Rock structure modeling method, device, equipment and medium based on drilling data

Abstract

The present application discloses a method, device, equipment and medium for modeling rock mass structure based on borehole data, which relates to the field of geotechnical engineering technology. Through the effective use of borehole data and a reconstruction algorithm that takes into account geological variability characteristics, even when there are inherent defects in the data, a model that more accurately reflects the actual structure of the rock mass can be established based on these data, achieving high-precision fitting with the borehole data while achieving high-precision retention and reasonable extrapolation of regional rock mass structure variability characteristics, effectively solving the problem of insufficient model accuracy in areas with sparse borehole data using traditional methods, and achieving high-precision modeling of rock mass structure on a global scale. Moreover, the spatial variability of borehole data is fully considered when constructing the rock mass structure model. Whether in areas with relatively simple geological conditions or under complex geological conditions, a rock mass structure model that conforms to the actual situation can be flexibly generated and fitted based on the characteristics of the borehole data, and has strong generalization capabilities.

Description

Technical Field

[0001] The present application relates to the field of geotechnical engineering technology, and in particular to a method, device, equipment and medium for rock structure modeling based on drilling data. Background Art

[0002] With the advancement of geotechnical engineering, the refined characterization and modeling of rock mass structure are crucial. In engineering practice, discrete data acquired through geological exploration (including drill cores, geophysical profiles, and surface mapping information) needs to be systematically integrated through spatial algorithms. However, existing spatial interpolation algorithms fail to capture the spatial variability of various structural surfaces within the rock mass. This makes it difficult to accurately reproduce the spatial distribution of rock mass structure during 3D reconstruction, resulting in inaccurate rock mass mass zoning and regional assignment of mechanical parameters.

[0003] Traditional rock mass structure modeling methods rely primarily on geological exploration data, often employing interpolation methods such as the inverse distance power method and kriging to construct block models. However, in complex geological conditions, such as karst areas and fault zones, the detection accuracy of geophysical exploration methods such as geological radar can be severely affected, resulting in incomplete and inaccurate data. Existing rock mass structure modeling methods also suffer from issues such as a lack of data at the far end of the model, large inference biases, insufficient interpretability, and limited generalization capabilities, making it difficult to fully and accurately reflect the true three-dimensional rock mass structure characteristics and geological correlations. Summary of the Invention

[0004] In view of this, the present application provides a rock structure modeling method, device, equipment and medium based on drilling data, the main purpose of which is to solve problems such as lack of remote data of the model, large inference deviation, insufficient interpretability and limited generalization ability.

[0005] According to a first aspect of the present application, a rock structure modeling method based on drilling data is provided, the method comprising:

[0006] Obtaining original drilling data, modeling the original drilling data after uniform discretization processing, and obtaining an initial drilling attribute model;

[0007] calculating a plurality of distance-weighted coefficients of variation for characterizing the spatial variability of the initial model of borehole attributes, and generating a plurality of heterogeneous block model samples using a bubbling method based on the initial model of borehole attributes and the plurality of distance-weighted coefficients of variation;

[0008] Taking the initial model of the borehole attributes as a constraint condition, the weight coefficient matrix of the multiple heterogeneous block model samples is solved by the least squares method, and the multiple heterogeneous block model samples are weightedly superimposed and fitted using the weight coefficient matrix to generate a target rock mass structural attribute model.

[0009] According to a second aspect of the present application, a rock structure modeling device based on drilling data is provided, the device comprising:

[0010] An initial model building module is used to obtain original drilling data, model the original drilling data after uniform discretization processing, and obtain an initial drilling attribute model;

[0011] a heterogeneous sample generation module, configured to calculate a plurality of distance-weighted variation coefficients for characterizing the spatial variability of the initial borehole attribute model, and to generate a plurality of heterogeneous block model samples using a bubbling method based on the initial borehole attribute model and the plurality of distance-weighted variation coefficients;

[0012] The weight fitting modeling module is used to solve the weight coefficient matrix of the multiple heterogeneous block model samples by the least squares method with the initial model of the borehole attributes as a constraint condition, and use the weight coefficient matrix to perform weighted superposition fitting on the multiple heterogeneous block model samples to generate a target rock mass structure attribute model.

[0013] According to a third aspect of the present application, a device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the methods described in the first aspect when executing the computer program.

[0014] According to a fourth aspect of the present application, a medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the methods in the first aspect are implemented.

[0015] By means of the above technical solution, the technical solution provided by the embodiment of the present application has at least the following advantages:

[0016] The present application provides a method, device, equipment and medium for rock structure modeling based on borehole data. The present application obtains original borehole data, models the original borehole data after uniform discretization processing, obtains an initial model of borehole properties, and then calculates multiple distance-weighted variation coefficients used to characterize the spatial variability of the initial model of borehole properties. Based on the initial model of borehole properties and the multiple distance-weighted variation coefficients, a bubbling method is used to generate multiple heterogeneous block model samples. Then, with the initial model of borehole properties as a constraint condition, the weight coefficient matrix of the multiple heterogeneous block model samples is solved by the least squares method. The weight coefficient matrix is ​​used to perform weighted superposition fitting on the multiple heterogeneous block model samples to generate a target rock structure property model. By effectively utilizing borehole data and reconstructing algorithms that consider geological variability, a model that more accurately reflects the actual rock mass structure can be established based on this data, even when the data has inherent flaws. This achieves high-precision fitting with the borehole data while also retaining and reasonably extrapolating the regional rock mass structure variability. This effectively addresses the problem of insufficient model accuracy in areas with sparse borehole data, effectively achieving high-precision modeling of rock mass structure on a global scale. Furthermore, by fully considering the spatial variability of borehole data when constructing the rock mass structure model, it can better adapt to the rock mass structure characteristics under different geological conditions. Whether in areas with relatively simple geological conditions or under complex geological conditions, it can flexibly generate and fit a rock mass structure model that conforms to the actual situation based on the characteristics of the borehole data, demonstrating strong generalization capabilities.

[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present application. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0019] Figure 1 A schematic flow chart of a method for rock mass structure modeling based on drilling data provided in an embodiment of the present application is shown;

[0020] Figure 2 A schematic flow chart of another method for rock mass structure modeling based on drilling data provided in an embodiment of the present application is shown;

[0021] Figure 3A schematic diagram of uniform discretization processing provided by an embodiment of the present application is shown;

[0022] Figure 4 A schematic diagram of an initial model of drilling properties provided by an embodiment of the present application is shown;

[0023] Figure 5 A schematic diagram of borehole spatial variability provided by an embodiment of the present application is shown;

[0024] Figure 6 A schematic diagram of RQD data partitioning of a mine borehole provided in an embodiment of the present application is shown;

[0025] Figure 7 A schematic diagram of a target rock mass structural property model provided in an embodiment of the present application is shown;

[0026] Figure 8 A schematic diagram of a multi-block stacking method provided in an embodiment of the present application is shown;

[0027] Figure 9 A schematic diagram of the weighted coefficient of variation of RQD distance of a new model provided in an embodiment of the present application is shown;

[0028] Figure 10 A schematic diagram of a cross-validation drilling RQD comparison provided in an embodiment of the present application is shown;

[0029] Figure 11 A schematic diagram of a process for a rock mass structure refined modeling method based on drilling data provided in an embodiment of the present application is shown;

[0030] Figure 12 A schematic diagram of a rock mass structure modeling method based on drilling data provided in an embodiment of the present application is shown;

[0031] Figure 13 A schematic diagram of another rock mass structure modeling based on drilling data provided by an embodiment of the present application is shown;

[0032] Figure 14 A schematic diagram of the device structure of a device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0033] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0034] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. Throughout the description of this application, "plurality" means two or more, unless otherwise specifically defined.

