A method and system for analyzing rock mass crack correlation in tunnels

Abstract

The present invention belongs to the technical field of rock mass crack analysis in tunnels, and specifically discloses a method and system for correlation analysis of rock mass cracks in tunnels, the method comprising: respectively obtaining three-dimensional occurrence information of crack surfaces on the surface of a rock mass face to be predicted, and occurrence information of cracks in a borehole; inputting the obtained information into a trained crack correlation prediction model to obtain coplanar crack surface numbers; the present invention can not only more accurately obtain crack occurrence information by comprehensively analyzing digital feature attributes of cracks, but also conduct in-depth research on the correlation between cracks, thereby comprehensively improving the accuracy and reliability of tunnel engineering design and construction.

Description

Technical Field

[0001] The present invention relates to the technical field of rock mass crack analysis in tunnels, and in particular to a rock mass crack correlation analysis method and system in tunnels. Background Art

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

[0003] During tunnel construction and operation, the presence of rock cracks poses a potential threat to engineering structures, so in-depth understanding of the distribution and properties of cracks and their relationship with rock stability has become the core of research. The study of crack distribution can provide a scientific basis for tunnel design and engineering construction, thereby reducing engineering risks and improving engineering safety.

[0004] At present, the methods for obtaining the occurrence of cracks in tunnels mainly include: geological exploration, seismic survey and underground radar, as well as image machine vision and other technical means. Geological exploration obtains rock information of the strata around the tunnel through core sampling, drilling and other methods, thereby inferring the occurrence characteristics of the cracks. Seismic survey uses the characteristics of seismic waves propagating underground to infer the stratum structure and existence of cracks around the tunnel by recording and analyzing the reflection and refraction of seismic waves. Underground radar transmits electromagnetic waves underground and receives reflected signals to obtain information on changes in the strata and cracks; image machine vision mainly identifies the surface features of cracks through images and provides spatial information for underground cracks.

[0005] However, although traditional geological radar, seismic survey and other technologies can obtain some characteristic information of cracks, it is difficult to accurately obtain the three-dimensional shape and fine structure information of cracks, resulting in a relatively vague understanding of the nature of cracks. Secondly, some detection methods have poor transmittance to geological media and are easily affected by factors such as groundwater and rock type. They cannot obtain the correlation characteristics between different crack surfaces, affecting the accuracy of crack information. Image machine vision technology can only identify exposed surface crack information and lacks the exploration of the correlation between cracks inside the rock mass. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes a method and system for correlation analysis of rock cracks in tunnels. By correlating analysis through means such as image collection of tunnel surface and boreholes and tunnel crack model tests, the correlation of internal cracks in the rock mass in the tunnel is determined, thereby improving the safety of tunnel engineering construction.

[0007] In some embodiments, the following technical solutions are adopted:

[0008] A method for analyzing rock mass crack correlation in a tunnel, comprising:

[0009] The three-dimensional occurrence information of the fracture surface on the face of the rock mass to be predicted and the occurrence information of the fractures in the borehole are obtained respectively; the obtained information is input into the trained fracture correlation prediction model to obtain the coplanar fracture surface numbers;

[0010] The training process of the crack correlation prediction model is as follows:

[0011] Based on the tunnel face surface image and point cloud information, the tunnel face surface crack surface and three-dimensional occurrence information are obtained;

[0012] Based on the on-site borehole image, cracks in the borehole image are identified, and the occurrence information of the cracks is calculated, and the attribute representation of the discontinuous crack surface inside the rock body is obtained based on the occurrence information;

[0013] Perform fracture correlation analysis on the fracture surfaces and three-dimensional occurrence information on the tunnel face surface, as well as the attribute representation of the discontinuous fracture surfaces inside the rock mass, to determine whether different fracture surfaces belong to the same coplanarity data set; thereby obtaining a first data set of multiple coplanar discontinuous fracture surfaces;

[0014] Building a tunnel entity model according to the fracture data and drilling data actually collected on site, and obtaining a second data set of multiple coplanar discontinuous fracture surfaces through the tunnel entity model;

[0015] The first data set and the second data set are used as training data sets to train the fracture correlation prediction model.

[0016] Among them, based on the surface image and point cloud information of the tunnel face, the surface crack surface and three-dimensional occurrence information of the tunnel face are obtained, specifically: the surface image and point cloud information of the tunnel face are obtained, the trace cracks are identified by using the trained machine vision semantic segmentation model, and the trace cracks are fitted to obtain the surface crack surface and three-dimensional occurrence information of the tunnel face; the three-dimensional occurrence information at least includes: fracture surface inclination, inclination angle, trace length, center coordinates and tunnel face number.

