately use them for the generation of the guiding field. Instead, it first extracts the corresponding drilling process parameter features at known geological sample point locations and associates them with the geological or engineering event types at those locations. This constructs a parameter fingerprint database within the modeling system. Any identified physical beacon point is adopted as a valid source only if its own parameter features match the records in the fingerprint database that characterize real geological discontinuities. This design prevents the modeling process from mistakenly interpreting engineering artifacts during drilling as geological structures, thus ensuring that the structural authenticity of the final generated guiding field and its constrained geological model is supported by the data source level.

[0029] 3. This invention also utilizes the drilling process parameters within each borehole segment defined by physical beacon points. By calculating their statistical characteristic values ​​and performing independent three-dimensional spatial interpolation based on these statistical characteristic values, a novel continuous attribute field characterizing the differences in physical properties within the rock mass is generated within the geological structure model. This process operates in parallel with the process of constructing the geological structure boundary based on the guiding field. This allows the same drilling process data source to have its drastically changing parts used to constrain the macroscopic structural boundary of the model, while its relatively stable changing parts are used to characterize the microscopic physical property texture within the model. The final result is no longer a static lithology model with only geometric shape, but an engineering geological object with a richer information dimension that simultaneously carries structural boundaries and the distribution of internal physical properties. Attached Figure Description

[0030] Figure 1 This is a flowchart of the three-dimensional geological modeling method based on physical beacons and guiding fields of the present invention;

[0031] Figure 2 This is a schematic diagram of the overall process of the present invention from field drilling to engineering application;

[0032] Figure 3 This is a diagram of the architecture of the multi-terminal collaborative intelligent geological structure analysis system of the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0034] This invention provides an intelligent analysis method and system for geological structures in mining areas based on 3D visualization modeling. Its overall operation is designed as a sequentially executed and logically coupled data processing pipeline. This pipeline begins with the acquisition and characterization of raw data from the drilling process, followed by signal discrimination and spatial field construction. Ultimately, it transforms physical process information into geometric constraints for a 3D spatial interpolation algorithm. Specifically, the method includes a parameter fingerprint database construction step, a potential beacon point identification step, a beacon point discrimination step, a guiding field generation step, and an interpolation weight adjustment step. The output of the parameter fingerprint database construction step provides the decision-making basis for the beacon point discrimination step, while the set of effective physical beacon points purified by the beacon point discrimination step serves as the sole source input for the guiding field generation step. Finally, the geologically accurate guiding field constructed by the guiding field generation step imposes decisive constraints on the standard spatial interpolation process in the interpolation weight adjustment step.

[0035] In a typical application scenario, such as geological modeling for deep ore body exploration, a common technical challenge is that relying solely on discrete borehole core samples for spatial interpolation makes it difficult to accurately characterize geological discontinuities with significant differences in mechanical properties, such as faults and dikes. This is because the interpolation algorithm treats the region between sample points as a mathematical space with continuously changing physical properties. To address this challenge, the first step of this invention is to extract local feature vectors of drilling process parameters at sample point locations with known geological information, such as borehole depths where core logging has been completed. These vectors are then correlated with the geological event types at that location to construct a parameter fingerprint database for identifying signal origins. Specifically, the local feature vector contains the mean, variance, and normalized gradient of the drilling process parameters within a preset depth window at the corresponding location. The system employs a multidimensional vector model. For example, within a depth window centered on a core sample point and 0.5 meters wide, it collects the corresponding drilling rate and torque sequences, calculates their mean, variance, and normalized gradient, forming a six-dimensional feature vector. This vector is then bound and stored with a label representing a real geological discontinuity event—the granite-surrounding rock interface—as determined by the core logging. Conversely, if the engineering log records engineering artifacts such as drill bit clogging within this depth window, the calculated corresponding feature vector is bound with the engineering artifact label. By performing this operation on all known sample points, the system internally establishes a query database—a parameter fingerprint database—that maps the form of specific parameter signals to their physical origins. This parameter fingerprint database, constructed by this procedure, provides a basis for subsequently identifying the true geological significance of parameter changes at any location.