[0035] The following describes exemplary embodiments of the present application in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0036] Traditional rock mass structure modeling methods primarily rely on geological exploration data, often employing interpolation methods such as the inverse distance power method and kriging to construct block models. However, geological exploration processes have significant limitations. While drilling, as the primary method for acquiring rock mass information, plays a key role in rock mass information collection, its limitations are also significant. Traditional methods can only capture rock mass information within a localized area, insufficiently considering the spatial variability of deep rock mass structure and the potential geological correlations between borehole attributes. Furthermore, in complex geological conditions, such as karst areas and fault zones, the detection accuracy of geophysical exploration methods such as geological radar is severely affected, resulting in incomplete and inaccurate data. These factors make it difficult for the generated three-dimensional models to fully and accurately reflect rock mass structural characteristics and geological correlations. Model accuracy is particularly reduced at model boundaries far from sampling points. Therefore, in practical applications, traditional modeling methods often fail to meet the desired accuracy requirements, significantly limiting their effectiveness and reliability in engineering practice.

[0037] With the rapid development of computer technology, especially major breakthroughs in artificial intelligence, deep learning-based rock modeling techniques have emerged. These techniques demonstrate significant advantages in processing complex geological data and improving modeling accuracy. However, they suffer from drawbacks such as large data requirements, insufficient model interpretability, and limited generalization capabilities. These shortcomings restrict their widespread application in geotechnical engineering, making them difficult to meet the practical needs of both theoretical research and engineering practice.

[0038] To address this issue, this application proposes a rock structure modeling method based on borehole data. The method first preprocesses the borehole data and characterizes its spatial variability. An algorithm is then used to generate multiple heterogeneous block models, which are then reconstructed and fitted. The resulting fitted model not only matches the borehole data but also fully matches the spatial variability of the corresponding locations. This allows for a globally optimal rock structure refinement model, resolving issues such as low accuracy, large data requirements, insufficient interpretability, and limited generalization capabilities of existing models. The application's execution entity may be a rock structure modeling system, which relies on the computing power of a server to provide services to users. The server may be a standalone server or a server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0039] The present application embodiment provides a rock structure modeling method based on drilling data, such as Figure 1 As shown, the method includes:

[0040] 101. Obtain original drilling data, model the original drilling data after uniform discretization processing, and obtain an initial drilling attribute model.

[0041] In the embodiment of the present application, the original drilling data carries the real physical and mechanical properties of the rock mass and is a direct mapping of the rock mass characteristics at the engineering site. Through uniform discretization processing and modeling, the actual rock mass properties at the drilling location can be accurately converted into a modeled expression, providing a real and reliable basic data source for subsequent rock mass structure modeling. From the data input layer, it ensures that the model construction is based on objective engineering facts, avoids information distortion caused by data abstraction, and maintains the physical consistency between the model and the actual rock mass.

[0042] 102. Calculate multiple distance-weighted variation coefficients for characterizing the spatial variability of the initial model of borehole attributes, and generate multiple heterogeneous block model samples using the bubbling method based on the initial model of borehole attributes and the multiple distance-weighted variation coefficients.

[0043] In the embodiment of the present application, the distance-weighted coefficient of variation is an indicator for measuring the spatial variability of the initial model of borehole properties. Its core logic is to quantify the degree of variation of borehole properties on the basis of considering spatial distance. By assigning different weights to adjacent borehole properties at different distances, a coefficient is calculated to reflect the variation of the property value in space with distance, so as to characterize the variability characteristics of the rock mass properties in spatial distribution, such as unevenness and discreteness. The distance-weighted coefficient of variation introduces spatial distance weights. Compared with the ordinary coefficient of variation, it is more in line with the actual situation of the spatial distribution of rock mass properties, and can accurately capture the changing laws of rock mass properties in different spatial positions, so that the subsequently generated block model samples can more realistically reflect the actual heterogeneity of the rock mass, and improve the model's restoration of the complex spatial characteristics of the rock mass.

[0044] The bubbling method is a modeling method for simulating the heterogeneous distribution of rock masses. Based on the existing initial model of borehole properties and the multiple distance-weighted coefficients of variation calculated previously, the bubbling method simulates a random distribution process similar to "bubbling". It can be understood as randomly generating blocks that conform to the spatial property variation characteristics according to the variability law, and constructing multiple block model samples with different heterogeneous distributions. These samples can reflect the diverse and possible non-uniform distribution of rock mass properties in space.

[0045] 103. Using the initial model of borehole properties as a constraint, the weight coefficient matrix of multiple heterogeneous block model samples is solved by the least squares method. The weight coefficient matrix is ​​used to perform weighted superposition fitting on multiple heterogeneous block model samples to generate the target rock mass structural property model.

[0046] In the embodiment of the present application, the core logic of the least squares method is to minimize the residual sum of squares between the model prediction value and the actual drilling data, so that the target model after weighted superposition can fit the rock properties reflected by the actual drilling to the greatest extent, ensure the high-precision fitting of the model to the actual engineering data, and improve the consistency of the model with the on-site rock mass. The rock structure is highly heterogeneous, and multiple heterogeneous block model samples can reflect the possible distribution of rock properties from different angles. By solving the weight coefficient matrix and weighted superposition, the advantages of multiple samples, such as the accurate characterization of different regions and different variability characteristics of the rock mass by different samples, can be scientifically integrated, abandoning the one-sidedness of a single model and constructing a target model that integrates multi-dimensional information and more comprehensively reflects the rock structure properties.

