Method, device and storage medium for searching for dominant occurrence of rock joints

Abstract

The present application discloses a method, device and storage medium for searching for the dominant occurrence of rock joints. First, obtain n groups of measurement data of joint occurrence and carry out coordinate conversion, converting them from polar coordinates to Cartesian coordinates; use the average arc length method to remove outliers; then, use the measurement data after the outliers are removed as the data source to construct the initial nest matrix and the corresponding objective function value; use the Levy flight operation in the cuckoo algorithm to iterate the initial nest matrix; obtain the optimal nest matrix and the corresponding objective function value according to the preset number of iterations and the ratio of abandoned nests; finally, determine the optimal number of clustering centers and their corresponding dominant occurrence of rock joints by measuring the clustering effect of the optimal nest matrix. It can be seen that the present application is based on the K-means clustering algorithm, and uses the cuckoo search algorithm to perform cluster analysis on the dominant occurrence of joints, thereby improving the accuracy and representativeness of the searched dominant occurrence of rock joints.

Description

Technical Field

[0001] The invention relates to the field of geotechnical engineering investigation, and in particular to a method, a device and a storage medium for searching for the dominant occurrence of rock joints. Background Art

[0002] From the perspective of rock engineering, joints cause the rock mass to lose its original continuity and solid integrity, further changing the permeability and engineering properties of the rock mass. They are essential information for determining the layout of chambers and the direction of excavation in the project.

[0003] Joints are structural features produced in a certain part of the earth's crust under the action of a certain tectonic stress field at a certain period of time. By understanding their regularity, we can infer the tectonic stress field and tectonic movement mode when they were formed, and can better reflect the structural characteristics. The occurrence is the primary characteristic of joints, which determines the spatial development form of joints. Therefore, for joints with small development scale, large number of groups, and complex data, statistically analyzing their occurrence distribution and obtaining the dominant occurrence of joints can reflect the development law of regional rock mass joints, which is of great significance for rock excavation engineering. Summary of the invention

[0004] Based on this, the embodiments of the present application provide a method, device and storage medium for searching for the dominant occurrence of rock joints, which effectively improve the accuracy of the searched rock joints compared with traditional technologies.

[0005] In a first aspect, a method for searching for the dominant occurrence of rock joints is provided, the method comprising:

[0006] Obtaining n groups of joint attitude measurement data, and converting the measurement data to obtain a measurement data matrix in a Cartesian coordinate system; the measurement data includes dip data and dip angle data;

[0007] The average arc length method is used to calculate the outliers in the measurement data matrix, and the outliers are eliminated;

[0008] The measured data matrix after outliers are removed is used as the data source, the number of cluster centers K is set, the objective function is constructed, the initial nest matrix is ​​constructed based on the population number P, and the corresponding objective function value is calculated. The objective function value is used to measure the overall clustering effect of the joints;

[0009] The initial nest matrix is ​​iterated by using the Levy flight operation in the cuckoo algorithm; the optimal nest matrix and the objective function value corresponding to the optimal nest matrix are obtained by processing according to the preset number of iterations N and the nest abandonment ratio γ;

[0010] The optimal number of cluster centers and their corresponding dominant occurrence of rock joints are determined by measuring the clustering effect of the optimal nest matrix.

[0011] Optionally, the measurement data includes dip data and dip angle data, and is characterized in that the acquisition of n groups of measurement data of joint occurrences and the conversion of the measurement data to obtain a measurement data matrix in a Cartesian coordinate system include:

[0012] The n groups of dip data and inclination data of joint occurrences form an n×2 matrix, and coordinate transformation is performed using the first formula to obtain a measurement data matrix, wherein the measurement data matrix is ​​an n×3 matrix;

[0013] The first formula can be expressed as:

[0014]

[0015] Where i represents the occurrence of the i-th group of joints, θ i Indicates the inclination data (0≤θ≤360°), Indicates the inclination data x i ,y i ,z i They respectively represent the x-axis, y-axis, and z-axis coordinates of the unit normal vector of the i-th group of joints.