[0036] When extracting local feature vectors, the sliding depth window size is determined by performing a window size optimization procedure on the calibration borehole data. This procedure traverses a preset size range, such as 0.2 meters to 2.0 meters, with a step size of 0.1 meters. For each size value, the Davies-Bouldin index of the feature vector clusters of various geological event types in the parameter fingerprint database is calculated. This index is a statistic that measures the ratio of inter-cluster separation to intra-cluster compactness. The window size that minimizes this index value is ultimately selected as the operating parameter for subsequent analysis. When performing beacon point identification, the system calculates the minimum Euclidean distance between the vector to be identified and all vectors in the database and compares it with a maximum classification distance threshold. This threshold is determined by the minimum distance between all known event feature vector clusters in the database. The large intrinsic radius is obtained by multiplying by a coefficient of 3. If the minimum distance is greater than this threshold, the potential beacon point is marked as pending and stored in an independent database for manual review, without being classified into any existing geological event type. When the modeling data comes from multiple drilling rigs, a cross-rig data scale consistency calibration procedure is performed before identifying potential physical beacons. This procedure first designates one or more lithologically homogeneous and laterally stable strata as standard strata in the regional geological data. Then, it extracts the statistical mean of the torque and drilling rate parameter sequences of each rig when traversing the standard strata, designates one rig as the benchmark rig, and then calculates the ratio of the parameter mean of each other rig within the standard strata to the parameter mean of the benchmark rig, obtaining the rig's specific correction factor C. i In subsequent data processing, all raw drilling process parameters P from device i raw All will be processed through operation P calibrated =P raw ·C i The transformation is performed to unify all multi-source heterogeneous data onto the same response benchmark.

[0037] Furthermore, in the 3D modeling process, the system first acquires the raw time-series data of drilling rate and torque that continuously vary along the depth of all simulated boreholes to be built. Within a sliding depth window of a defined size, it calculates the local rate of change of drilling rate and torque. When the local rate of change of any parameter exceeds a preset threshold, that depth location is marked as a potential physical beacon point. Given the spatial heterogeneity of rock mass mechanical properties in different geological regions, a globally fixed threshold may be too sensitive in areas of strong contrast and too insensitive in areas of weak contrast. Therefore, the threshold used to identify potential physical beacon points is configured for spatial adaptive adjustment, and its adjustment procedure is as follows: First, calculate the local time-series data of each... Within the vertical borehole, parameters such as torque are considered, and their overall variance is calculated across the entire borehole depth. This variance is then used as a regional geological contrast index characterizing the differences in the overall mechanical properties of the strata traversed by the borehole. Subsequently, using this index value from all boreholes as discrete samples, a continuous threshold modulation field is generated across the entire 3D modeling space using a standard spatial interpolation algorithm, such as inverse distance weighted interpolation. Finally, when determining whether the local rate of change at any location exceeds the limit, the dynamic threshold is obtained by multiplying a global reference threshold by the field value at that location in the threshold modulation field. For example, if the global reference threshold is set to 1.0, and the modulation field value at a certain location is 1.5 (strong contrast zone), then the threshold is determined. The value is automatically increased to 1.5; conversely, if the field value at another location is 0.7 (a weak contrast zone), the judgment threshold is automatically reduced to 0.7. In this way, the system can automatically adjust its sensitivity for identifying geological discontinuities based on differences in the local geological environment, thereby obtaining a set of potential physical beacon points. However, the signal origin of the potential physical beacon point set obtained at this stage is still uncertain, and may include engineering artifacts caused by non-geological factors such as drill bit wear or improper operation. To ensure the reliability of the signal source ultimately used for geometric constraints, the system initiates a beacon point identification step, extracting the local feature vector of the drilling process parameters at each potential physical beacon point, and then, based on the parameters established in the above steps... The fingerprint database is used for causal identification. This identification process calculates the vector distance (e.g., Euclidean distance) between the local feature vector to be identified and each feature vector already stored in the database. Based on the nearest neighbor principle, it classifies the feature vectors into the geological event types with the smallest distance. Only potential physical beacon points whose parametric features are associated with real geological discontinuities in the parameter fingerprint database are adopted as valid physical beacon points, while the rest are filtered out. Through this inherent anti-falsification mechanism, the modeling process avoids misinterpreting engineering artifacts during drilling as geological structures, thus ensuring that the structural authenticity of the final generated guiding field and its constrained geological model is supported by the data source.