[0047] An embodiment of the present application provides a rock structure modeling method based on borehole data. Compared with the prior art, the embodiment of the present application obtains original borehole data, models the original borehole data after uniform discretization processing, obtains an initial model of borehole properties, and then calculates multiple distance-weighted variation coefficients used to characterize the spatial variability of the initial model of borehole properties. Based on the initial model of borehole properties and the multiple distance-weighted variation coefficients, a bubbling method is used to generate multiple heterogeneous block model samples. Then, with the initial model of borehole properties as a constraint condition, the weight coefficient matrix of the multiple heterogeneous block model samples is solved by the least squares method. The weight coefficient matrix is ​​used to perform weighted superposition fitting on the multiple heterogeneous block model samples to generate a target rock structure property model. By effectively utilizing borehole data and reconstructing algorithms that consider geological variability, a model that more accurately reflects the actual rock mass structure can be established based on this data, even when the data has inherent flaws. This achieves high-precision fitting with the borehole data while also retaining and reasonably extrapolating the regional rock mass structure variability. This effectively addresses the problem of insufficient model accuracy in areas with sparse borehole data, effectively achieving high-precision modeling of rock mass structure on a global scale. Furthermore, by fully considering the spatial variability of borehole data when constructing the rock mass structure model, it can better adapt to the rock mass structure characteristics under different geological conditions. Whether in areas with relatively simple geological conditions or under complex geological conditions, it can flexibly generate and fit a rock mass structure model that conforms to the actual situation based on the characteristics of the borehole data, demonstrating strong generalization capabilities.

[0048] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, the embodiment of the present application provides another rock structure modeling method based on drilling data, such as Figure 2 As shown, the method includes:

[0049] 201. Obtain original drilling data, perform uniform discretization processing on the original drilling data, and obtain multiple discretized blocks.

[0050] In the embodiment of the present application, the original drilling data is preprocessed to remove data with missing data and data anomalies in the original drilling data to obtain the target drilling data. Taking drilling 1 as an example, the original drilling data is as follows Table 1 and Table 2:

[0051] Table 1

[0052]

[0053] Table 2

[0054]

[0055] Next, the target drilling data is converted into multiple three-dimensional drilling coordinates and the attribute value corresponding to each three-dimensional drilling coordinate. The multiple three-dimensional drilling coordinates and the attribute values ​​corresponding to the multiple three-dimensional drilling coordinates are uniformly discretized. For example, the data after uniform discretization of drilling 1 is shown in Table 3 below. Finally, the following is obtained: Figure 3 The figure shows multiple discretized blocks of equal size. The cube on the left represents the original rock mass space and the distribution of borehole data, providing an abstract representation of the three-dimensional space of the engineering site. The discrete points within it correspond to 3D spatial data samples collected from actual boreholes. The coordinates of each point reflect the sample's physical location in the rock mass space, and the point's attributes implicitly reflect the rock mass characteristics at that location. In rock mass structural modeling scenarios, these discrete points serve as raw data for subsequent analysis. However, their large number and scattered distribution increase computational cost and difficulty when used directly in complex model construction due to high data dimensionality and redundancy. The cube on the right represents the result of uniform discretization of the original data on the left, demonstrating the process of data simplification and structuring. After discretization, the original scattered borehole data is integrated into regularly arranged discretized blocks at uniform spatial intervals. Each block can be considered a basic analysis unit, and its attribute values, derived through statistics on the original discrete point data, represent the comprehensive characteristics of the rock mass properties within that spatial block. The cube on the right converts complex and disordered raw data into structured data that meets the needs of the modeling algorithm. This facilitates subsequent operations such as rock mass spatial variability analysis, block model construction, and weighted fitting, significantly reducing the complexity of modeling calculations while preserving the spatial distribution trend of rock mass properties. The formula meanvalue in the figure is a calculation method for discretized block property values. By taking the arithmetic mean of the n original discrete point data within the discretized block, a characteristic value representing the rock mass property of the discretized block is obtained. This effectively integrates the information of multiple raw data within the block and uses simple characteristic values ​​to reflect the rock mass property level in the local space, simplifying the discretized block data while retaining the statistical characteristics of the original data.

[0056] Table 3

[0057]

[0058] The first column of data in Table 3 represents the X coordinate, the second column of data represents the Y coordinate, and the third column of data represents the Z coordinate. The three columns of data together represent the three-dimensional drilling coordinates, and the RQD value is the attribute value corresponding to the three-dimensional drilling coordinates.

[0059] 202. Use multiple discretized blocks for modeling to obtain the initial model of drilling properties.

[0060] In an embodiment of the present application, for each discretized block, the multiple three-dimensional drilling coordinates included in the discretized block are determined, the centroid coordinates of the discretized block are calculated using the multiple three-dimensional drilling coordinates included in the discretized block, the attribute values ​​corresponding to the multiple three-dimensional drilling coordinates included in the discretized block are determined, and the average of the attribute values ​​corresponding to the multiple three-dimensional drilling coordinates included in the discretized block is used as the attribute value of the discretized block. Based on this process, each discretized block is processed separately to obtain the centroid coordinates and attribute values ​​of each discretized block. Then, the centroid coordinates and attribute values ​​of the multiple discretized blocks are used for modeling, and the following is obtained: Figure 4 The initial model of borehole properties is shown. By refining the complex data of each discretized block into two sets of key features, namely the centroid coordinates and mean attributes, the data dimension is greatly simplified. This not only preserves the spatial location and attribute characteristics of the block, but also reduces the computational complexity of subsequent modeling.

[0061] 203. Calculate multiple distance-weighted coefficients of variation for characterizing the spatial variability of the initial model of borehole attributes.

[0062] In the embodiments of the present application, the characterization of spatial variability is an important part of the method. First, a spherical domain with a radius of r centered on the discretized block is selected, i.e., a three-dimensional spatial neighborhood, and the set of discretized blocks involved in the calculation within the neighborhood is determined. Then, multiple discretized blocks included in the initial model of the drilling attributes are determined, and the distance-weighted coefficient of variation of each discretized block is calculated using the following formula 1:

[0063] Formula 1:

[0064]

[0065]

[0066]

[0067] in, Indicates the The distance-weighted coefficient of variation of the discretized blocks, Indicates the The weighted standard deviation of a discretized block is used to measure the degree of dispersion of attribute values ​​within the neighborhood relative to the weighted mean, reflecting the fluctuation amplitude of spatial variability. Indicates the The weighted average of the discretized blocks is calculated using the distance weight The attribute values ​​of the discretized blocks in the neighborhood Weighted averaging can highlight the influence of the properties of the near-center blocks and better fit the spatial characteristics of the rock mass properties with short-range correlation and long-range weakening. n represents the total number of discretized blocks in the sphere with radius r, represents the weight of the i-th discretized block in the sphere, represents the property value of the i-th discretized block in the sphere, Indicates the distance from the i-th discretized block to the i-th discretized block in the sphere The distance of the discretized blocks, where The discretized blocks are the center points of the sphere. Finally, the distance-weighted variation coefficients of multiple discretized blocks are used to characterize the spatial variability of the initial model of borehole attributes.

[0068] In an optional embodiment, the distance weighted variation coefficient can be plotted as follows: Figure 5 The borehole spatial variability diagram shown in the figure can intuitively reflect the variability of RQD attribute values ​​in different directions and distances in each area. This variability information can provide a reference for the geological correlation between boreholes for subsequent model generation. Among them, the region is divided into rock mass areas according to the attributes and spatial locations of the boreholes using the K-Means clustering algorithm, as shown in the figure. Figure 6 The RQD data partitioning diagram of a mine borehole is shown. The entire rock mass area is divided into 10 different areas based on the relationship between the spatial position of the borehole rock mass structural attribute values ​​and the rock mass attributes.