[0016] Optionally, the calculating the outliers in the measurement data matrix by using the average arc length method includes:

[0017] Calculating the arc length distance between each group of joints and other joints in the measurement data matrix, and calculating the average value to obtain an n×1 column vector;

[0018] The mean μ and standard deviation σ of all elements in the column vector are calculated, and then the row numbers of the elements in the column vector that are greater than μ+3σ are screened, which are the outliers in the measurement data matrix.

[0019] Optionally, the calculating the arc length distance between each group of joints and other joints in the measurement data matrix comprises:

[0020] The arc length distance between each group of joints and other joints is calculated by the second formula, wherein the second formula specifically includes:

[0021] d(e i ,e j )=arccos|e i ·e j T |

[0022] Among them, e i represents the unit normal vector of the i-th group of joints, e j represents the unit normal vector of the jth group of joints, d(e i ,e j ) represents the arc length distance between the ith group of joints and the jth group of joints.

[0023] Optionally, the objective function value F is expressed by a third formula, and the third formula specifically includes:

[0024]

[0025] Among them, K represents the number of cluster centers, S j represents the unit normal vector of the joints contained in the jth class, c j represents the coordinate value of the jth cluster center, e i represents the unit normal vector of the i-th group of joints.

[0026] Optionally, the iterating the initial nest matrix using the Levy flight operation in the cuckoo algorithm includes:

[0027] Input the nest matrix E of the current training batch t and the optimal nest matrix E t_best , according to the fourth formula, we can get the new nest matrix E of the next training batch t+1 , the fourth formula specifically includes:

[0028]

[0029] Among them, α represents the step size scaling factor, Levy (β) represents the Levy flight path, and β represents the flight factor, and its value range is (1, 3).

[0030] Optionally, the Levy flight path adopts Mantegna's method, uses normal distribution to generate random steps that obey the Levy distribution, and obtains the Levy flight path according to the fourth formula. The fifth formula includes:

[0031]

[0032] Among them, Levy(β) represents the Levy flight path, β represents the range of flight factor, E t represents the nest matrix of the current training batch, E t_best represents the optimal nest matrix of the current training batch, v represents a random variable that satisfies the standard normal distribution, μ represents a random variable that satisfies the normal distribution, and their mathematical expectations are both 0. The standard deviations of v and μ are 1 and σ respectively.

[0033] The method of determining the optimal number of cluster centers and the corresponding rock joint dominant occurrence by measuring the clustering effect of the optimal nest matrix includes:

[0034] The Ray-Turi index υ is used to measure the clustering effect, and the optimal number of cluster centers is determined according to the sixth formula, which includes:

[0035]

[0036] Among them, K represents the number of cluster centers, n represents the number of joints, S j represents the unit normal vector of the joints contained in the jth class, c i represents the coordinate value of the i-th cluster center, c j Represents the coordinate value of the jth cluster center.

[0037] In a second aspect, a rock mass joint dominant occurrence search device is provided, the device comprising:

[0038] An acquisition module is used to acquire n groups of joint occurrence measurement data, and convert the measurement data to obtain a measurement data matrix in a Cartesian coordinate system; the measurement data includes dip data and dip angle data;

[0039] A processing module, used for calculating the outliers in the measurement data matrix by using the average arc length method, and removing the outliers;

[0040] A construction module is used to use the measurement data matrix after outliers are removed as the data source, set the number of cluster centers K, construct the objective function, construct the initial nest matrix based on the population number P and calculate the corresponding objective function value, and the objective function value is used to measure the overall clustering effect of the joints;

[0041] A calculation module is used to iterate the initial nest matrix using the Levy flight operation in the cuckoo algorithm; and obtain the optimal nest matrix and the objective function value corresponding to the optimal nest matrix according to the preset number of iterations N and the nest abandonment ratio γ;

[0042] The determination module is used to determine the optimal number of cluster centers and their corresponding dominant rock joint occurrences by measuring the clustering effect of the optimal nest matrix.

[0043] In a third aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for searching for the dominant occurrence of rock joints described in any one of the first aspects is implemented.