[0038] After obtaining a verified set of all valid physical beacon points with high credibility in both spatial location and geological significance, the guiding field generation module uses this set as a source to generate a scalar geological authenticity guiding field in the 3D modeling space through distance transformation. Specifically, the field value of any point in this guiding field is uniquely defined as the Euclidean distance from that point to the nearest valid physical beacon point in space. This is a standard distance field that can be efficiently calculated by mature algorithms such as the fast traversal method. This guiding field, in the form of a continuous scalar function, depicts the possibility of geological discontinuities in the entire 3D space. The lower the field value, the closer the point is to a verified real geological boundary, and vice versa. This geological authenticity guiding field provides a foundation for subsequently correcting the weight function of the spatial interpolation algorithm. The spatial constraint field originates from the physical process. Ultimately, when using discrete geological sample points, such as lithology or grade data from borehole cores, to perform three-dimensional spatial interpolation to construct the final geological entity model, the weight adjustment module, based on the generated geological authenticity guiding field, applies a penalty adjustment to the original weights of any sample point acting on the interpolated point. This penalty adjustment is entirely based on the field value of the geological authenticity guiding field along the path connecting the sample point and the interpolated point, and does not depend on the lithology or grade attribute values ​​of the geological sample point itself. The specific adjustment procedure is as follows: the original weights are multiplied by a penalty factor to obtain the adjusted weights. The penalty factor is uniquely determined by the path integral of the geological authenticity guiding field value along the path connecting the sample point and the interpolated point, and the penalty factor is a monotonically decreasing function of the path integral value. For example, it can be achieved by using... The form is adjusted, where W adj For the adjusted weights, W orig Here are the original weights, and k is the penalty coefficient, ∫ path (x)dx is the line integral of the guiding field value G(x) along the path connecting the sample point and the interpolated point. When the connecting path crosses the region with a low guiding field value, i.e. the potential geological boundary, the integral value increases and the penalty factor decreases, thereby weakening the cross-boundary influence of the sample point. In this way, the boundary of the geological body model is no longer a mathematical fitting result determined purely by the spatial position of the sample point, but rather an objective expression that reflects the discontinuity of the rock mass mechanical properties, which has already incorporated physical process information during the construction process.

[0039] It should be noted that this invention can also execute in parallel a procedure for diagnosing the health status of the drill string during data acquisition, thereby further improving the reliability of the input data. This procedure calculates an index E characterizing drilling efficiency in real time based on the acquired drilling rate (ROP) and torque (T), where E = ROP / T. By performing low-pass filtering on the time series of the drilling efficiency index E to extract its long-term trend, when the negative slope of this long-term trend exceeds a wear threshold determined based on historical data statistics, the system determines that the drill string has entered a wear state and compensates and corrects the currently acquired drilling rate and torque according to this state to eliminate wear. In addition to the interference of systematic parameter drift caused by drill bit wear on the identification of geological discontinuities, after obtaining an effective set of physical beacon points, the present invention can further perform a step of identifying structural hotspot areas. This step is based on the three-dimensional spatial location set of all effective physical beacon points, and uses a kernel density estimation algorithm to generate a beacon point spatial density field in the three-dimensional modeling space. Areas with a field value higher than a preset density threshold in this density field are identified as structural hotspot areas. These areas represent the most developed and complex geological structures and can provide high-priority decision guidance for subsequent in-depth exploration or engineering design.

[0040] Furthermore, this invention can also utilize drilling process parameters within each borehole segment defined by effective physical beacon points. By calculating the statistical characteristic values ​​of the parameters within each segment, such as the variance or coefficient of variation of torque, and performing independent three-dimensional spatial interpolation based on these statistical characteristic values, an additional continuous attribute field characterizing the differences in physical properties within the rock mass is generated within the geological structure model. This process operates in parallel with the process of constructing the geological structure boundary based on the guiding field. This allows the same drilling process data source to have its drastically changing parts used to constrain the geometric boundary of the model, while its relatively stable changing parts are used to characterize the physical property texture within the model. Ultimately, an engineering geological object with richer information dimensions is obtained, which simultaneously carries both structural boundaries and the distribution of internal physical properties. Meanwhile, to address the common problem of borehole trajectory deviation in practical engineering, this invention may also include a borehole trajectory deviation correction step. This step first identifies, based on a parameter fingerprint database, a reference beacon point corresponding to one or more laterally distributed stable geological marker layers among all valid physical beacon points. Subsequently, a theoretical reference surface of the geological marker layer is fitted based on the design coordinates of these reference beacon points, and the deviation between the design depth of each reference beacon point and the corresponding depth of the theoretical reference surface is calculated. Then, a three-dimensional depth deviation field is generated by interpolation. Finally, the spatial coordinates of the valid physical beacon points are corrected based on the depth deviation field, thereby eliminating the spatial positioning error of beacon points caused by borehole curvature and ensuring the geometric accuracy of the guiding field and the final model.