[0069] Based on the initial model of borehole attributes and multiple distance-weighted coefficients of variation, the bubbling method is used to generate multiple heterogeneous block model samples.

[0070] In an embodiment of the present application, target drilling data is obtained, and a clustering algorithm is used to partition the initial model of drilling attributes according to the target drilling data to obtain multiple partitions of the initial model of drilling attributes, wherein the clustering algorithm can be a K-Means clustering algorithm. Different regions are divided according to the spatial position relationship of the model and the attribute size distribution, so that different regions will be given different weights in the subsequent linear superposition process, which can improve the final model superposition accuracy. Then, the attribute value of each discretized block in the initial model of drilling attributes and the spatial size of the initial model of drilling attributes are obtained. Then, the bubbling method is used to calculate the multiple partitions of the initial model of drilling attributes, the attribute value of each discretized block in the initial model of drilling attributes, the spatial size of the initial model of drilling attributes, and multiple distance-weighted coefficients of variation to generate multiple heterogeneous block model samples, wherein the number of partitions of each heterogeneous block model sample is the same as the number of partitions of the initial model of drilling attributes. During this process, bubble generation needs to refer to the spatial size, number of partitions, attribute values ​​and coefficient of variation of the drilling model to ensure that the sample conforms to both random simulation heterogeneity and the spatial characteristics of the actual rock mass, so that each sample retains the number of partitions and spatial framework of the initial model and can simulate the complex heterogeneous distribution of the rock mass.

[0071] The bubbling method is a stochastic simulation-based approach for generating heterogeneous block model samples. The basic idea is to randomly generate a series of overlapping "bubbles" in three-dimensional space. By adjusting the size, number, and property distribution of the bubbles, the heterogeneous properties of the rock mass can be simulated. This method generates multiple block model samples with internal heterogeneity based on the spatial dimensions of the borehole model, the RQD attribute values ​​of the regional boreholes, and spatial variability data. Each block model sample should have the same partitions as the original borehole model. Specific bubbling method parameters can be found in Table 4:

[0072] Table 4

[0073]

[0074] pass Adjust the bubble radius discreteness, By controlling the number of bubbles, the rock heterogeneity of different engineering scenarios can be simulated in a targeted manner. The size matches the actual engineering space range to ensure the engineering relevance of the model. Adjust the discretization accuracy, By controlling the number of samples, we can select the model that best fits the actual drilling data and solve the problem of traditional single models having too strong assumptions and poor adaptability.

[0075] 205. Taking the initial model of borehole attributes as the constraint condition, the weight coefficient matrix of multiple heterogeneous block model samples is solved by the least square method.

[0076] In an embodiment of the present application, a plurality of discretized blocks included in the initial model of drilling attributes and a plurality of discretized blocks included in each heterogeneous block model sample are obtained, wherein the total number of the plurality of discretized blocks included in the initial model of drilling attributes is the same as the total number of the plurality of discretized blocks included in each heterogeneous block model sample, and the plurality of discretized blocks included in the initial model of drilling attributes and the plurality of discretized blocks included in each heterogeneous block model sample are in one-to-one correspondence, that is, the spatial positions and partitions are consistent. Then, the plurality of partitions included in the initial model of drilling attributes and the plurality of partitions included in each heterogeneous block model sample are determined, wherein the plurality of partitions of the initial model of drilling attributes and the plurality of partitions of each heterogeneous block model sample are in one-to-one correspondence. Then, the residual sum of squares of each partition is calculated.

[0077] Specifically, for each partition, the multiple discretized blocks of the initial model of the drilling attribute in the partition and the multiple discretized blocks of each heterogeneous block model sample in the partition are determined, and the residual sum of squares of the partition is calculated using the attribute values ​​of the multiple discretized blocks of the initial model of the drilling attribute in the partition and the attribute values ​​of the multiple discretized blocks of each heterogeneous block model sample in the partition. The calculation formula is the following formula 2:

[0078] Formula 2:

[0079]

[0080]

[0081] in, Represents the residual sum of squares of its partition, which is used to measure the overall deviation of the sample from the initial model. m represents the number of discretized blocks in its partition. It represents the error between the kth discretized block in its partition in multiple heterogeneous block model samples and the kth discretized block in its partition in the initial model of borehole attributes, represents the attribute value of the kth discretized block of the initial model of drilling attributes in its partition, Represents the predicted attribute value of the kth discretized block in its partition of multiple heterogeneous block model samples, represents the weight coefficient of the i-th heterogeneous block model sample in its partition, s represents the total number of multiple heterogeneous block model samples, represents the property value of the kth discretized block in the i-th heterogeneous block model sample within its partition. Finally, the least squares method is used to solve the sum of squared residuals for multiple partitions to obtain a weight coefficient matrix, which maximizes the fit of the target model to the actual drill hole data and improves model accuracy. Because rock mass properties naturally exhibit spatial variability, for example, fault zones differ from intact rock zones in their zoning characteristics, optimizing weights independently for each partition allows the sample weights in each region to better align with its own geological laws.

[0082] 206. The target rock mass structural property model is generated by weighted superposition fitting of multiple heterogeneous block model samples using a weight coefficient matrix.

[0083] In the embodiment of the present application, the attribute value matrix of each heterogeneous block model sample in each partition is determined, and the weight coefficient matrix and the attribute value matrix of each heterogeneous block model sample in each partition are used to perform weight coefficient superposition fitting to generate the following: Figure 7 The target rock mass structural property model shown is calculated using the following formula 3:

[0084] Formula 3:

[0085] in, represents the target attribute value matrix of the jth partition in the target rock mass structural attribute model, s represents the total number of multiple heterogeneous block model samples, represents the weight coefficient of the i-th heterogeneous block model sample in the j-th partition, The matrix represents the attribute values ​​of the i-th heterogeneous block model sample in the j-th partition. The partition results are obtained based on the spatial attribute characteristics of the rock mass, and then the weights are obtained through an optimization algorithm. The weights of different partitions are optimized independently, and the partition characteristics are strictly preserved during superposition. This ensures that the target model is highly consistent with the geological variability of the actual rock mass in space, avoiding feature loss caused by global averaging.

[0086] In an alternative embodiment, Figure 8 As shown, the left side of the figure shows n heterogeneous block model samples, each sample represents a possible distribution of rock mass structure. Each sample corresponds to a weight coefficient The weight coefficient is calculated using an optimization algorithm such as the least squares method, reflecting the sample's contribution to the final model. The larger the weight, the greater the sample's influence on the final model. The weighted summation of multiple block samples is then combined to generate a rock mass structural model, integrating the advantages of multiple samples. Attribute data from the preliminary model is then extracted using a virtual borehole and compared with the original borehole data to verify that the model matches the rock mass characteristics of the actual borehole.