[0044] The technical solution provided in the embodiment of the present application obtains n groups of measurement data of joint occurrence, and performs coordinate transformation, converting the measurement data matrix from polar coordinate form to Cartesian coordinate form; using the average arc length method to remove outliers; then, using the measurement data matrix after outliers are removed as the data source, setting the number of cluster centers, constructing the objective function, constructing the initial nest matrix based on the population size and calculating the corresponding objective function value; using the Levy flight operation in the cuckoo algorithm to iterate the initial nest matrix; processing according to the preset number of iterations and the ratio of abandoned nests to obtain the optimal nest matrix and the corresponding objective function value; finally, determining the optimal number of cluster centers and their corresponding rock joint dominant occurrence by measuring the clustering effect of the optimal nest matrix. It can be seen that the beneficial effects of the present application include:

[0045] (1) Based on the fact that the outliers of the joint occurrence are far from the average arc length of the other joints, the outliers are screened and eliminated;

[0046] (2) Based on the K-means clustering algorithm, the cuckoo search algorithm is used to perform cluster analysis on the dominant occurrence of joints to enhance the global search capability and algorithm robustness;

[0047] (3) The dominant joint occurrence in the best classification is replaced by the adjacent joint occurrence from the calculation center, and the effectiveness index Ray-Turi index is used to determine the best classification K, which enhances the representativeness of the dominant joint occurrence. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0049] Figure 1 A flow chart of a method for searching for the dominant occurrence of rock joints provided in an embodiment of the present application;

[0050] Figure 2 A schematic diagram of a Cartesian coordinate joint unit normal vector provided in an embodiment of the present application;

[0051] Figure 3 A schematic diagram of a flow chart of training iterations of the rock joint dominant occurrence search method provided in an embodiment of the present application;

[0052] Figure 4 A pole diagram of joint measurement data obtained in an optional embodiment of the present application;

[0053] Figure 5A schematic diagram of removed outliers of joint measurement data obtained in an optional embodiment of the present application;

[0054] Figure 6 A graph showing the relationship between the number of iterations and the sum of the distances between each joint and the center of the cluster to which it belongs when the number of clusters is 2 in an optional embodiment of the present application;

[0055] Figure 7 A graph showing the relationship between the number of iterations and the sum of the distances between each joint and the center of the cluster to which it belongs when the number of clusters is 3 in an optional embodiment of the present application;

[0056] Figure 8 A graph showing the relationship between the number of iterations and the sum of the distances between each joint and the center of the cluster to which it belongs when the number of clusters is 4 in an optional embodiment of the present application;

[0057] Fig. 9 A graph showing the relationship between the number of iterations and the sum of the distances between each joint and the center of the cluster to which it belongs when the number of clusters is 5 in an optional embodiment of the present application;

[0058] Fig.10 A block diagram of a rock joint dominant occurrence search device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0060] From the perspective of rock engineering, joints make the rock mass lose its original continuity and solid integrity, further changing the permeability and engineering performance of the rock mass. They are essential information for determining the layout of chambers and the direction of excavation in engineering.

[0061] Joints are structural features produced in a certain part of the earth's crust under the action of a certain tectonic stress field at a certain period of time. By understanding their regularity, we can infer the tectonic stress field and tectonic movement mode when they were formed, and can better reflect the structural characteristics. The occurrence is the primary characteristic of joints, which determines the spatial development form of joints. Therefore, for joints with small development scale, large number of groups, and complex data, statistically analyzing their occurrence distribution and obtaining the dominant occurrence of joints can reflect the development law of regional rock mass joints, which is of great significance for rock excavation engineering.

[0062] Traditionally, the pole isodensity map is used to determine the dominant joint occurrence. Based on the pole map, contour lines are used to represent the characteristics of the pole density distribution in the map. The pole density concentration area reflects the preferred orientation of the structural development. This method mainly relies on artificial point selection to determine the dominant joint occurrence. On the one hand, it is not objective, and on the other hand, it is not accurate enough. The obtained joint dominant occurrence is an interval value.