[0041] Example 1: In a task of constructing a three-dimensional geological model for slope stability analysis and deep resource exploration in an open-pit to underground mining area, engineers constructed a rock mass model based on core sample data obtained from more than 100 boreholes using the standard inverse distance weighted interpolation method. The visualization results of the model show that the contact boundary between the hard granite mass and the weak alteration zone, which are key factors affecting slope stability, is a smooth mathematically fitted surface. However, during the subsequent construction of the tunnel, small-scale geological disasters such as landslides and water inrushes were encountered near the smooth boundary indicated by the model. The actual geological conditions revealed that there is a fractured structural transition zone several meters wide that is not represented by the model.

[0042] To address the engineering risks arising from the limitations of the modeling method, the project team applied the technical solution of this invention. First, the system acquired the drilling rate and torque parameters that continuously varied along the depth during the drilling process of the aforementioned boreholes. On these continuous data streams, the system calculated the local rate of change of the parameters through a sliding depth window, identifying multiple potential physical beacon points whose local rate of change exceeded the dynamic threshold. These points were spatially densely distributed in the contact area between hard granite and weak alteration zones, and also sporadically distributed at other depths in some boreholes. Some of these points were caused by drill bit clogging in weak strata. At this point, the system's parameter fingerprint database was constructed and... The beacon point identification mechanism begins to operate. The parameter fingerprint database has been pre-established at a small number of known geological and engineering event points by extracting the local feature vectors of the corresponding drilling process parameters. The system then extracts the local feature vector at each potential physical beacon point and compares it with the parameter fingerprint database. Potential physical beacon points that match engineering artifact fingerprints such as smeared drills are filtered out, while those that match lithological interface fingerprints are confirmed as valid physical beacon points. This collaborative operation produces a set of valid physical beacon points that have had engineering artifact signals filtered out and spatially delineate the contours of the fractured structural transition zone.

[0043] Furthermore, using this effective set of physical beacon points as the source, the system generates a geological authenticity guiding field across the entire 3D modeling space through distance transformation. When performing spatial interpolation using the original discrete core sample points, this guiding field penalizes the interpolation weights. For any interpolation point located within hard granite, when the interpolation algorithm attempts to reference the influence of a sample point located within a weak alteration zone, the path connecting these two points will inevitably traverse the region of low field value in the geological authenticity guiding field, i.e., the location of the fractured tectonic transition zone, according to the weight adjustment procedure. The path integral value of the guiding field increases, leading to a decrease in the penalty factor and a significant reduction in the weight of the sample points in the alteration zone. This mechanism resolves the contradiction between maintaining global smoothness in spatial interpolation and the requirement for boundary sharpening in geological structures. It does not change the mathematical core of the interpolation algorithm, but rather changes the way the algorithm operates in space through the guidance of external physical information. This allows the algorithm to maintain continuous filling within lithological units, while exhibiting discontinuity when crossing boundaries defined by the guiding field. In the final output 3D geological model, the original smooth mathematically fitted surface is replaced by an independent geological entity of a fractured structural transition zone with a clear spatial range and geometric shape. The spatial location of this entity shows spatial consistency with the areas where collapses and water inrushes actually occurred during subsequent tunnel excavation, thus providing a decision-making basis consistent with engineering reality for slope stability analysis and support design.

[0044] Furthermore, after the geometric boundaries of the aforementioned fractured structural transition zone geological entity were determined, the system further analyzed the differences in physical properties within the rock mass. Specifically, the system extracted parameter segments of all boreholes traversing the interior of the fractured structural transition zone entity and calculated the variance of the drill bit torque within each segment. These variance values ​​distributed along the borehole path were used as new discrete sample points. Through three-dimensional spatial interpolation, a continuous torque variance attribute field was independently generated within the geometric model of the entity. In the final visualization model, this attribute field was used to control the color rendering within the model. Areas with high torque variance values ​​were rendered in red, and areas with low torque variance values ​​were rendered in blue. This visualization result allows engineers not only to identify the spatial extent of the transition zone but also to intuitively identify the core section with the highest degree of fracture within it. This provides a more refined zoning basis for the subsequent design of key parameters such as anchor density and depth when performing rock anchor support.

