Rock mass damage evaluation method and system based on improved electrical detection technology
By improving electrical resistivity tomography (EDT) technology, eliminating outliers, constructing a three-dimensional geological trend field and performing interpolation, and combining quantitative correlation models and density clustering algorithms, the problems of data interference and subjective bias in rock mass damage evaluation were solved, realizing automated and quantitative evaluation of rock mass damage.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NORTH CHINA UNIVERSITY OF SCIENCE AND TECHNOLOGY
- Filing Date
- 2026-03-25
- Publication Date
- 2026-06-19
Smart Images
Figure CN122241282A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of geotechnical engineering safety technology, and in particular to a method and system for evaluating rock mass damage based on improved electrical resistivity tomography (OTT) technology. Background Technology
[0002] In underground engineering projects such as mining, tunnel construction, and slope stabilization, the degree of rock mass fracturing and damage directly affects the stability of the project and construction safety. Therefore, accurately grasping the damage state and fracturing level of the rock mass is crucial for engineering design and risk control. Existing rock mass damage assessment methods mainly rely on geological surveys, borehole sampling, and sonic logging. However, these traditional methods have significant limitations: they are time-consuming, labor-intensive, and costly, and their detection range is usually very limited, often only reflecting the local geological conditions around the borehole, resulting in a "one-hole view" and making it difficult to achieve a continuous three-dimensional evaluation of the entire engineering area.
[0003] With the development of geophysical exploration technology, high-density electrical resistivity tomography (ED tomography) has been increasingly applied to rock mass structure detection due to its non-destructive and wide-coverage characteristics. However, existing high-density E tomography technologies still have many shortcomings when applied to the refined evaluation of rock mass damage. On the one hand, the field environment is complex, and E tomography data is easily interfered with by factors such as groundwater, metal facilities, and instrument noise. Traditional filtering methods are difficult to effectively distinguish noise from actual geological anomalies, resulting in difficulty in ensuring data quality. On the other hand, E tomography measuring points are usually laid out along the survey line, and the data distribution is discrete and uneven. Conventional spatial interpolation methods, such as kriging or inverse distance weighting, often ignore the anisotropy of geological structures when constructing three-dimensional models, resulting in distorted interpolation results and difficulty in accurately reconstructing the spatial distribution of rock resistivity.
[0004] Furthermore, existing technologies mostly remain at the level of qualitative interpretation of resistivity images, lacking a rigorous physical quantitative correlation model between resistivity values and rock mass damage variables, making it difficult to directly apply the evaluation results to subsequent mechanical calculations. At the same time, the classification of rock mass fracture levels often relies on the subjective experience of engineers, lacking objective clustering basis based on data characteristics, resulting in significant human bias in the evaluation results, making it difficult to meet the needs of modern engineering for refined and quantitative evaluation. Summary of the Invention
[0005] To address the problems existing in the prior art, this invention provides a rock mass damage evaluation method and system based on improved electrical resistivity tomography (ERT) technology. This method enables automated three-dimensional quantitative evaluation of rock mass damage status from ERT data, providing an objective basis for engineering stability analysis.
[0006] To achieve the above objectives, this invention provides a rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology, comprising: Apparent resistivity and video divergence values were collected from multiple measuring points in the target rock area, and outlier removal was performed on the multiple measuring points to obtain a clean dataset. Construct a three-dimensional geological trend field for the target rock area; The purified dataset is combined with the three-dimensional geological trend field, and a weighted external drift kriging algorithm is used for three-dimensional spatial interpolation to generate a three-dimensional data volume of rock resistivity. The damage variable values of each grid point in the three-dimensional data volume of rock resistivity are calculated using a pre-constructed quantitative correlation model to generate a three-dimensional damage variable data volume; the quantitative correlation model is constructed based on the differential effective medium theory and the extended Archie formula. Determine the spatial distribution feature vector of each grid point in the three-dimensional damage variable data volume; Density clustering algorithm is used to classify the spatial distribution feature vectors of each grid point, and the classification results are mapped to the rock mass fracture grade.
[0007] Optionally, outlier removal processing is performed on multiple measurement points, including: Measurement points whose apparent resistivity values exceed the preset resistivity threshold range are marked as Category I anomalies and removed. Using the current measurement point as the center, select neighboring measurement points within a preset spatial radius and depth window to form a local dataset; Calculate the mean and standard deviation of the apparent resistivity values in the local dataset; Measurement points in the local dataset whose apparent resistivity values deviate from the mean and exceed a preset multiple of the standard deviation are marked as second-type anomalies and removed; Establish a two-dimensional empirical distribution map of apparent resistivity and video divergence; Based on the two-dimensional empirical distribution map, measurement points that meet the conditions of having an apparent resistivity value lower than the first threshold and a video dispersion value lower than the second threshold are marked as third-type anomalies and removed.