[0087] 207. Calculate the complex correlation coefficient based on the target rock mass structural attribute model and the original rock mass structural attribute model. If the detection determines that the complex correlation coefficient meets the complex correlation coefficient determination rule, store the target rock mass structural attribute model.

[0088] In the embodiment of the present application, the original rock mass structure attribute model is obtained, and the complex correlation coefficient is calculated based on the target rock mass structure attribute model and the original rock mass structure attribute model. The calculation formula is the following formula 4:

[0089] Formula 4:

[0090]

[0091]

[0092]

[0093] in, represents the multiple correlation coefficient, represents the residual sum of squares, represents the regression sum of squares, represents the error between the pth original block in the original rock mass structural attribute model and the pth target block in the target rock mass structural attribute model, m represents the total number of multiple target blocks in the target rock mass structural attribute model, represents the property value of the pth original block in the original rock mass structure property model, represents the attribute value of the pth target block in the target rock mass structure attribute model, It represents the mean of the attribute values ​​of multiple original blocks in the original rock mass structural attribute model. Then, the complex correlation coefficient detection rule is obtained. If the detection determines that the complex correlation coefficient meets the complex correlation coefficient judgment rule, the target rock mass structural attribute model is stored. The complex correlation coefficient ranges from 0 to 1. The closer the value is to 1, the better the correlation of the generated model.

[0094] In an optional embodiment, multiple target distance weighted variation coefficients for characterizing the spatial variability of the target rock mass structural attribute model are calculated to generate Figure 9 The new model RQD distance weighted coefficient of variation diagram is shown, and it can be seen that the spatial variability is similar to Figure 5 The shown borehole spatial variability maps are similar, indicating that the generated rock mass structure model achieves sufficient fitting accuracy and conforms to the spatial variability information of the borehole data.

[0095] In an optional implementation scheme, a cross-validation method is used to test the target rock mass structural attribute model. Specifically, part of the drilling data is used as verification data, and the remaining data is used as modeling data. The target rock mass structural attribute model is constructed using the modeling data, and then the model is used to predict the RQD attribute value of the verification data. The predicted value is compared with the actual value to obtain the following results: Figure 10 The cross-validation borehole RQD comparison shown in the figure shows that the RQD curves for the original and generated boreholes show similar overall fluctuation trends. For example, the curve for original borehole 1 exhibits multiple peaks and valleys, and the curve for generated borehole 1 also exhibits corresponding peak-valley fluctuations. The fluctuation pattern of the curve for original borehole 2 also shows a similar trend in the curve for generated borehole 2. The complex correlation coefficient between the original and generated boreholes reaches above 0.6, significantly improving the complex correlation coefficient compared to other interpolation methods. This demonstrates that the model captures the overall pattern of rock mass RQD variation with depth, reflecting the spatially heterogeneous distribution of rock mass integrity. Furthermore, characteristic points in the curves, such as the locations of significant peaks and valleys, correspond well between the original and generated curves, and the complex correlation coefficient shows a better performance. Taking the sample point numbers of the original borehole 1 as an example, the curve for generated borehole 1 also shows peaks and valleys at similar locations, indicating that the model can identify and simulate key locations within the rock mass where rock integrity changes suddenly.

[0096] In summary, this application proposes a rock mass structure refinement modeling method based on drilling data, such as Figure 11As shown in the figure, a borehole spatial distribution model is first established. By collecting and integrating information such as the three-dimensional coordinates of on-site boreholes, the spatial position framework of the rock boreholes is outlined. Then, the borehole model is uniformly discretized, and the continuous borehole data are split into regular discrete blocks according to a uniform spatial interval. Then, the engineering area is partitioned according to the borehole location and attribute value. Based on the association between rock mass attributes and spatial location, clustering and other algorithms are used to divide areas with similar geological characteristics. Finally, the spatial variability of the engineering rock mass is characterized based on the borehole data, and the spatial fluctuation law of the rock mass attributes in different partitions is quantified with the help of methods such as the distance-weighted coefficient of variation.

[0097] Based on the previous data and characteristics, the rock structure modeling parameters (such as the radius standard deviation and the number of blocks of the bubbling method) are first selected to clarify the modeling rules; then, multiple block model samples with non-uniform property distribution are generated, and the bubbling method is used to simulate the rock heterogeneity, so that the samples cover a variety of geological scenarios; then, the sample weights are optimized with the drilling data as a constraint, and the least squares method is used to make the weighted samples accurately fit the measured drilling data, highlighting the contribution of key samples to the model; finally, the rock structure model is generated by superimposing the fitting, and the weighted samples are integrated to construct a model that preliminarily reflects the spatial distribution of rock properties.

[0098] After the model is built, a test of fitting accuracy and spatial variability is conducted, comparing the deviation between the model's predicted values ​​and the measured values ​​in the borehole to assess whether the model fits reality. If the test fails, the feedback correction parameter mechanism is triggered, and the modeling parameters are adjusted in reverse (such as optimizing sample weights and adjusting discretization accuracy), and the parameter selection-sample generation-weighted fitting process is re-executed. If the test passes, the model is confirmed to be valid, and a globally optimal rock mass structure refinement model is established, providing a high-precision foundation for engineering analysis. This application uses borehole attribute values ​​and spatial variability as constraints, and uses an algorithm to generate multiple rock mass structure model samples with non-uniform internal attributes for reconstruction and fitting. This results in a highly accurate rock mass structure model that can compensate for the accuracy loss caused by data defects and maximize the mining and utilization of effective information in the borehole data. Based on the effective use of borehole data and the reconstruction algorithm of spatial variability characteristics, it achieves high-precision fitting with borehole data in terms of accuracy, while fully preserving the variability characteristics of the regional rock mass structure and achieving reasonable extrapolation; in terms of data requirements, it significantly reduces the reliance on densely sampled data; in terms of model interpretability, it establishes a clear modeling process; in terms of generalization ability, it can adapt to modeling needs under different geological conditions. This application shows significant advantages in rock mass structure modeling in the field of geotechnical engineering, providing a more efficient and reliable technical approach that balances high-precision modeling with the maintenance of geological relevance.

[0099] An embodiment of the present application provides a rock structure modeling method based on borehole data. Compared with the prior art, the embodiment of the present application obtains original borehole data, models the original borehole data after uniform discretization processing, obtains an initial model of borehole properties, and then calculates multiple distance-weighted variation coefficients used to characterize the spatial variability of the initial model of borehole properties. Based on the initial model of borehole properties and the multiple distance-weighted variation coefficients, a bubbling method is used to generate multiple heterogeneous block model samples. Then, with the initial model of borehole properties as a constraint condition, the weight coefficient matrix of the multiple heterogeneous block model samples is solved by the least squares method. The weight coefficient matrix is ​​used to perform weighted superposition fitting on the multiple heterogeneous block model samples to generate a target rock structure property model. By effectively utilizing borehole data and reconstructing algorithms that consider geological variability, a model that more accurately reflects the actual rock mass structure can be established based on this data, even when the data has inherent flaws. This achieves high-precision fitting with the borehole data while also retaining and reasonably extrapolating the regional rock mass structure variability. This effectively addresses the problem of insufficient model accuracy in areas with sparse borehole data, effectively achieving high-precision modeling of rock mass structure on a global scale. Furthermore, by fully considering the spatial variability of borehole data when constructing the rock mass structure model, it can better adapt to the rock mass structure characteristics under different geological conditions. Whether in areas with relatively simple geological conditions or under complex geological conditions, it can flexibly generate and fit a rock mass structure model that conforms to the actual situation based on the characteristics of the borehole data, demonstrating strong generalization capabilities.