[0017] Furthermore, the occurrence information of the fracture is calculated, specifically:

[0018] Based on the location of the borehole where the fracture is located and the radius of the borehole, the inclination and dip angle of the fracture are calculated;

[0019] According to the inclination angle of the crack, the plane normal vector of the plane equation corresponding to the crack surface is established;

[0020] The depth of the borehole where the fracture is located is taken as the depth of the fracture surface;

[0021] The fracture surface in the borehole fracture image is expanded into a sine curve along the edge of the borehole, and discrete points in the borehole fracture image are randomly extracted. The distance between the discrete points and the best fitting plane is fitted using a normal curve, and the standard deviation of the normal distance from the discrete points to the best fitting plane is calculated as the dispersion.

[0022] In other embodiments, the following technical solutions are adopted:

[0023] A rock mass crack correlation analysis system in a tunnel, comprising:

[0024] The data acquisition module is used to respectively acquire the three-dimensional occurrence information of the fracture surface of the rock face to be predicted and the occurrence information of the fractures in the borehole;

[0025] A coplanar prediction module is used to input the acquired information into the trained fracture correlation prediction model to obtain the coplanar fracture surface numbers;

[0026] The training process of the crack correlation prediction model is as follows:

[0027] Based on the tunnel face surface image and point cloud information, the tunnel face surface crack surface and three-dimensional occurrence information are obtained;

[0028] Based on the on-site borehole image, cracks in the borehole image are identified, and the occurrence information of the cracks is calculated, and the attribute representation of the discontinuous crack surface inside the rock body is obtained based on the occurrence information;

[0029] Perform fracture correlation analysis on the fracture surfaces and three-dimensional occurrence information on the tunnel face surface, as well as the attribute representation of the discontinuous fracture surfaces inside the rock mass, to determine whether different fracture surfaces belong to the same coplanarity data set; thereby obtaining a first data set of multiple coplanar discontinuous fracture surfaces;

[0030] Building a tunnel entity model according to the fracture data and drilling data actually collected on site, and obtaining a second data set of multiple coplanar discontinuous fracture surfaces through the tunnel entity model;

[0031] The first data set and the second data set are used as training data sets to train the fracture correlation prediction model.

[0032] In other embodiments, the following technical solutions are adopted:

[0033] A terminal device comprises a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store a plurality of instructions, wherein the instructions are suitable for being loaded by the processor and executing the above-mentioned rock mass fracture correlation analysis method in a tunnel.

[0034] In other embodiments, the following technical solutions are adopted:

[0035] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the above-mentioned rock mass fracture correlation analysis method in a tunnel.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] (1) The present invention integrates the data of cracks on the tunnel face surface and cracks in the borehole, and obtains three-dimensional occurrence information of the cracks. The tunnel crack surface is divided into a first data set of multiple coplanar continuous crack surfaces through theoretical analysis of coplanarity, which can fully explore the correlation between different crack surfaces. The results of the theoretical analysis are verified by constructing a tunnel entity model, providing experimental data for verifying the crack correlation theory, and providing a new way to further study the crack correlation. This helps to more comprehensively understand the relationship and evolution law between cracks, thereby improving the accurate assessment of the impact of cracks in tunnel design and construction.

[0038] At the same time, a second data set of multiple coplanar discontinuous crack surfaces of the tunnel entity model is constructed, and the first data set and the second data set are fused as a training set to train the crack correlation prediction model; the training data of the present invention combines the field collection data and the experimental model data, thereby improving the accuracy and reliability of the crack correlation.

[0039] (2) The present invention uses a machine vision semantic segmentation model to identify trace fractures, and fits the trace fractures to obtain the fracture surface and three-dimensional occurrence information of the face; uses a rock fracture recognition model to obtain the fractures in the borehole image, and calculates the occurrence information of the fractures; thus, the three-dimensional shape and fine structure of the fractures are accurately obtained, thereby improving the high-precision recognition of the fractures.

[0040] (3) The present invention realizes the automatic fitting of fracture correlation through the autonomous learning of deep learning model, reduces the burden of manual selection and labeling, and improves the efficiency of fracture correlation analysis. At the same time, it integrates different data sources and enhances the comprehensiveness and comprehensiveness of fracture correlation analysis.