[0045] Example 2: To objectively verify the effectiveness of the technical solution of this invention in improving the geometric accuracy of geological models, the following numerical simulation experiment was designed and executed. The purpose of the experiment was to quantitatively compare the spatial position error of the model constructed using the method of this invention with that constructed using traditional methods when reconstructing known geological discontinuities. The experimental platform was based on a three-dimensional numerical computing environment, in which a digital geological reference model with a size of 500m×500m×500m was constructed. This model contains two types of lithology: one is the background surrounding rock, and the other is an inclined plate-shaped hard rock intrusion defined by the plane equation z=0.5x+100 that runs through the entire model. The upper and lower interfaces of this intrusion are the real geological boundaries that need to be accurately reconstructed in this experiment. In this geological reference model, 20 virtual boreholes with random spatial positions but uniform overall distribution were arranged. The system performed data simulation sampling along the path of each virtual borehole to generate two sets of datasets: the first set is discrete geological data. The first set of data consists of two sets of data: sample point data, which is lithological information recorded every 1.0 m along the borehole path; and continuous drilling process parameter data. The generation rules are as follows: in the background surrounding rock, the baseline values ​​for drilling rate (ROP) and torque (T) are set to 15 m / h and 2000 Nm, respectively; in hard rock intrusions, they are set to 5 m / h and 6000 Nm, respectively. Gaussian noise with a mean of zero and a standard deviation of 5% of the baseline value is superimposed on the two baseline value sequences to simulate data fluctuations under real-world conditions. Based on these two sets of data, two experimental groups were set up: a control group and an experimental group. The control group used only discrete geological sample point data and employed the standard inverse distance weighting method for three-dimensional spatial interpolation to construct a geological model. The experimental group first used continuous drilling process parameter data and, according to the procedures in the specific implementation method, generated a geological authenticity guiding field. Subsequently, when interpolating the same discrete geological sample points using the same inverse distance weighting method, this guiding field was applied for weight penalty adjustment.

[0046] After generating corresponding 3D geological models for both experimental groups, 10 spatially fixed detection points were pre-set on the real geological boundary plane. The normal position error of the corresponding boundary constructed by the two models at these detection points was calculated, which is the shortest distance from the model boundary point to the real boundary plane. The experimental results showed that the average boundary position error of the model constructed by the control group was 7.93m, and the error of some detection points reached 11.2m. The generated boundary shape showed obvious deformation and displacement in areas far from the borehole data constraints. In contrast, the average boundary position error of the model constructed by the experimental group was 0.39m, and the error of all detection points did not exceed 0.6m. The generated boundary shape showed a clear linear characteristic consistent with the real boundary plane. The underlying mechanism of this data difference is that... The interpolation process in the control group relies solely on the spatial distance of sample points, leading to overly smooth mathematical fitting in sparse sample areas and thus deviating from the true boundary. In contrast, the low-field-value guiding field region generated by the drastic changes in drilling process parameters in the experimental group provides a spatial constraint for the interpolation algorithm. The weight adjustment mechanism weakens the influence of sample points crossing this region, resulting in a high gradient change in the interpolation result at the boundary, thus maintaining consistency with the position of the true geological boundary. The results of this numerical simulation experiment confirm that, under the same discrete sample data conditions, by applying the method of this invention and constraining the spatial interpolation by introducing a geological authenticity guiding field generated from drilling process data, the constructed geological model has a significantly improved geometric accuracy in reproducing geological discontinuities compared to traditional interpolation methods.

[0047] Example 3: This example combines Figures 1 to 3 This section describes the intelligent analysis method and system for geological structures in mining areas based on 3D visualization modeling, such as... Figure 1As shown, this graph begins with the raw data input, namely, acquiring the drilling rate and torque that continuously change along the borehole. This data stream is used for two purposes: firstly, for online diagnosis of drill string health and parameter correction to compensate for systematic parameter drift caused by drill string wear; secondly, it enters the potential physical beacon point identification module, which completes the identification by calculating the local rate of change of drilling parameters and marking locations exceeding thresholds. Simultaneously, an independent parameter fingerprint database construction process correlates known geological events with the local feature vectors of drilling parameters. This fingerprint database provides a decision-making basis for subsequent beacon point identification steps, namely, identifying artifacts based on the fingerprint database and retaining only genuine geological discontinuities. These identified effective physical beacons... Punctuation points are further differentiated in their uses: firstly, they serve as sources for generating geological authenticity guidance fields through spatial distance transformation; secondly, they are used for identifying structural hotspot areas. This identification employs a kernel density estimation algorithm to identify dense areas of beacon points and outputs a structural hotspot map, providing decision-making guidance for intensified exploration or engineering design. Finally, the geological authenticity guidance field and discrete geological sample points, such as static geometric and attribute data from borehole cores and tunnel logging, jointly enter the interpolation weight adjustment step. Based on the guidance field, a penalized adjustment is applied to the weights across geological boundaries, thereby generating a highly realistic three-dimensional geological model that integrates the structural morphology and internal attributes constrained by physical processes.