[0100] Further, as Figure 1 In a specific implementation of the method, the present application provides a rock structure modeling device based on drilling data, such as Figure 12 As shown, the device includes: an initial model building module 301, a heterogeneous sample generating module 302 and a weight fitting modeling module 303.

[0101] The initial model building module 301 is used to obtain original drilling data, model the original drilling data after uniform discretization processing, and obtain an initial drilling attribute model;

[0102] A heterogeneous sample generation module 302 is configured to calculate a plurality of distance-weighted variation coefficients for characterizing the spatial variability of the initial borehole attribute model, and generate a plurality of heterogeneous block model samples using a bubbling method based on the initial borehole attribute model and the plurality of distance-weighted variation coefficients;

[0103] The weight fitting modeling module 303 is used to solve the weight coefficient matrix of the multiple heterogeneous block model samples by the least squares method with the initial model of the drilling properties as a constraint condition, and use the weight coefficient matrix to perform weighted superposition fitting on the multiple heterogeneous block model samples to generate a target rock structure property model.

[0104] In a specific application scenario, the initial model construction module 301 is used to preprocess the original drilling data, remove data with missing data and data anomalies in the original drilling data, and obtain target drilling data; convert the target drilling data into multiple three-dimensional drilling coordinates and attribute values ​​corresponding to each of the three-dimensional drilling coordinates; uniformly discretize the multiple three-dimensional drilling coordinates and the attribute values ​​corresponding to the multiple three-dimensional drilling coordinates to obtain multiple discretized blocks; for each of the discretized blocks, determine the multiple three-dimensional drilling coordinates included in the discretized block, calculate and determine the center of mass coordinates of the discretized block using the multiple three-dimensional drilling coordinates included in the discretized block, determine the attribute values ​​corresponding to the multiple three-dimensional drilling coordinates included in the discretized block, and use the average of the attribute values ​​corresponding to the multiple three-dimensional drilling coordinates included in the discretized block as the attribute value of the discretized block; process each of the discretized blocks separately to obtain the center of mass coordinates and attribute values ​​of each discretized block; and use the center of mass coordinates and attribute values ​​of the multiple discretized blocks for modeling to obtain the initial model of the drilling attribute.

[0105] In a specific application scenario, the heterogeneous sample generation module 302 is used to determine a plurality of discretized blocks included in the initial model of borehole attributes, and calculate the distance-weighted coefficient of variation of each discretized block.

[0106]

[0107]

[0108]

[0109]

[0110] in, Indicates the The distance-weighted coefficient of variation of the discretized blocks, Indicates the The weighted standard deviation of the discretized blocks, Indicates the The weighted average of discretized blocks, n represents the total number of discretized blocks in the sphere with radius r, represents the weight of the i-th discretized block in the sphere, represents the property value of the i-th discretized block in the sphere, Indicates the distance from the i-th discretized block to the i-th discretized block in the sphere The distance of the discretized blocks, where the The plurality of discretized blocks are the center points of the spherical domain; and the distance-weighted variation coefficients of the plurality of discretized blocks are used to characterize the spatial variability of the initial model of the drilling attributes.

[0111] In a specific application scenario, the heterogeneous sample generation module 302 is used to obtain target drilling data, use a clustering algorithm to partition the drilling attribute initial model according to the target drilling data, and obtain multiple partitions of the drilling attribute initial model; obtain the attribute value of each discretized block in the drilling attribute initial model, and the spatial size of the drilling attribute initial model; use the bubbling method to calculate the multiple partitions of the drilling attribute initial model, the attribute value of each discretized block in the drilling attribute initial model, the spatial size of the drilling attribute initial model, and the multiple distance-weighted variation coefficients to generate the multiple heterogeneous block model samples, and the number of partitions of each of the heterogeneous block model samples is the same as the number of partitions of the drilling attribute initial model.

[0112] In a specific application scenario, the weighted fitting modeling module 303 is used to obtain a plurality of discretized blocks included in the initial model of the drilling attribute, and a plurality of discretized blocks included in each of the heterogeneous block model samples, wherein the total number of the plurality of discretized blocks included in the initial model of the drilling attribute is the same as the total number of the plurality of discretized blocks included in each of the heterogeneous block model samples, and the plurality of discretized blocks included in the initial model of the drilling attribute is in one-to-one correspondence with the plurality of discretized blocks included in each of the heterogeneous block model samples; determine a plurality of partitions included in the initial model of the drilling attribute, and each partition The heterogeneous block model samples include multiple partitions, wherein the multiple partitions of the initial model of the borehole attributes correspond one to one to the multiple partitions of each heterogeneous block model sample; the residual sum of squares of each partition is calculated, and the residual sum of squares of the multiple partitions is solved by the least squares method to obtain the weight coefficient matrix; the attribute value matrix of each heterogeneous block model sample in each partition is determined, and the weight coefficient matrix and the attribute value matrix of each heterogeneous block model sample in each partition are used to perform weight coefficient superposition fitting to generate the target rock structure attribute model,

[0113]

[0114] in, represents the target attribute value matrix of the jth partition in the target rock mass structural attribute model, s represents the total number of the multiple heterogeneous block model samples, represents the weight coefficient of the i-th heterogeneous block model sample in the j-th partition, Represents the attribute value matrix of the i-th heterogeneous block model sample in the j-th partition.

[0115] In a specific application scenario, the weighted fitting modeling module 303 is used to determine, for each partition, a plurality of discretized blocks of the initial model of the drilling attribute in the partition, and a plurality of discretized blocks of each of the heterogeneous block model samples in the partition; and calculate the residual sum of squares of the partition using the attribute values ​​of the initial model of the drilling attribute in the partition, and the attribute values ​​of the plurality of discretized blocks of each of the heterogeneous block model samples in the partition.

[0116]

[0117]

[0118]

[0119] in, represents the residual sum of squares of the partition, m represents the number of discretized blocks in the partition, represents the error between the kth discretized block of the plurality of heterogeneous block model samples in the partition and the kth discretized block of the initial drilling attribute model in the partition, represents the attribute value of the kth discretized block of the initial drilling attribute model in the partition, represents the predicted attribute value of the kth discretized block of the plurality of heterogeneous block model samples in the partition, represents the weight coefficient of the i-th heterogeneous block model sample in the partition, s represents the total number of the multiple heterogeneous block model samples, Represents the property value of the kth discretized block of the i-th heterogeneous block model sample in the partition.