[0008] Optionally, a three-dimensional geological trend field is constructed for the target rock area, including: Obtain borehole data, geological profiles, and lithological distribution information for the target rock area; Based on the borehole data, geological profiles, and lithological distribution information, the integrity of the rock mass at different spatial locations within the target rock area is determined. A three-dimensional scalar field is constructed based on the integrity of the rock mass to serve as a three-dimensional geological trend field.
[0009] Optionally, a weighted external drift kriging algorithm is used for three-dimensional spatial interpolation, including: The three-dimensional geological trend field is introduced as an external drift variable into the Kriging equations to constrain the geological trend of the interpolation results. A comprehensive weight is assigned to each measuring point in the clean dataset; the comprehensive weight is calculated based on the signal-to-noise ratio, local data density, and detection direction of the measuring point. Based on the Kriging equations with the introduction of external drift variables and the comprehensive weights of each measuring point, three-dimensional spatial interpolation is performed to generate a three-dimensional data volume of rock resistivity.
[0010] Optionally, the classification results can be mapped to rock mass fracture grades, including: Based on the damage variable values of all grid points within each cluster in the classification results, determine the damage statistical characteristic values corresponding to each cluster. Based on the preset rock mass fracture level classification rules, the damage statistical characteristic values of each cluster are compared with multiple level thresholds to determine the rock mass fracture level to which each cluster belongs.
[0011] Optionally, the method further includes: The three-dimensional data volume of rock resistivity, the three-dimensional data volume of damage variables, and the rock fracture level are visualized and output.
[0012] Optionally, the formula for calculating the spatial radius is: in, Indicates spatial radius, The average spacing between measuring points. The scaling factor is 1.5-2.5; The formula for calculating the depth window is: in, Represents the depth window. For the vertical resolution of electrical resistivity tomography, The proportionality coefficient is 1-2.
[0013] Optionally, the formula for calculating the comprehensive weight is: In the formula, For the first The overall weight of each measurement point For the first Signal-to-noise ratio weights for each measurement point For the first Data density weights for each measurement point For the first The detection direction weight of each measuring point The first weighting coefficient, This is the second weighting coefficient. The third weighting coefficient, .
[0014] Optionally, the expression for the quantized correlation model is: in, Represents rock mass damage variables. Resistivity at grid points The resistivity is the reference value for the intact rock mass. This represents the cementation index.
[0015] This invention also provides a rock mass damage evaluation system based on improved electrical resistivity tomography (OTT) technology, comprising: The data acquisition and preprocessing unit is used to acquire the apparent resistivity and video divergence values of multiple measuring points in the target rock area, and to perform outlier removal processing on the multiple measuring points to obtain a cleaned dataset. 3D modeling and interpolation unit, used for: Construct a three-dimensional geological trend field for the target rock area; The purified dataset is combined with the three-dimensional geological trend field, and a weighted external drift kriging algorithm is used for three-dimensional spatial interpolation to generate a three-dimensional data volume of rock resistivity. The quantitative damage evaluation unit is used to calculate the damage variable values of each grid point in the three-dimensional data volume of rock resistivity using a pre-constructed quantitative correlation model, and generate a three-dimensional damage variable data volume; the quantitative correlation model is constructed based on the differential effective medium theory and the extended Archie formula; Intelligent hierarchical unit, used for: Determine the spatial distribution feature vector of each grid point in the three-dimensional damage variable data volume; Density clustering algorithm is used to classify the spatial distribution feature vectors of each grid point, and the classification results are mapped to the rock mass fracture grade.
[0016] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: The rock mass damage evaluation method based on improved electrical resistivity detection technology provided by this invention effectively removes various interferences in complex field environments by collecting apparent resistivity and video dispersion values from multiple measuring points and eliminating outliers, thus obtaining a highly reliable clean dataset and solving the problem of weak anti-interference ability of existing methods. By constructing a three-dimensional geological trend field and combining it with the clean dataset, and using the weighted external drift kriging algorithm for interpolation, geological prior knowledge is incorporated into the modeling process, making the generated resistivity data volume conform to the anisotropic law of geological structure, overcoming the defects of insufficient accuracy and easy distortion of conventional interpolation methods. Based on the differential effective medium theory and the extended Archie formula, a quantitative correlation model is constructed to establish the physical relationship between resistivity and damage variables, realizing the direct quantitative conversion from electrical parameters to mechanical damage state, filling the gap between qualitative evaluation and mechanical calculation. By determining the spatial distribution feature vector of each grid point and using density clustering algorithm for classification mapping, the objective and automatic classification of rock mass fracture level is realized, avoiding the bias caused by the reliance on human subjective experience in existing methods. This invention automates the entire process from data acquisition to crushing level output, providing a reliable and efficient solution for engineering stability analysis and risk warning. Attached Figure Description
[0017] The above and other objects, features and advantages of the present invention will become more apparent from the more detailed description of exemplary embodiments of the invention in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same parts.