[0063] Mahtab and Shanley first introduced clustering algorithms into the calculation of joint advantage. From the perspective of clustering technology, density-based clustering requires users to give a density search radius, which is not easy to determine; clustering based on agglomerative hierarchy starts with points as individual clusters, and each step merges two closest clusters, requiring the proximity of selected clusters; the partition-based K_means method requires users to set the number of clusters K, which can usually be obtained by observing the density map of poles. Therefore, due to the advantages of easy parameter assignment, simple implementation, and fast convergence, the partition-based K_means method is widely used in the determination of joint advantage. However, the K_means method has three disadvantages: it is sensitive to the selection of initial cluster centers, it is easy to converge to a local optimal solution, and the number of clusters K needs to be preset, that is, the clustering results are random, and continuous attempts are required to achieve better results. In order to solve the first two drawbacks mentioned above, many researchers have introduced particle swarm optimization algorithm, firefly algorithm, genetic algorithm and other algorithms to optimize K_means algorithm when using K_means to cluster joint occurrence analysis, and have achieved positive results. However, the algorithm is not stable enough, and the joints obtained by multiple operations have large differences in their advantageous occurrences. That is, the above algorithms are effective in helping K_means get rid of local optimality, and the global optimization effect is general. For the preset number of clusters K, which requires experts to judge the possible number of advantageous occurrences of regional rock joints, foreign scholars use Ray-Turi index, Xie-Beni index and other indicators to evaluate the clustering results obtained by different K values, and have achieved good results.

[0064] On the other hand, there are still two drawbacks in analyzing the whole process of obtaining the joint advantage state by the K_means method: first, abnormal joint occurrence values ​​are not eliminated. Since the joint occurrence data are usually obtained by surveyors in the field using a geological compass, human errors cannot be avoided, and the causes of joint formation are complicated. If abnormal joint values ​​are not eliminated, the final clustering effect cannot correctly obtain the joint advantage state; second, the cluster center is a calculated virtual center. After the K_means algorithm divides a certain category, the mean is calculated for the joints of that category, thereby determining the cluster center of the joints of that category, which means that the joint advantage state corresponding to the cluster center is not necessarily the real joint occurrence. The result of the joint occurrence clustering analysis is the beginning of understanding the spatial distribution of regional joints. The virtual joint advantage state is unacceptable for the subsequent analysis of regional joint distribution.

[0065] According to the problems existing in the above prior art, it can be seen that the purpose of a rock joint dominant occurrence search method provided in the embodiment of the present application is to:

[0066] (1) Before clustering the joint occurrence data, preprocessing is performed to screen and remove abnormal values ​​of joint occurrence;

[0067] (2) Combining the K-means clustering algorithm with the cuckoo search algorithm, the algorithm is applied to the clustering analysis of joint advantage occurrence to enhance the global search capability;

[0068] (3) For each iteration, the cluster center of the joint occurrence is replaced by the adjacent real joint occurrence.

[0069] Please refer to Figure 1 , which shows a flow chart of a method for searching for the dominant occurrence of rock joints provided in an embodiment of the present application. The method may include the following steps:

[0070] Step 101, obtaining n groups of joint attitude measurement data, and converting the measurement data to obtain a measurement data matrix in a Cartesian coordinate system.

[0071] The measurement data include dip data and dip angle data. The dip data and dip angle data of n groups of joint attitudes are obtained by measurement, and coordinate conversion is performed to convert them from polar coordinates to Cartesian coordinates into a measurement data matrix.

[0072] In the embodiment of the present application, the joint occurrence (measurement data) includes the dip data θ (0≤θ≤360°) and the dip data Two types of data, n groups of joint occurrences form a matrix A (n × 2) composed of Figure 2 As shown, the joint orientation can be expressed by its unit normal vector e i =(x i ,y i ,z i ) is represented as shown in the first formula (Formula 1). According to Formula (1), the matrix A is converted into a matrix B (n×3), with n rows and 3 columns;

[0073] The first formula can be expressed as:

[0074]

[0075] Where i represents the occurrence of the i-th group of joints, θ i Indicates the inclination data (0≤θ≤360°), Indicates the inclination data x i ,y i ,z i They respectively represent the x-axis, y-axis, and z-axis coordinates of the unit normal vector of the i-th group of joints.

[0076] Step 102: Calculate outliers in the measurement data matrix using the average arc length method and remove outliers.