[0120] In specific application scenarios, such as Figure 13 As shown, the device further includes: a testing module 304.

[0121] The verification module 304 is used to obtain the original rock mass structure attribute model, calculate the complex correlation coefficient based on the target rock mass structure attribute model and the original rock mass structure attribute model,

[0122]

[0123]

[0124]

[0125]

[0126] in, represents the complex correlation coefficient, represents the residual sum of squares, represents the regression sum of squares, represents the error between the pth original block in the original rock mass structural attribute model and the pth target block in the target rock mass structural attribute model, m represents the total number of multiple target blocks in the target rock mass structural attribute model, represents the attribute value of the pth original block in the original rock mass structure attribute model, represents the attribute value of the pth target block in the target rock mass structure attribute model, Representing the mean value of the attribute values ​​of multiple original blocks in the original rock mass structure attribute model; obtaining a complex correlation coefficient detection rule, and if the detection determines that the complex correlation coefficient meets the complex correlation coefficient determination rule, storing the target rock mass structure attribute model.

[0127] The present invention provides an apparatus that, compared with the prior art, obtains raw borehole data, models the raw borehole data after uniform discretization, obtains an initial borehole attribute model, then calculates multiple distance-weighted coefficients of variation used to characterize the spatial variability of the initial borehole attribute model, generates multiple heterogeneous block model samples using the bubbling method based on the initial borehole attribute model and the multiple distance-weighted coefficients of variation, and then uses the initial borehole attribute model as a constraint to solve the weight coefficient matrix of the multiple heterogeneous block model samples using the least squares method. The weight coefficient matrix is ​​used to perform weighted superposition fitting on the multiple heterogeneous block model samples to generate a target rock mass structural attribute model. By effectively utilizing the borehole data and a reconstruction algorithm that considers geological variability characteristics, a model that more accurately reflects the actual rock mass structure can be established based on the data even when the data has inherent defects, achieving high-precision fitting with the borehole data while achieving high-precision retention and reasonable extrapolation of regional rock mass structural variability characteristics, effectively solving the problem of insufficient model accuracy of traditional methods in areas with sparse borehole data, and achieving high-precision modeling of rock mass structure on a global scale. Moreover, the spatial variability of drilling data is fully considered when constructing the rock structure model, which can better adapt to the rock structure characteristics under different geological conditions. Whether in areas with relatively simple geological conditions or under complex geological conditions, it can flexibly generate and fit a rock structure model that conforms to the actual situation based on the characteristics of drilling data, and has strong generalization ability.

[0128] It should be noted that for other corresponding descriptions of the functional units involved in the rock structure modeling device based on drilling data provided in the embodiment of the present application, reference can be made to Figures 1 to 11 The corresponding description in will not be repeated here.

[0129] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0130] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0131] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

[0132] In an exemplary embodiment, see Figure 14 A device is also provided, comprising a bus, a processor, a memory, and a communication interface. The device may also include an input / output interface and a display device, wherein the various functional units can communicate with each other via the bus. The memory stores a computer program, and the processor is configured to execute the program stored in the memory and implement the rock structure modeling method based on borehole data in the above-described embodiment.

[0133] A medium stores a computer program, which, when executed by a processor, implements the steps of the rock structure modeling method based on drilling data.

[0134] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented through hardware or by using software plus the necessary general hardware platform. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, mobile hard disk, etc.) and includes a number of instructions for enabling a computer device (such as a personal computer, server, or network device) to execute the methods described in various implementation scenarios of the present application.

[0135] Those skilled in the art will understand that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily required to implement the present application.

[0136] Those skilled in the art will appreciate that the modules in the devices in the implementation scenario can be distributed in the devices of the implementation scenario according to the implementation scenario description, or can be modified accordingly and located in one or more devices different from the implementation scenario. The modules in the above implementation scenario can be combined into one module or further split into multiple submodules.

[0137] The above application serial numbers are for description only and do not represent the advantages or disadvantages of the implementation scenarios.

[0138] The above disclosure only describes several specific implementation scenarios of the present application. However, the present application is not limited thereto, and any changes that can be conceived by those skilled in the art should fall within the scope of protection of the present application.

Claims

1. A rock structure modeling method based on drilling data, characterized in that: include: Obtaining original drilling data, modeling the original drilling data after uniform discretization processing, and obtaining an initial drilling attribute model; calculating a plurality of distance-weighted coefficients of variation for characterizing the spatial variability of the initial model of borehole attributes, and generating a plurality of heterogeneous block model samples using a bubbling method based on the initial model of borehole attributes and the plurality of distance-weighted coefficients of variation; Using the initial borehole attribute model as a constraint condition, solving the weight coefficient matrix of the multiple heterogeneous block model samples by the least squares method, and performing weighted superposition fitting on the multiple heterogeneous block model samples using the weight coefficient matrix to generate a target rock mass structural attribute model; The method comprises the following steps: using the initial borehole attribute model as a constraint condition, solving the weight coefficient matrix of the plurality of heterogeneous block model samples by the least square method, and performing weighted superposition fitting on the plurality of heterogeneous block model samples using the weight coefficient matrix to generate a target rock mass structural attribute model. The method comprises the following steps: Obtaining a plurality of discretized blocks included in the initial drilling attribute model and a plurality of discretized blocks included in each of the heterogeneous block model samples, wherein the total number of the plurality of discretized blocks included in the initial drilling attribute model is the same as the total number of the plurality of discretized blocks included in each of the heterogeneous block model samples, and the plurality of discretized blocks included in the initial drilling attribute model and the plurality of discretized blocks included in each of the heterogeneous block model samples are in one-to-one correspondence; Determining a plurality of partitions included in the initial model of drilling properties and a plurality of partitions included in each of the heterogeneous block model samples, wherein the plurality of partitions of the initial model of drilling properties and the plurality of partitions of each of the heterogeneous block model samples are in one-to-one correspondence; Calculating the residual sum of squares of each partition, and solving the residual sum of squares of the multiple partitions using the least squares method to obtain the weight coefficient matrix; Determine the attribute value matrix of each heterogeneous block model sample in each partition, use the weight coefficient matrix and the attribute value matrix of each heterogeneous block model sample in each partition to perform weight coefficient superposition fitting to generate the target rock mass structure attribute model, in, represents the target attribute value matrix of the jth partition in the target rock mass structural attribute model, s represents the total number of the multiple heterogeneous block model samples, represents the weight coefficient of the i-th heterogeneous block model sample in the j-th partition, Represents the attribute value matrix of the i-th heterogeneous block model sample in the j-th partition.