[0018] Figure 1 This is a schematic flowchart illustrating the rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology, as shown in an embodiment of the present invention. Figure 2 The following is a flowchart of the evaluation method and a schematic diagram of the data preprocessing logic as shown in the embodiments of the present invention; Figure 3 This is a schematic diagram illustrating the principle of the improved weighted external drift Kriging algorithm as shown in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating damage model calibration and DBSCAN clustering analysis in an embodiment of the present invention; Figure 5 This is a schematic diagram of the module structure of a rock mass damage evaluation system based on improved electrical resistivity tomography (OTT) technology, as shown in an embodiment of the present invention. Detailed Implementation
[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0020] Please see Figure 1 and Figure 2 , Figure 1 This is a schematic diagram of the process for evaluating rock mass damage based on improved electrical resistivity tomography (OTT) technology.
[0021] Rock mass damage assessment methods based on improved electrical resistivity tomography (OTT) techniques include: S101: Collect apparent resistivity and video dispersion values from multiple measuring points in the target rock area, and perform outlier removal processing on multiple measuring points to obtain a clean dataset.
[0022] In applications, high-density electrical resistivity survey lines can be laid out in the target rock area according to a pre-set survey network. Data acquisition is performed using devices such as the Wenner or Wenner-Schlumberger instruments to obtain the apparent resistivity and video dispersion values at each measuring point. Due to the complex on-site environment, the raw data often contains outliers caused by factors such as groundwater, metal facilities, poor electrode grounding, or instrument noise. If these outliers are not processed, they will directly affect the accuracy of subsequent 3D modeling and damage assessment. Therefore, outlier identification and removal are necessary to obtain a clean dataset with a high signal-to-noise ratio.
[0023] Specifically, the outlier removal process for multiple measurement points mentioned above includes: Measurement points whose apparent resistivity values exceed the preset resistivity threshold range are marked as Category I anomalies and removed. Using the current measurement point as the center, select neighboring measurement points within a preset spatial radius and depth window to form a local dataset; Calculate the mean and standard deviation of apparent resistivity values in a local dataset; Points in the local dataset whose apparent resistivity values deviate from the mean and exceed a preset multiple of the standard deviation are marked as Category II anomalies and removed. Establish a two-dimensional empirical distribution map of apparent resistivity and video divergence; Based on the two-dimensional empirical distribution map, measurement points that meet the conditions of having an apparent resistivity value lower than the first threshold and a video dispersion value lower than the second threshold are marked as third-type anomalies and removed.
[0024] See Figure 2In application, a three-level cascaded judgment logic can be used to systematically identify and remove outliers from the original data. First, a geological hard threshold filter is performed. Based on the regional geological data of the target rock area, the statistical results of borehole core resistivity tests, and the lithological distribution characteristics, a reasonable range of apparent resistivity corresponding to each lithology is pre-determined, i.e., a lower threshold is set. and upper limit threshold Will satisfy or The measured data points were marked as Category I anomalies and removed. For example, based on geological data of a mining area, the surrounding rock is mainly diorite. Combined with rock physics testing, a reasonable range for resistivity was set as follows: to Extreme points outside this range are removed. This geological hard threshold filter can quickly remove extreme points that significantly exceed the normal range of lithological variation, such as ultra-high resistivity anomalies caused by metal facilities or ultra-low resistivity anomalies caused by surface water.
[0025] Secondly, local dynamic statistical filtering is performed. Considering that geological bodies typically exhibit continuity within a local area, the resistivity value of the actual measuring point should be similar to that of its neighboring points. A spatial radius is set horizontally, centered on each measuring point to be detected. (For example (approximately twice the electrode distance), within the vertically set depth window. (For example A local dataset is constructed by selecting all neighboring measurement points within the three-dimensional neighborhood. The mean of the apparent resistivity values in this local dataset is calculated. with standard deviation If the preset conditions are met (e.g.) Then the corresponding measuring point Mark them as Category II anomalies and remove them; in this way, isolated jump points caused by poor electrode grounding, local random interference, etc. can be effectively identified.
[0026] The formula for calculating the spatial radius is: in, Indicates spatial radius, The average spacing between measuring points. The scaling factor is 1.5-2.5; The formula for calculating the depth window is: in, Represents the depth window. For the vertical resolution of electrical resistivity tomography, The proportionality coefficient is 1-2.
[0027] Finally, a consistency check of video divergence properties was performed. Video divergence is a parameter reflecting the induced polarization effect of the rock mass and has a significant response to water-bearing structures. The apparent resistivity and video divergence of the target rock area were established. The two-dimensional empirical distribution map, for the measurement points retained after the first two stages of filtering, satisfies... and If the anomaly is determined to be primarily caused by fissure water or aquifers, it is considered a water-induced disturbance rather than a genuine rock mass fracture signal. These anomalies are then classified as Category III anomalies and removed. This inspection effectively distinguishes between low-resistivity fracture zones and water-filled fissures, preserving rock mass anomaly information of engineering significance.
[0028] in, Represents resistivity. Take the lower percentile of the resistivity distribution (e.g., 10%). Indicates video dispersion. Take the lower percentile (e.g., 15%) of the video dispersion distribution.
[0029] S102: Construct a three-dimensional geological trend field for the target rock area.