[0077] In the embodiment of the present application, the average arc length normal unit vector spherical minimum arc length d is used to measure the similarity between joints, as shown in the second formula (Formula 2), the arc length distance between each joint in the matrix B and other joints is calculated to form a matrix C (n×n), and the average value of each row in the matrix C is calculated. Construct a column vector D (n×1), calculate the mean μ and standard deviation σ of all elements of vector D, and then select the row numbers of the elements in column vector D that are greater than μ+3σ, which are the outliers in matrix B, and delete the outliers in matrix A. Calculate the arc length distance between each group of joints and other joints by the second formula, which specifically includes:

[0078] d(e i ,e j ) = arccose i ·e j T | (2)

[0079] Among them, e i represents the unit normal vector of the i-th group of joints, e j represents the unit normal vector of the jth group of joints, d(e i ,e j ) represents the arc length distance between the ith group of joints and the jth group of joints.

[0080] Step 103, using the measurement data matrix after outliers are removed as the data source, setting the number of cluster centers and the number of populations to construct an initial nest matrix and the corresponding objective function value.

[0081] Among them, the objective function value is used to measure the overall clustering effect of joints.

[0082] Step 1031, in the embodiment of the present application, the number of cluster centers K and the population size P are set, the B matrix that eliminates outliers is used as the data source, the data value range is determined, data is randomly generated within the data range, and an initial nest matrix E0 (K×3×P) is constructed. Each cluster center has 3 coordinates, K cluster centers form a group of solutions, and the initial nest contains P groups of possible solutions.

[0083] Step 1032, objective function value calculation, calculate the distance between each joint in the n groups of joints and the K initial cluster centers, then divide the joints into the nearest cluster centers to complete the initial division of joints into K categories; then calculate the sum of the distances between each joint in the category and the cluster center to which it belongs, and add up the total distances of the K categories, which is defined as the objective function F, as shown in Formula 3. When the K value is determined, the F value measures the overall clustering effect of the joints. The smaller the F value, the more compact the category, and the better the clustering effect. Set the E0 objective function value to be the column vector K(P×1), and the minimum value F in F0 is 0_min The corresponding number of rows of best is the number of pages of the best nest in the initial nest matrix, that is, E0_best (;,;,best), and retain.

[0084]

[0085] Among them, K represents the number of cluster centers, S j represents the unit normal vector of the joint contained in the jth class, c j represents the coordinate value of the jth cluster center, e i represents the unit normal vector of the i-th group of joints;

[0086] About E 0_best (;,;,best), best is a number corresponding to the minimum value F in the column matrix F0 0_min The number of rows; at the same time, E0 is a three-dimensional matrix, rows (K), columns (3), pages (P), where P represents the number of populations, possible solutions; E 0_best (;,;, best) is the nest corresponding to the minimum value, that is, the best nest, the best solution.

[0087] Step 1033, optimal nest neighbor replacement, calculate the relationship between each joint in n groups of joints and the optimal nest E 0_best The arc length of each cluster center in (;,;,best) is replaced by the joint closest to the cluster center and the cluster center to form E 0_best_replace (;,;, best), and recalculate F 0_min_replace .

[0088] Step 104, using the Levy flight operation in the cuckoo algorithm to iterate the initial nest matrix, and processing according to the preset number of iterations and the nest abandonment ratio to obtain the optimal nest matrix and the corresponding objective function value.

[0089] In the embodiment of the present application, the joint dominant occurrence is searched iteratively. The K-means clustering algorithm is combined with the cuckoo search algorithm and applied to the joint occurrence clustering; firstly, the number of iterations N and the nest abandonment ratio γ are set.

[0090] like Figure 3 , the specific steps of the iterative search process include:

[0091] ①Enter Nest E t , E t_best , in E t Based on E t_best The position of the new nest matrix E is calculated by using the Levy flight path to iterate the step length and direction. t+1 (K×3×P), as shown in formula (4)

[0092]

[0093] Among them, α represents the step size scaling factor, which is generally 1, Levy (β) represents the Levy flight path, β represents the flight factor value range (1,3), which is generally 1.5. The Levy flight path adopts Mantegna's method and uses normal distribution to generate random step sizes that obey the Levy distribution, as shown in formula (5):

[0094]

[0095] Among them, Levy(β) represents the Levy flight path, β represents the range of flight factor, E t represents the nest matrix of the current training batch, E t_best represents the optimal nest matrix of the current training batch; v represents a random variable that satisfies the standard normal distribution, μ represents a random variable that satisfies the normal distribution, and their mathematical expectations are both 0, and the standard deviations of v and μ are 1 and σ respectively.