[0048] like Figure 2 As shown in the figure, the horizontal axis represents the borehole depth in meters (m), and the vertical axis represents the parameter values. The drilling rate (ROP) is indicated by a solid line in m / h, and the torque (T) is indicated by a dashed line in T (×100 Nm). It can be seen from the figure that the drilling rate ROP and torque T remain relatively stable in most borehole depth ranges. However, near depths of approximately 30 m and 66 m, the drilling rate ROP decreases significantly, while the torque T increases sharply. This drastic change in parameters indicates that the drill bit has encountered a geological body with significantly different properties from the surrounding rock. These points of dramatic change are precisely the objects that the potential physical beacon point identification step aims to capture and mark.

[0049] like Figure 3As shown, the system architecture consists of four parts. The drilling process data acquisition unit inside the on-site drilling equipment sends real-time data streams to the edge computing node. This node includes a drill string health status diagnosis module and a real-time data preprocessing module, and uploads the preprocessed data to the cloud-based data and model server. This server is the core of the system, and it deploys an intelligent analysis and computing engine and a unified data storage. The intelligent analysis and computing engine includes a parameter fingerprint library construction module, a potential / valid beacon point identification and screening module, a guide field generation and weight adjustment module, and a structural hotspot area identification module. The unified data storage manages the original drilling database, parameter fingerprint library, and 3D geological model library. Finally, the engineer's workstation interacts with the server through its onboard 3D visualization and analysis software and task management and calibration interface, via model requests and task assignments, to execute specific analysis tasks and obtain model results.

[0050] Example 4: Before applying the method of this invention to a new geological modeling project in a mining area, a systematic calibration and verification procedure needs to be performed to ensure that the core algorithm parameters match the specific geological and engineering conditions of the target mining area. This procedure aims to deterministically quantify the global benchmark threshold used to identify potential physical beacon points and the penalty factor k used to adjust the interpolation weights. Simultaneously, it verifies the ability of the parameter fingerprint database to distinguish between common drilling artifacts and real geological discontinuities in the mining area. To this end, engineers first select 3 to 5 boreholes in the mining area that have been drilled and have complete core geological logging and drilling logs. As calibration samples, the selection criteria for these boreholes require that they must traverse representative major lithological units and their contact zones within the mining area, and that the engineering logs record at least two typical drilling process artifact events, such as drill bit clogging or parameter anomalies caused by drill bit wear. After the initial state definition is completed, the calibration procedure is initiated. The system first collects all continuous drilling process parameters of these calibration boreholes and performs depth alignment with the core geological logging and engineering logs. As a result, the system obtains a set of parameter signal samples containing hundreds of clearly labeled events, where each sample is marked as a true geological discontinuity or a specific type of engineering artifact.

[0051] Based on this sample set, the system constructs and verifies the parameter fingerprint database. This involves extracting local feature vectors at the location of each event label and analyzing the distribution and inter-class separability of these vectors in the multi-dimensional feature space to confirm that the selected features can effectively separate signals from different causes. Furthermore, to determine the global baseline threshold, the system calculates the local rate of change for all sample points labeled as true geological discontinuities, forming one distribution. It also calculates the local rate of change for all other sample points, including background noise and engineering artifacts, forming another distribution. The global baseline threshold is set in the overlapping region of the two distributions, ensuring that the recognition rate of true discontinuities is greater than 95% and the false positive rate for engineering artifacts is less than 5%. The calibration process for the penalty factor k in the weight adjustment procedure is as follows: Within a local 3D space containing known complex boundaries, a geologically accurate guiding field is generated using established effective physical beacon points. Spatial interpolation modeling is then repeatedly performed with a series of incremental k values ​​ranging from 0.1 to 10.0. By geometrically comparing the model boundary generated for each k value with the true boundary of the calibration sample, the k value that minimizes the average geometric deviation between the model boundary and the true boundary is ultimately selected as the working parameter for the mining area. After completing the above series of calibration steps, a parameter fingerprint library specifically for the newly established mining area, as well as the quantified global benchmark threshold and penalty factor k, are solidified into a system configuration file. This provides a validated algorithmic foundation adapted to local working conditions for all subsequent intelligent geological structure analysis tasks within the mining area.