[0030] In application, it is also necessary to construct a three-dimensional geological trend field that reflects the spatial variation law of the geological background and rock mass integrity of the target rock area. This trend field does not come directly from electrical resistivity tomography data, but is an independent scalar field established based on prior knowledge of regional geology. Its role is to provide geological constraints for subsequent weighted external drift kriging interpolation, so that the interpolation results are mathematically optimal while conforming to the distribution trend of actual geological structures, avoiding the distortion that may be caused by relying solely on discrete measurement point data.
[0031] For example, constructing a three-dimensional geological trend field for the target rock area includes: Obtain borehole data, geological profiles, and lithological distribution information for the target rock area; Based on borehole data, geological profile maps, and lithological distribution information, the integrity of the rock mass at different spatial locations within the target rock area is determined. A three-dimensional scalar field is constructed based on the integrity of the rock mass to serve as a three-dimensional geological trend field.
[0032] In the specific construction process, the first step is to collect various geological data of the target rock area, including borehole core logging data, geological profile maps, lithology distribution maps, and existing regional geological survey reports. By organizing and analyzing this data, information such as the spatial distribution of different lithologies within the rock area, the strike and occurrence of major faults or fracture zones, and the variation patterns of rock mass weathering can be obtained. Based on this, combined with engineering experience and geostatistical methods, a quantitative index reflecting the integrity of the rock mass is assigned to each spatial location within the rock area. For example, values can be assigned based on the RQD index of the borehole core, rock quality description, or lithology category; intact rock masses are assigned values close to 1, while fractured or severely weathered areas are assigned values close to 0. For areas without direct borehole control, interpolation or assignment is performed based on the strata extension trends inferred from geological profiles and the extent of fault influence, thus forming a three-dimensional gridded scalar field covering the entire target rock area, i.e., a three-dimensional geological trend field.
[0033] Each grid point in this three-dimensional geological trend field has a value, typically set between 0 and 1. A value closer to 1 indicates a more intact rock mass at that location, while a value closer to 0 indicates a more fragmented or weak rock mass. This scalar field can be either continuously changing or exhibit a stepped distribution based on lithological zoning.
[0034] S103: Combine the purified dataset with the three-dimensional geological trend field, and use the weighted external drift kriging algorithm to perform three-dimensional spatial interpolation to generate a three-dimensional data volume of rock resistivity.
[0035] Conventional Kriging interpolation methods typically rely solely on the spatial correlation of data points for estimation, making it difficult to effectively utilize prior geological knowledge. This can lead to discrepancies between the interpolation results and actual geological structures at critical locations such as faults and fracture zones, resulting in trend deviations or continuity distortions. To overcome this deficiency, this invention introduces a three-dimensional geological trend field as an external drift variable into the Kriging equations. This ensures that the interpolation process is constrained by geological laws, guaranteeing that the three-dimensional resistivity data volume of the rock mass conforms to the overall trend of lithological distribution and structural orientation.
[0036] Specifically, a weighted external drift kriging algorithm is used for 3D spatial interpolation, including: The three-dimensional geological trend field is introduced as an external drift variable into the Kriging equations to constrain the geological trend of the interpolation results. A comprehensive weight is assigned to each measurement point in the clean dataset; the comprehensive weight is calculated based on the signal-to-noise ratio, local data density, and detection direction of the measurement point. Based on the Kriging equations with the introduction of external drift variables and the comprehensive weights of each measuring point, three-dimensional spatial interpolation is performed to generate a three-dimensional data volume of rock resistivity.
[0037] See Figure 3In the application, the previously constructed three-dimensional geological trend field is first... External drift kriging is added as an external drift variable to the drift term of the kriging system. The basic principle of external drift kriging is to represent the trend of the regionalized variable as a linear combination of known external variables under the conditions of unbiased estimation and minimization of estimation variance, thereby using these external variables to guide the interpolation direction. Building upon this, to reflect the differences in data quality at different measurement points, this invention further purifies the dataset for each valid measurement point. Assign a comprehensive weight The comprehensive weight is not the weight coefficient in traditional kriging, but a priori weight used to adjust the influence of each measuring point during interpolation. The determination of the comprehensive weight considers three factors: the signal-to-noise ratio (SNR) of the measuring point, the local data density, and the probe direction. A higher SNR indicates less interference and more reliable data quality, thus requiring a larger weight. Higher local data density indicates denser information in the area, allowing for a reduction in the weight of individual points to avoid overfitting. The probe direction reflects the favorable position of the survey line relative to the geological structure; for example, survey lines perpendicular to the main structural trend often provide more critical structural information.
[0038] The formula for calculating the overall weight is: In the formula, For the first The overall weight of each measurement point For the first Signal-to-noise ratio weights for each measurement point For the first Data density weights for each measurement point For the first The detection direction weights of each measuring point; The first weighting coefficient, This is the second weighting coefficient. This is the third weighting coefficient; the three weighting coefficients can be flexibly set according to the actual situation, or fixed values based on experience can be used. For example, considering the decisive influence of the signal-to-noise ratio on the accuracy of results in rock mass exploration, it can be set to... , , To highlight the contribution weight of high signal-to-noise ratio data.