[0096] ②Enter Nest E t+1 Repeat the above steps 1032 and 1033 to obtain E t+1_best , F t+1 , F t+1_min , compare F t+1_min With F t_min size, always use the smaller objective function value and the corresponding optimal nest, and update E t , F t , F t_min E′ t , F′ t , F′ t_min , that is, always keep the current minimum objective function value and the best nest to the next training batch;

[0097] ③ Regarding E′ t , randomly update the elements of 1-γ within the data range, γ is generally 0.25, and the remaining 25% remain the same, forming a regeneration nest matrix E′ t+1 ;

[0098] ④ Input nest E′ t+1 Repeat the above steps 1032 and 1033 to obtain E′ t+1_best , F′ t+1 , F′ t+1_min , compare F′ t_min and F′ t+1_min size, always use the smaller objective function value and the corresponding optimal nest, and update E′ t , F t ', F′ t_min E″ t , F t ”,F″ t_min , thus completing a cycle;

[0099] ⑤ Repeat steps ① to ④ until the number of iterations N is met, stop the iteration, and get the optimal cluster center matrix E N_best (K×3) is the real joint with the smallest F value within the number of iterations.

[0100] Step 105, determining the optimal number of cluster centers and the corresponding dominant occurrence of rock joints by measuring the clustering effect of the optimal nest matrix.

[0101] In the embodiment of the present application, it specifically includes determining the optimal K value. The Ray-Turi index υ is used to measure the clustering effect for different K values. The smaller the index υ, the better the clustering effect. The K corresponding to the smallest index v is selected as the best classification. The index υ is shown in formula (6):

[0102]

[0103] Among them, K represents the number of cluster centers, n represents the number of joints, S j represents the unit normal vector of the joints contained in the jth class, c i represents the coordinate value of the i-th cluster center, c j Represents the coordinate value of the jth cluster center.

[0104] Finally, after multiple calculations, the best classification K is selected, the corresponding cluster center coordinates are converted into polar coordinates, and the best classification effect diagram of the joint occurrence is drawn.

[0105] The following is an example of performing an actual search using the rock mass joint advantage occurrence search method of the present application:

[0106] (1) Select a rock mass in the field and use a geological compass to obtain sufficient data on the dip α and dip angle β of the outcrop joints to form a matrix A and draw its pole diagram, as shown in Figure 4 shown.

[0107] (2) The elements in matrix A are converted into Cartesian coordinates (x, y, z) to form matrix B. The similarity between joints is measured by the average arc length normal unit vector spherical minimum arc length d. The arc length distance between each joint and other joints is formed into matrix C (n×n). The average value of each row in matrix C is calculated to form column vector D (n×1), with mean μ and standard deviation σ. Then, the row numbers in column vector D whose elements are greater than μ+3σ are selected, which are the abnormal occurrence values ​​in matrix B: (345, 87) and (162, 77), as shown in Figure 5 As shown:

[0108] (3) Set the number of clusters K = 2, the number of iterations N = 500, the population size P = 25, and the abandonment coefficient γ = 0.25; set the number of clusters K = 3, 4, 5, and the relationship between the number of iterations and the sum of distances is as follows: Figure 6 , 7 , 8, and 9, the effectiveness indicators corresponding to different K values ​​are compared, and the calculation is performed multiple times, as shown in Table 1:

[0109] Table 1 Effectiveness indicators for different K values

[0110]

[0111] (4) According to the clustering effectiveness index, when the number of clusters K = 2, the joint occurrence clustering performance is better. At this time, the dominant joint occurrences are 299°∠31° and 80°∠75°. The second is when the number of clusters K = 5. At this time, the dominant joint occurrences are 83°∠66°, 233°∠78°, 275°∠67°, 342°∠62°, and 284°∠25°.