[0052] Example 5: In another implementation scenario of the method of the present invention, in order to eliminate the influence of borehole trajectory deviation, a common problem in deep hole drilling, on the geometric accuracy of the final model, the system performs a spatial position calibration procedure based on geological marker layers before generating the geological authenticity guidance field. This procedure first pre-sets parameter fingerprints of one or more geological marker layers that are horizontally stable in the mining area based on regional geological data. Then, the system uses the constructed parameter fingerprint library to automatically identify the reference beacon points that match the fingerprint of the geological marker layer among all identified valid physical beacon points. Based on the design three-dimensional coordinates of these reference beacon points, the system generates the theoretical reference surface of the geological marker layer through a surface fitting algorithm. Subsequently, the system calculates the deviation between the design depth of each reference beacon point and the theoretical reference surface at the corresponding plane position depth, and constructs a three-dimensional depth deviation field by spatial interpolation of all deviation values. Finally, the system compensates and corrects the spatial coordinates, especially the depth coordinates, of all valid physical beacon points based on this depth deviation field, thereby obtaining a set of spatially corrected beacon points for subsequent generation of the guidance field.

[0053] After obtaining the set of valid physical beacon points with spatial location correction, the system can also perform a task of identifying and classifying structurally complex areas in parallel. This task uses the three-dimensional spatial locations of all corrected valid physical beacon points as input data and employs a kernel density estimation algorithm to generate a spatial density field of beacon points in the entire three-dimensional modeling space. The field value of any point in this density field represents the density of valid physical beacon points within a unit volume around that point. Furthermore, by setting a density threshold, areas with field values ​​in the density field higher than the threshold are identified as structural hotspots. These areas intuitively reflect the spatial range where geological structures are most developed and lithological changes are most frequent. The identification results can be used as an independent data layer and overlaid on the final three-dimensional geological structure model to provide decision support for subsequent exploration engineering deployment or mining design.

[0054] Example 6: In a continuous deep drilling operation scenario, when the drill bit passes through a thick homogeneous granite body, its wear condition gradually changes due to long-term continuous cutting. This leads to a slow and systematic drift in the collected drilling rate and torque parameters, which is unrelated to the lithological changes. To address the potential artifacts caused by changes in the state of the data acquisition tool, the method of this invention initiates a set of online diagnostic and parameter calibration procedures for the health status of the drill bit in parallel before performing potential physical beacon point identification.

[0055] The procedure first calculates an index E characterizing drilling efficiency based on the real-time acquired drilling rate (ROP) and torque (T), where E = ROP / T. The time series of this index is fed into a low-pass filter to filter out short-term fluctuations caused by minor differences in the internal structure of the rock, thereby extracting its long-term trend reflecting drill bit wear. The system monitors the negative slope of this long-term trend. When the absolute value of the negative slope continuously exceeds a wear judgment threshold set based on historical drill bit replacement data, the system determines that the drill bit has entered a wear state. Once in the wear state, a correction module is activated. This module calculates a dynamic compensation coefficient based on the decrease in the current drilling efficiency index E relative to the initial healthy state baseline value, and uses this coefficient to amplify or reduce the subsequently input raw drilling rate and torque data. Through this online correction mechanism, the systematic deviation introduced by drill bit wear is removed from the source data stream, so that the subsequent physical beacon point identification steps can be performed based on a calibrated data foundation, avoiding the misinterpretation of the equipment wear process as a gradual geological structure.

[0056] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for intelligent analysis of geological structures of a mining area based on three-dimensional visualization modeling, characterized in that, The method comprises the following steps: Step a: extracting local feature vectors of drilling process parameters corresponding to each sample point at a sample point position with known geological information, and associating the local feature vectors with the type of geological event at the position to construct a parameter fingerprint library for identifying signal causes; Step b: obtaining drilling rates and torques that vary continuously along the depth of the borehole, calculating the local change rates of the drilling rates and torques within a determined sliding depth window, and marking a position as a potential physical beacon point if the local change rate exceeds a threshold value; Step c: extracting local feature vectors of drilling process parameters at each potential physical beacon point, and identifying them according to the parameter fingerprint library established in step a, and only using potential physical beacon points whose parameter features are associated with true geological discontinuity events as effective physical beacon points; Step d: taking all effective physical beacon points as sources, generating a scalar geological reality guide field in a three-dimensional modeling space through distance transformation; Step e: when interpolating in a three-dimensional space using discrete geological sample points, the weight acting on the interpolated point for each sample point is adjusted in a punitive manner according to the field value of the geological reality guide field on the path connecting the sample point and the interpolated point.