[0039] In the interpolation calculation process, the aforementioned comprehensive weights are combined with the external drift kriging method. On the one hand, the external drift variable guides the interpolation results to spatially follow known geological trends; on the other hand, when solving the kriging equations, the measurement point data are proportionally adjusted according to their comprehensive weights, or incorporated into the calculation of the variogram, thereby achieving differentiated processing of data of different quality. After the calculation by this improved algorithm, the optimal unbiased resistivity value can be obtained at each grid node of the target rock area, ultimately generating a continuous, uniform, and geologically sound three-dimensional resistivity data volume.
[0040] S104: Calculate the damage variable values of each grid point in the three-dimensional rock resistivity data volume using a pre-built quantitative correlation model to generate a three-dimensional damage variable data volume.
[0041] In existing technologies, qualitative interpretations are often based solely on resistivity levels, such as simply classifying low-resistivity regions as fracture zones. This approach lacks rigorous physical basis, makes it difficult to establish a quantitative relationship with rock mass mechanical parameters, and fails to meet the needs of engineering stability analysis. To overcome this limitation, this invention pre-constructs a quantitative correlation model based on differential effective medium theory and the extended Archie formula, establishing a mathematical relationship between resistivity and damage variables from the perspective of microscopic pore structure evolution.
[0042] The differential effective medium theory describes the asymptotic process of physical properties in multiphase composite materials as a function of component volume fraction, and is suitable for simulating resistivity changes caused by the initiation and propagation of microcracks in rocks. The extended Archie formula, based on the classic Archie formula, further considers the influence of non-uniform pore structure and crack development on the conductivity of rocks. Combining the two yields a quantitative correlation model of the following form: in, Represents rock mass damage variables. Resistivity at grid points The resistivity is the reference value for the intact rock mass. It is the cementation index, also known as the porosity index.
[0043] See Figure 4 Model parameters and The calibration can be obtained through indoor rock sample experiments. The calibration method is as follows: prepare a set of samples with known damage variables. Rock samples; resistivity of each rock sample was measured. ;Pair the data Substituting the parameters into the quantized correlation model, the optimal parameters are obtained by fitting using the nonlinear least squares method. .
[0044] Specifically, representative core samples can be collected from the target rock area. A set of rock samples with different degrees of damage can be prepared using artificial means or mechanical loading equipment, and the actual resistivity value of each sample can be measured. The experimental data (damage variables) for each set can then be... resistivity Substituting the above quantitative correlation model into the curve fitting using the nonlinear least squares method, the optimal parameters applicable to this target rock area can be solved. and Once calibrated, this quantization correlation model can be used for every grid point in the three-dimensional resistivity data volume.
[0045] In the specific calculation, all grid points in the three-dimensional rock resistivity data volume are traversed, and the resistivity value of each grid point is substituted into the calibrated quantitative correlation model to calculate the corresponding damage variable value. The damage variable values of all grid points are organized according to their original spatial locations, forming a three-dimensional damage variable data volume with the same grid structure as the resistivity data volume. Each value in this data volume quantitatively reflects the degree of damage to the rock mass at that location, realizing a direct conversion from geophysical exploration data to rock mass mechanical state parameters, and providing a quantitative input basis for the subsequent objective classification of rock mass fracture levels.
[0046] S105: Determine the spatial distribution feature vector of each grid point in the three-dimensional damage variable data volume.
[0047] Each grid point in the three-dimensional damage variable data volume corresponds to a numerical value that quantitatively describes the degree of rock mass damage. However, relying solely on the damage value of a single grid point is insufficient to fully characterize the spatial structural features of rock mass fracture, such as the scale, morphology, internal heterogeneity, and spatial relationships between different fractured regions. This information is crucial for objectively classifying the rock mass fracture grade. Therefore, it is necessary to determine a spatial distribution feature vector for each grid point in the three-dimensional damage variable data volume that comprehensively reflects its local damage characteristics and geometric properties.
[0048] Specifically, for each grid point in the three-dimensional damage variable data volume Calculate a five-dimensional feature vector .in, The damage variable value of this grid point itself directly reflects the degree of damage to the rock mass at that location; It is the local mean of the damage values of all grid points within a certain three-dimensional neighborhood centered at this point, used to characterize the overall degree of fragmentation in this local area; This is the local standard deviation of the damage values within the local neighborhood, used to reflect the dispersion or non-uniformity of the damage values within this region; The volume of the connected region to which the grid point belongs is determined by three-dimensional connectivity analysis, which identifies the connected region consisting of all grid points whose damage values exceed a certain threshold, and calculates the volume of the region to characterize the size of the fracture zone. The sphericity of the connected region is the ratio of the surface area of a sphere with the same volume as the region to the actual surface area of the region. It reflects the geometry of the fractured region. The closer the sphericity is to 1, the closer the region is to an equiaxed shape. The smaller the sphericity, the flatter or strip-like the region is. These features characterize the spatial distribution of rock mass damage from different perspectives. Combining them constitutes the spatial distribution feature vector of the grid point.