[0112] It can be seen that the beneficial effects of the rock joint dominant occurrence search method proposed in this application are:

[0113] (1) Based on the fact that the outliers of the joint occurrence are far from the average arc length of the other joints, the outliers are screened and eliminated;

[0114] (2) Based on the K-means clustering algorithm, the cuckoo search algorithm is used to perform cluster analysis on the dominant occurrence of joints to enhance the global search capability and algorithm robustness;

[0115] (3) The dominant joint occurrence in the best classification is replaced by the adjacent joint occurrence from the calculation center, and the effectiveness index Ray-Turi index is used to determine the best classification K, which enhances the representativeness of the dominant joint occurrence.

[0116] Please refer to Fig.10 , which shows a block diagram of a rock mass joint dominant occurrence search device 200 provided in an embodiment of the present application. Fig.10 As shown, the device 200 may include: an acquisition module 201 , a processing module 202 , a construction module 203 , a calculation module 204 and a determination module 205 .

[0117] The acquisition module 201 is used to acquire n groups of joint occurrence measurement data and convert the measurement data to obtain a measurement data matrix in a Cartesian coordinate system; the measurement data includes dip data and dip angle data;

[0118] Processing module 202, used to calculate outliers in the measurement data matrix using the average arc length method and remove outliers;

[0119] Construction module 203, used to use the measurement data matrix after outliers are removed as the data source, set the number of cluster centers K, construct the objective function, construct the initial nest matrix based on the population number P and calculate the corresponding objective function value, the objective function value is used to measure the overall clustering effect of the joints;

[0120] The calculation module 204 is used to iterate the initial nest matrix using the Levy flight operation in the cuckoo algorithm; and obtain the optimal nest matrix and the objective function value corresponding to the optimal nest matrix according to the preset number of iterations N and the nest abandonment ratio γ;

[0121] The determination module 205 is used to determine the optimal number of cluster centers and the corresponding dominant occurrence of rock joints by measuring the clustering effect of the optimal nest matrix.

[0122] For the specific definition of the rock mass joint advantageous occurrence search device, please refer to the definition of the rock mass joint advantageous occurrence search method in the above text, which will not be repeated here. Each module in the above rock mass joint advantageous occurrence search device can be implemented in whole or in part by software, hardware and their combination. The above modules can be embedded in or independent of the processor in the device in the form of hardware, or can be stored in the memory in the device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0123] In one embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned rock mass joint dominant occurrence search method are implemented.

[0124] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described 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.

[0125] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A method for searching for the dominant occurrence of rock joints, characterized in that: The method comprises: Obtaining n groups of joint attitude measurement data, and converting the measurement data to obtain a measurement data matrix in a Cartesian coordinate system; the measurement data includes dip data and dip angle data; The average arc length method is used to calculate the outliers in the measurement data matrix, and the outliers are eliminated; The measured data matrix after outliers are removed is used as the data source, the number of cluster centers K is set, the objective function is constructed, the initial nest matrix is ​​constructed based on the population number P, and the corresponding objective function value is calculated. The objective function value is used to measure the overall clustering effect of the joints; The initial nest matrix is ​​iterated by using the Levy flight operation in the cuckoo algorithm; the optimal nest matrix and the objective function value corresponding to the optimal nest matrix are obtained by processing according to the preset number of iterations N and the nest abandonment ratio γ; The optimal number of cluster centers and their corresponding dominant rock joint occurrences are determined by measuring the clustering effect of the optimal nest matrix. The method of calculating the outliers in the measurement data matrix by using the average arc length method includes: Calculating the arc length distance between each group of joints and other joints in the measurement data matrix, and calculating the average value to obtain an n×1 column vector; Calculate the mean μ and standard deviation σ of all elements in the column vector, and then select the row numbers of the elements in the column vector that are greater than μ+3σ, which are the outliers in the measurement data matrix; The arc length distance between each group of joints and other joints is calculated by the second formula, wherein the second formula specifically includes: of i ,and j )=arc i ·and j T Among them, e i represents the unit normal vector of the i-th group of joints, e j represents the unit normal vector of the jth group of joints, d(e i ,e j ) represents the arc length distance between the ith group of joints and the jth group of joints.