2. The method according to claim 1, characterized in that The modeling of the original drilling data after uniform discretization to obtain an initial drilling attribute model includes: Preprocessing the original drilling data to remove data with missing data and data anomalies in the original drilling data to obtain target drilling data; Converting the target drilling data into a plurality of three-dimensional drilling coordinates and an attribute value corresponding to each of the three-dimensional drilling coordinates; uniformly discretizing the plurality of three-dimensional drilling coordinates and the attribute values ​​corresponding to the plurality of three-dimensional drilling coordinates to obtain a plurality of discretized blocks; For each of the discretized blocks, determining a plurality of three-dimensional borehole coordinates included in the discretized block, calculating and determining the coordinates of the centroid of the discretized block using the plurality of three-dimensional borehole coordinates included in the discretized block, determining attribute values ​​corresponding to the plurality of three-dimensional borehole coordinates included in the discretized block, and taking an average of the attribute values ​​corresponding to the plurality of three-dimensional borehole coordinates included in the discretized block as the attribute value of the discretized block; Processing each of the discretized blocks separately to obtain the centroid coordinates and attribute values ​​of each of the discretized blocks; Modeling is performed using the centroid coordinates and attribute values ​​of the plurality of discretized blocks to obtain the initial model of the drilling attribute.

3. The method according to claim 1, characterized in that The calculating of a plurality of distance-weighted coefficients of variation for characterizing the spatial variability of the initial model of borehole attributes includes: Determine a plurality of discretized blocks included in the initial model of borehole attributes, calculate the distance-weighted coefficient of variation of each discretized block, in, Indicates the The distance-weighted coefficient of variation of the discretized blocks, Indicates the The weighted standard deviation of the discretized blocks, Indicates the The weighted average of discretized blocks, n represents the total number of discretized blocks in the sphere with radius r, represents the weight of the i-th discretized block in the sphere, represents the property value of the i-th discretized block in the sphere, Indicates the distance from the i-th discretized block to the i-th discretized block in the sphere The distance of the discretized blocks, where the A discretized block is the center point of the spherical domain; The distance-weighted coefficient of variation of the plurality of discretized blocks is used to characterize the spatial variability of the initial model of the borehole attributes.

4. The method according to claim 1, wherein The method of generating a plurality of heterogeneous block model samples using a bubbling method based on the initial model of the borehole attributes and the plurality of distance-weighted coefficients of variation includes: Acquire target drilling data, and partition the initial drilling attribute model according to the target drilling data using a clustering algorithm to obtain multiple partitions of the initial drilling attribute model; Obtaining the attribute value of each discretized block in the initial drilling attribute model and the spatial size of the initial drilling attribute model; The bubbling method is used to calculate multiple partitions of the drilling attribute initial model, the attribute value of each discretized block in the drilling attribute initial model, the spatial size of the drilling attribute initial model, and the multiple distance weighted variation coefficients to generate the multiple heterogeneous block model samples. The number of partitions of each of the heterogeneous block model samples is the same as the number of partitions of the drilling attribute initial model.

5. The method according to claim 1, wherein Calculating the residual sum of squares of each partition includes: For each of the partitions, determining a plurality of discretized blocks of the initial model of the borehole properties in the partition, and a plurality of discretized blocks of each of the heterogeneous block model samples in the partition; The residual sum of squares of the partition is calculated using the attribute values ​​of the multiple discretized blocks of the initial drilling attribute model in the partition and the attribute values ​​of the multiple discretized blocks of each heterogeneous block model sample in the partition. in, represents the residual sum of squares of the partition, m represents the number of discretized blocks in the partition, represents the error between the kth discretized block of the plurality of heterogeneous block model samples in the partition and the kth discretized block of the initial drilling attribute model in the partition, represents the attribute value of the kth discretized block of the initial drilling attribute model in the partition, represents the predicted attribute value of the kth discretized block of the plurality of heterogeneous block model samples in the partition, represents the weight coefficient of the i-th heterogeneous block model sample in the partition, s represents the total number of the multiple heterogeneous block model samples, Represents the property value of the kth discretized block of the i-th heterogeneous block model sample in the partition.

6. The method according to claim 1, characterized in that The method further comprises: Obtaining an original rock mass structural attribute model, calculating a complex correlation coefficient based on the target rock mass structural attribute model and the original rock mass structural attribute model, in, represents the complex correlation coefficient, represents the residual sum of squares, represents the regression sum of squares, represents the error between the pth original block in the original rock mass structural attribute model and the pth target block in the target rock mass structural attribute model, m represents the total number of multiple target blocks in the target rock mass structural attribute model, represents the attribute value of the pth original block in the original rock mass structure attribute model, represents the attribute value of the pth target block in the target rock mass structure attribute model, represents the mean value of the attribute values ​​of multiple original blocks in the original rock mass structure attribute model; A complex correlation coefficient detection rule is obtained, and if the detection determines that the complex correlation coefficient meets the complex correlation coefficient determination rule, the target rock mass structural attribute model is stored.

7. A rock structure modeling device based on drilling data, characterized in that: include: An initial model building module is used to obtain original drilling data, model the original drilling data after uniform discretization processing, and obtain an initial drilling attribute model; a heterogeneous sample generation module, configured to calculate a plurality of distance-weighted variation coefficients for characterizing the spatial variability of the initial borehole attribute model, and to generate a plurality of heterogeneous block model samples using a bubbling method based on the initial borehole attribute model and the plurality of distance-weighted variation coefficients; a weighted fitting modeling module, configured to solve a weight coefficient matrix of the plurality of heterogeneous block model samples by a least squares method using the initial model of the borehole attributes as a constraint condition, and perform weighted superposition fitting on the plurality of heterogeneous block model samples using the weight coefficient matrix to generate a target rock mass structural attribute model; The weight fitting modeling module is further used to obtain a plurality of discretized blocks included in the initial model of drilling attributes, and a plurality of discretized blocks included in each of the heterogeneous block model samples, wherein the total number of the plurality of discretized blocks included in the initial model of drilling attributes is the same as the total number of the plurality of discretized blocks included in each of the heterogeneous block model samples, and the plurality of discretized blocks included in the initial model of drilling attributes and the plurality of discretized blocks included in each of the heterogeneous block model samples are in one-to-one correspondence; determine a plurality of partitions included in the initial model of drilling attributes, and each of the heterogeneous The rock mass structure attribute model comprises a plurality of partitions, wherein the plurality of partitions of the initial model of the borehole attributes correspond one to one with the plurality of partitions of each of the heterogeneous block model samples; the residual sum of squares of each of the partitions is calculated, and the residual sum of squares of the plurality of partitions is solved by the least squares method to obtain the weight coefficient matrix; the attribute value matrix of each of the heterogeneous block model samples in each of the partitions is determined, and the weight coefficient matrix and the attribute value matrix of each of the heterogeneous block model samples in each of the partitions are used to perform weight coefficient superposition fitting to generate the target rock mass structure attribute model, in, represents the target attribute value matrix of the jth partition in the target rock mass structural attribute model, s represents the total number of the multiple heterogeneous block model samples, represents the weight coefficient of the i-th heterogeneous block model sample in the j-th partition, Represents the attribute value matrix of the i-th heterogeneous block model sample in the j-th partition.

8. A device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images