[0049] By assigning a multidimensional feature vector to each grid point, the original three-dimensional damage variable data volume is mapped into a feature space, where each grid point corresponds to a point. Subsequent cluster analysis will be based on these feature vectors, enabling grid points with similar local damage characteristics and geometric shapes to be automatically grouped into one category, thus laying the foundation for the objective classification of rock mass fracture levels.
[0050] S106: Density clustering algorithm is used to classify the spatial distribution feature vectors of each grid point, and the classification results are mapped to the rock mass fracture grade.
[0051] Specifically, the classification results are mapped to rock mass fracture grades, including: Based on the damage variable values of all grid points within each cluster in the classification results, determine the damage statistical characteristic values corresponding to each cluster. Based on the preset rock mass fracture level classification rules, the damage statistical characteristic values of each cluster are compared with multiple level thresholds to determine the rock mass fracture level to which each cluster belongs.
[0052] Each grid point in the three-dimensional damage variable data volume defines a five-dimensional spatial distribution feature vector. These feature vectors characterize the local features and geometric properties of rock mass damage from different perspectives. However, classifying the rock mass into engineering-significant fracture levels based on these feature vectors still requires an objective and automated classification method. Traditional methods often rely on engineers' subjective experience to set a single threshold, which is difficult to adapt to the multidimensionality and spatial variability of damage characteristics under complex geological conditions. To address this, this invention employs a density-based clustering algorithm to automatically classify the feature vectors of each grid point and maps the classification results to rock mass fracture levels, thereby achieving an objective classification of rock mass damage states.
[0053] In the application, the five-dimensional feature vectors of all grid points are first standardized. Because the components in the feature vectors have different dimensions and numerical ranges (e.g., damage values are between 0 and 1, while the volume of connected regions may span multiple orders of magnitude), direct clustering would lead to large numerical components dominating the clustering results. Standardization ensures that each dimension has a mean of 0 and a variance of 1, eliminating the influence of dimensions and ensuring that each feature has equal importance in the clustering process.
[0054] Then, the standardized set of feature vectors is used as input, such as Figure 4 The algorithm used is DBSCAN (Density-Based Spatial Clustering of Applications with Noise) for cluster analysis. The algorithm controls the density of clustering through two parameters: neighborhood radius and minimum number of neighborhood points. Appropriate parameter values are selected based on the distribution characteristics of the feature vectors, ensuring that grid points within the same fragmented region are grouped into the same cluster due to similar features, while different fragmented regions are effectively separated.
[0055] After clustering, each grid point is assigned a cluster label. Grid points with the same label form a cluster, representing a type of region with similar damage characteristics. For each cluster, the damage variable values of all grid points within the cluster are statistically analyzed, and the median is calculated as the damage statistical characteristic value of the cluster. Using the median instead of the mean can reduce the influence of a few extreme values and more robustly reflect the overall damage level of the cluster.
[0056] Finally, based on the preset rock mass fracture grade classification rules, the damage statistical characteristic values of each cluster are compared with multiple grade thresholds to determine the rock mass fracture grade to which each cluster belongs. Grade thresholds can be set according to engineering specifications or historical experience; for example, areas with damage variables less than 0.2 are classified as intact rock mass, 0.2~0.5 as slightly damaged, and greater than 0.5 as severely fractured. All grid points within each cluster are assigned the same fracture grade, thus transforming the three-dimensional damage variable data volume into a rock mass fracture grade distribution map with engineering semantics.
[0057] For example, three main clusters were identified: Cluster A (intact rock mass): mean damage Cluster B (minor damage): average damage value 0.2-0.5, distributed in areas far from the subsidence zone; Cluster C (severely broken): average damage value... Furthermore, the local variance is large, concentrated in the core area of the goaf roof. The above clustering results are mapped to Class I (stable), Class II (basically stable), and Class III (unstable) rock masses, respectively.
[0058] In one embodiment, the above method further includes: The three-dimensional data volume of rock resistivity, the three-dimensional data volume of damage variables, and the rock fracture level are visualized and output.
[0059] After classifying the rock mass fracture levels, the generated data volumes and classification results can be visualized. Specifically, the obtained three-dimensional rock resistivity data volume, three-dimensional damage variable data volume, and rock mass fracture levels are converted into intuitive visual images using graphical rendering technology. This allows engineers to quickly and accurately grasp the internal structural state and damage distribution patterns of the rock mass. The visualization output can include at least one of three-dimensional contour maps, two-dimensional slice maps, and isosurface maps. Three-dimensional contour maps are used to display the continuous spatial changes of resistivity or damage variables; two-dimensional slice maps are used to observe detailed features at any depth or on any cross-section; and isosurface maps are used to clearly delineate the boundaries of the fractured areas corresponding to specific damage thresholds.