2. The method according to claim 1, wherein the measurement data comprises inclination data and dip angle data, characterized in that: The method of obtaining n groups of joint occurrence measurement data and converting the measurement data to obtain a measurement data matrix in a Cartesian coordinate system includes: The n groups of dip data and inclination data of joint occurrences form an n×2 matrix, and coordinate transformation is performed using the first formula to obtain a measurement data matrix, wherein the measurement data matrix is ​​an n×3 matrix; The first formula can be expressed as: Where i represents the occurrence of the i-th group of joints, θ i Indicates the tendency data, Indicates the inclination data, x i ,y i ,z i They respectively represent the x-axis, y-axis, and z-axis coordinates of the unit normal vector of the i-th group of joints.

3. The method according to claim 1, characterized in that: The objective function value F is expressed by a third formula, which specifically includes: Among them, K represents the number of cluster centers, S j represents the unit normal vector of the joints contained in the jth class, c j represents the coordinate value of the jth cluster center, e i represents the unit normal vector of the i-th group of joints.

4. The method according to claim 1, characterized in that: The step of iterating the initial nest matrix by using the Levy flight operation in the cuckoo algorithm includes: Input the nest matrix E of the current training batch t and the optimal nest matrix E t_best , according to the fourth formula, we can get the new nest matrix E of the next training batch t+1 , the fourth formula specifically includes: Among them, α represents the step size scaling factor, Levy (β) represents the Levy flight path, and β represents the flight factor, and its value range is (1, 3).

5. The method according to claim 4, characterized in that The Levy flight path adopts Mantegna's method, uses normal distribution to generate random steps that obey the Levy distribution, and obtains the Levy flight path according to the fifth formula, which includes: Among them, Levy(β) represents the Levy flight path, β represents the range of flight factor, E t represents the nest matrix of the current training batch, E t_best represents the optimal nest matrix of the current training batch, v represents a random variable that satisfies the standard normal distribution, μ represents a random variable that satisfies the normal distribution, and their mathematical expectations are both 0. The standard deviations of v and μ are 1 and σ respectively.

6. The method according to claim 1, characterized in that The method of determining the optimal number of cluster centers and the corresponding rock joint dominant occurrence by measuring the clustering effect of the optimal nest matrix includes: The Ray-Turi index υ is used to measure the clustering effect, and the optimal number of cluster centers is determined according to the sixth formula, which includes: Among them, K represents the number of cluster centers, n represents the number of joints, S j represents the unit normal vector of the joints contained in the jth class, c i represents the coordinate value of the i-th cluster center, c j Represents the coordinate value of the jth cluster center.

7. A rock mass joint dominant occurrence search device, characterized in that: The device comprises: An acquisition module is used to acquire n groups of joint occurrence measurement data, and convert the measurement data to obtain a measurement data matrix in a Cartesian coordinate system; the measurement data includes dip data and dip angle data; A processing module, used for calculating the outliers in the measurement data matrix by using the average arc length method, and removing the outliers; A construction module is used to use the measurement data matrix after outliers are removed as the data source, set the number of cluster centers K, construct the objective function, construct the initial nest matrix based on the population number P and calculate the corresponding objective function value, and the objective function value is used to measure the overall clustering effect of the joints; A calculation module is used to iterate the initial nest matrix using the Levy flight operation in the cuckoo algorithm; and obtain the optimal nest matrix and the objective function value corresponding to the optimal nest matrix according to the preset number of iterations N and the nest abandonment ratio γ; A determination module is used to determine the optimal number of cluster centers and their corresponding dominant rock joint occurrences by measuring the clustering effect of the optimal nest matrix; The method of calculating the outliers in the measurement data matrix by using the average arc length method includes: Calculating the arc length distance between each group of joints and other joints in the measurement data matrix, and calculating the average value to obtain an n×1 column vector; Calculate the mean μ and standard deviation σ of all elements in the column vector, and then select the row numbers of the elements in the column vector that are greater than μ+3σ, which are the outliers in the measurement data matrix; The arc length distance between each group of joints and other joints is calculated by the second formula, wherein the second formula specifically includes: of i ,and j )=arc i ·and j T Among them, e i represents the unit normal vector of the i-th group of joints, e j represents the unit normal vector of the jth group of joints, d(e i ,e j ) represents the arc length distance between the ith group of joints and the jth group of joints.

8. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the method for searching for the dominant occurrence of rock joints as described in any one of claims 1 to 6 is implemented.

Images