2. The method according to claim 1, wherein, The punitive adjustment of the weight in step e is specifically: multiplying the original weight by a penalty factor to obtain the adjusted weight, the penalty factor is uniquely determined by the path integral of the field value of the geological reality guide field on the path connecting the sample point and the interpolated point, and the penalty factor is a monotonically decreasing function of the path integral value. 3.The mine geological structure intelligent analysis method based on three-dimensional visualization modeling of claim 1, wherein, The threshold value for identifying potential physical beacon points in step b is adaptively adjusted in space; the spatial adaptive adjustment includes: calculating the overall variance of drilling process parameters in each borehole as a regional geological contrast index, generating a threshold modulation field based on the regional geological contrast index of all boreholes, multiplying a global reference threshold value by the field value of the threshold modulation field at the corresponding position to obtain a dynamic threshold value for judgment at the position.

4. The method according to claim 1, wherein, The local feature vectors in steps a and c are a multi-dimensional vector containing the mean, variance and normalized gradient of drilling process parameters within the depth window at the corresponding position, and the identification of the parameter fingerprint library is achieved by calculating the vector distance between the local feature vector to be identified and the local feature vectors stored in the library, and classifying it to the geological event type with the smallest distance.

5. The method according to claim 1, wherein, The field value of any point in the geological reality guide field in step d is defined as the Euclidean distance from the point to the nearest effective physical beacon point in space.

6. The method according to claim 1, wherein, Further comprising a correction step of borehole trajectory deviation between steps c and d, the correction step includes: identifying reference beacon points corresponding to one or more laterally stable geological marker layers from all effective physical beacon points according to the parameter fingerprint library, fitting a theoretical reference surface of the geological marker layer based on the design coordinates of the reference beacon points, calculating the deviation of the design depth of each reference beacon point from the corresponding depth of the theoretical reference surface, and interpolating to generate a depth deviation field, and correcting the spatial coordinates of the effective physical beacon points according to the depth deviation field.

7. The method according to claim 1, wherein, Further comprising the following steps: Extracting drilling process parameters within each borehole segment defined by the adjacent valid physical beacon points identified in step c, calculating statistical characteristic values of drilling process parameters within each borehole segment, and performing three-dimensional spatial interpolation based on the statistical characteristic values, an additional continuous attribute field representing the difference in physical properties within the rock mass is generated inside the geological structure model independently of the generation of the geological reality guide field.

8. The method according to claim 1, wherein, Also included is the step of performing in parallel the following steps: calculating in real time an indicator E representing drilling efficiency based on the rate of penetration and the torque, wherein ROP is the rate of penetration, T is the torque; low-pass filtering the time series of the drilling efficiency indicator E to extract its long-term trend, and determining that the drilling tool is in a worn state when the negative slope of the long-term trend exceeds a wear determination threshold determined based on historical data, and compensating the rate of penetration and the torque obtained in step b according to this state.

9. The method according to claim 1, wherein, Further comprising a step of identifying a hot spot area after step c, the identifying step comprising: based on the three-dimensional spatial position set of all valid physical beacon points, generating a beacon point spatial density field in the three-dimensional modeling space using a kernel density estimation algorithm, and identifying the area in the beacon point spatial density field where the field value is higher than a determined density threshold as the hot spot area.

10. A three-dimensional visualization modeling-based intelligent analysis system for geological structures of a mining area, characterized in that, Comprise: A parameter fingerprint library construction module for extracting local feature vectors of drilling process parameters corresponding to each sample point at a known geological information sample point position, and associating the local feature vectors with the geological event type at the position to construct a parameter fingerprint library for signal origin discrimination; A potential beacon point identification module for obtaining continuously varying drilling rate and torque along the borehole depth, calculating the local variation rate of drilling rate and torque within a determined sliding depth window, and marking the position where the local variation rate exceeds the threshold as a potential physical beacon point; A beacon point discrimination module for extracting local feature vectors of drilling process parameters at each potential physical beacon point, and discriminating according to the parameter fingerprint library, only using potential physical beacon points whose parameter characteristics are associated with real geological discontinuity events as valid physical beacon points; A guide field generation module for taking all valid physical beacon points as sources to generate a scalar geological reality guide field in the three-dimensional modeling space through distance transformation; A weight adjustment module for adjusting the weight of each sample point acting on the interpolated point when performing three-dimensional spatial interpolation using discrete geological sample points, according to the field value of the geological reality guide field on the path connecting the sample point and the interpolated point.

Images