[0060] In practical applications, dedicated visualization modules can be developed based on open-source graphics libraries such as VTK or PyVista. After loading the 3D data volume, high-resolution images can be generated through color mapping, transparency adjustment, lighting rendering, and other means.
[0061] Corresponding to the aforementioned application function implementation method embodiments, the present invention also provides a rock mass damage evaluation system based on improved electrical resistivity tomography (OTT) technology and corresponding embodiments.
[0062] Please see Figure 5 , Figure 5 This is a schematic diagram of the module structure of a rock mass damage evaluation system based on improved electrical resistivity tomography (OTT) technology.
[0063] A rock mass damage assessment system based on improved electrical resistivity tomography (OTT) technology includes: The data acquisition and preprocessing unit 51 is used to acquire the apparent resistivity and video divergence values of multiple measuring points in the target rock area, and to perform outlier removal processing on multiple measuring points to obtain a clean dataset. 3D modeling and interpolation unit 52, used for: Construct a three-dimensional geological trend field for the target rock area; The purified dataset is combined with a three-dimensional geological trend field, and a weighted external drift kriging algorithm is used for three-dimensional spatial interpolation to generate a three-dimensional data volume of rock resistivity. The quantitative damage evaluation unit 53 is used to calculate the damage variable values of each grid point in the three-dimensional data volume of rock resistivity using a pre-constructed quantitative correlation model, and generate a three-dimensional damage variable data volume; the quantitative correlation model is constructed based on the differential effective medium theory and the extended Archie formula; Intelligent hierarchical unit 54, used for: Determine the spatial distribution feature vector of each grid point in the three-dimensional damage variable data volume; Density clustering algorithm is used to classify the spatial distribution feature vectors of each grid point, and the classification results are mapped to the rock mass fracture grade.
[0064] In one embodiment, in terms of outlier removal processing for multiple measurement points, the data acquisition and preprocessing unit 51 is specifically used for: Measurement points whose apparent resistivity values exceed the preset resistivity threshold range are marked as Category I anomalies and removed. Using the current measurement point as the center, select neighboring measurement points within a preset spatial radius and depth window to form a local dataset; Calculate the mean and standard deviation of apparent resistivity values in a local dataset; Points in the local dataset whose apparent resistivity values deviate from the mean and exceed a preset multiple of the standard deviation are marked as Category II anomalies and removed. Establish a two-dimensional empirical distribution map of apparent resistivity and video divergence; Based on the two-dimensional empirical distribution map, measurement points that meet the conditions of having an apparent resistivity value lower than the first threshold and a video dispersion value lower than the second threshold are marked as third-type anomalies and removed.
[0065] In one embodiment, in constructing a three-dimensional geological trend field of a target rock area, the three-dimensional modeling and interpolation unit 52 is specifically used for: Obtain borehole data, geological profiles, and lithological distribution information for the target rock area; Based on borehole data, geological profile maps, and lithological distribution information, the integrity of the rock mass at different spatial locations within the target rock area is determined. A three-dimensional scalar field is constructed based on the integrity of the rock mass to serve as a three-dimensional geological trend field.
[0066] In one embodiment, in the context of using a weighted external drift kriging algorithm for 3D spatial interpolation, the 3D modeling and interpolation unit 52 is specifically used for: The three-dimensional geological trend field is introduced as an external drift variable into the Kriging equations to constrain the geological trend of the interpolation results. A comprehensive weight is assigned to each measurement point in the clean dataset; the comprehensive weight is calculated based on the signal-to-noise ratio, local data density, and detection direction of the measurement point. Based on the Kriging equations with the introduction of external drift variables and the comprehensive weights of each measuring point, three-dimensional spatial interpolation is performed to generate a three-dimensional data volume of rock resistivity.
[0067] In one embodiment, the intelligent grading unit 54 is specifically used for mapping the classification results to rock mass fracturing grades: Based on the damage variable values of all grid points within each cluster in the classification results, determine the damage statistical characteristic values corresponding to each cluster. Based on the preset rock mass fracture level classification rules, the damage statistical characteristic values of each cluster are compared with multiple level thresholds to determine the rock mass fracture level to which each cluster belongs.
[0068] In one embodiment, the system further includes: The visualization output unit is used to visualize and output the three-dimensional data volume of rock resistivity, the three-dimensional damage variable data volume, and the rock fracture level.
[0069] Regarding the system in the above embodiments, the specific manner in which each unit module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated further here.
[0070] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.
Claims
1. A method for evaluating rock mass damage based on improved electrical resistivity tomography (OTT) technology, characterized in that, include: Apparent resistivity and video divergence values were collected from multiple measuring points in the target rock area, and outlier removal was performed on the multiple measuring points to obtain a clean dataset. Construct a three-dimensional geological trend field for the target rock area; The purified dataset is combined with the three-dimensional geological trend field, and a weighted external drift kriging algorithm is used for three-dimensional spatial interpolation to generate a three-dimensional data volume of rock resistivity. The damage variable values of each grid point in the three-dimensional data volume of rock resistivity are calculated using a pre-constructed quantitative correlation model to generate a three-dimensional damage variable data volume; the quantitative correlation model is constructed based on the differential effective medium theory and the extended Archie formula. Determine the spatial distribution feature vector of each grid point in the three-dimensional damage variable data volume; Density clustering algorithm is used to classify the spatial distribution feature vectors of each grid point, and the classification results are mapped to the rock mass fracture grade.
2. The rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 1, characterized in that, Outlier removal processing is performed on multiple measurement points, including: Measurement points whose apparent resistivity values exceed the preset resistivity threshold range are marked as Category I anomalies and removed. Using the current measurement point as the center, select neighboring measurement points within a preset spatial radius and depth window to form a local dataset; Calculate the mean and standard deviation of the apparent resistivity values in the local dataset; Measurement points in the local dataset whose apparent resistivity values deviate from the mean and exceed a preset multiple of the standard deviation are marked as second-type anomalies and removed; Establish a two-dimensional empirical distribution map of apparent resistivity and video divergence; Based on the two-dimensional empirical distribution map, measurement points that meet the conditions of having an apparent resistivity value lower than the first threshold and a video dispersion value lower than the second threshold are marked as third-type anomalies and removed.
3. The rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 2, characterized in that, Constructing a three-dimensional geological trend field for the target rock area includes: Obtain borehole data, geological profiles, and lithological distribution information for the target rock area; Based on the borehole data, geological profiles, and lithological distribution information, the integrity of the rock mass at different spatial locations within the target rock area is determined. A three-dimensional scalar field is constructed based on the integrity of the rock mass to serve as a three-dimensional geological trend field.
4. The rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 1, characterized in that, Three-dimensional spatial interpolation is performed using a weighted external drift kriging algorithm, including: The three-dimensional geological trend field is introduced as an external drift variable into the Kriging equations to constrain the geological trend of the interpolation results. A comprehensive weight is assigned to each measuring point in the clean dataset; the comprehensive weight is calculated based on the signal-to-noise ratio, local data density, and detection direction of the measuring point. Based on the Kriging equations with the introduction of external drift variables and the comprehensive weights of each measuring point, three-dimensional spatial interpolation is performed to generate a three-dimensional data volume of rock resistivity.
5. A rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 1, characterized in that, The classification results are mapped to rock mass fracture grades, including: Based on the damage variable values of all grid points within each cluster in the classification results, determine the damage statistical characteristic values corresponding to each cluster. Based on the preset rock mass fracture level classification rules, the damage statistical characteristic values of each cluster are compared with multiple level thresholds to determine the rock mass fracture level to which each cluster belongs.
6. The rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 1, characterized in that, The method further includes: The three-dimensional data volume of rock resistivity, the three-dimensional data volume of damage variables, and the rock fracture level are visualized and output.
7. A rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 2, characterized in that, The formula for calculating the spatial radius is: in, Indicates spatial radius, The average spacing between measuring points. The scaling factor is 1.5-2.5; The formula for calculating the depth window is: in, Represents the depth window. For the vertical resolution of electrical resistivity tomography, The proportionality coefficient is 1-2.
8. A rock mass damage evaluation method based on improved electrical resistivity tomography (OTT) technology according to claim 4, characterized in that, The formula for calculating the overall weight is as follows: In the formula, For the first The overall weight of each measurement point For the first Signal-to-noise ratio weights for each measurement point For the first Data density weights for each measurement point For the first The detection direction weight of each measuring point The first weighting coefficient, This is the second weighting coefficient. The third weighting coefficient, .
9. A method for evaluating rock mass damage based on improved electrical resistivity tomography (OTT) technology according to claim 1, characterized in that, The expression for the quantitative correlation model is: in, Represents rock mass damage variables. Resistivity at grid points The resistivity is the reference value for the intact rock mass. This represents the cementation index.
10. A rock mass damage evaluation system based on improved electrical resistivity tomography (OTT) technology, characterized in that, include: The data acquisition and preprocessing unit is used to acquire the apparent resistivity and video divergence values of multiple measuring points in the target rock area, and to perform outlier removal processing on the multiple measuring points to obtain a cleaned dataset. 3D modeling and interpolation unit, used for: Construct a three-dimensional geological trend field for the target rock area; The purified dataset is combined with the three-dimensional geological trend field, and a weighted external drift kriging algorithm is used for three-dimensional spatial interpolation to generate a three-dimensional data volume of rock resistivity. The quantitative damage evaluation unit is used to calculate the damage variable values of each grid point in the three-dimensional data volume of rock resistivity using a pre-constructed quantitative correlation model, and generate a three-dimensional damage variable data volume; the quantitative correlation model is constructed based on the differential effective medium theory and the extended Archie formula; Intelligent hierarchical unit, used for: Determine the spatial distribution feature vector of each grid point in the three-dimensional damage variable data volume; Density clustering algorithm is used to classify the spatial distribution feature vectors of each grid point, and the classification results are mapped to the rock mass fracture grade.