[0041] (4) By comprehensively analyzing the digital characteristic attributes of the cracks, the present invention can not only obtain the crack occurrence information more accurately, but also conduct in-depth research on the correlation between the cracks, thereby comprehensively improving the accuracy and reliability of tunnel engineering design and construction. This helps to reduce engineering risks and improve engineering safety.

[0042] Other features and additional advantages of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1This is a flow chart of a method for analyzing rock mass fracture association in a tunnel according to an embodiment of the present invention;

[0044] Figure 2 Schematic diagram of the dispersion in an embodiment of the present invention;

[0045] Figure 3 A schematic diagram of a process for determining whether different fracture surfaces belong to the same coplanarity data set in an embodiment of the present invention;

[0046] Figure 4 Schematic diagram of coplanar discontinuous crack surfaces in an embodiment of the present invention;

[0047] Figure 5 This is a schematic diagram of fracture surface fitting in an embodiment of the present invention;

[0048] FIG6( a ) is a schematic diagram of a tunnel entity model segmented in an embodiment of the present invention;

[0049] FIG6( b ) is a schematic diagram of a three-dimensional tunnel model with crack distribution in an embodiment of the present invention;

[0050] FIG6( c ) is a schematic diagram showing the intersection of a borehole and a crack in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0052] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0053] Embodiment 1

[0054] In one or more embodiments, a method for analyzing rock mass fracture correlation in a tunnel is disclosed, combining Figure 1 , specifically including the following processes:

[0055] S101: Acquire the tunnel face surface image and point cloud information, use the trained machine vision semantic segmentation model to identify the trace cracks, and fit the trace cracks to obtain the tunnel face surface crack surface and three-dimensional occurrence information;

[0056] In this embodiment, the existing technology is used to identify the trace cracks using a machine vision semantic segmentation model.

[0057] The identified trace fractures are fitted with a disk model to obtain the fracture surface and three-dimensional occurrence of the face; the three-dimensional occurrence information includes: fracture surface inclination, dip angle, trace length, center coordinates, face number, etc. This process can be achieved by existing technology.

[0058] S102: Obtain an image of the borehole on site, use the rock mass fracture recognition model established by transfer learning to obtain fractures in the borehole image, calculate the occurrence information of the fractures, and obtain the attribute representation of the discontinuous fracture surface inside the rock mass based on the occurrence information;

[0059] The rock fracture recognition model of this embodiment can adopt the currently more mature semantic segmentation model, that is, use the deep learning semantic segmentation model to identify the borehole fractures. If it is applied to other tunnel projects, a transfer learning model is established to identify the borehole fractures.

[0060] The information on the occurrence of the calculated fractures specifically includes:

[0061] (1) Based on the borehole location and borehole radius of the fracture, calculate the inclination θ and dip angle α of the fracture;

[0062] θ=(x2 / 2πd)×360° (1)

[0063] α=tan -1 (h / x2-x1)=tan -1 (h / d) (2)

[0064] (2) Establishing the plane normal vector N of the plane equation corresponding to the fracture surface according to the inclination angle of the fracture;

[0065]

[0066] If a point P(x0, y0, z0) on a plane has a normal unit vector (A, B, C), the plane equation is as shown in formula (4). By breaking down formula (4), we can get D, as shown in formula (5);

[0067] cos(θ)sin(α)(x-x0)-sin(θ)sin(α)(y-y0)+cos(α)(z-z0)=0 (4)

[0068] D=-cos(θ)sin(α)x0+sin(θ)sin(α)y0-cos(α)z0 (5)

[0069] Among them, θ represents the inclination, α represents the dip angle, h and d represent the borehole position and borehole radius of the fracture, respectively, A, B, C, and D are the parameters of the plane equation Ax+By+Cz+D=0 corresponding to the fracture surface, and x1 and x2 are shown as follows Figure 2 As shown, the fracture surface in the borehole fracture image is expanded into a sine curve along the edge of the borehole, discrete points in the borehole fracture image are randomly extracted, and the distance between the discrete points and the best fitting plane is fitted using a normal curve.

[0070] (3) The depth of the borehole where the crack is located is taken as the depth H of the crack surface; the coordinates of the center of the crack surface are used to represent the crack position.

[0071] (4) The fracture surface in the borehole fracture image is expanded into a sine curve along the edge of the borehole, and discrete points in the borehole fracture image are randomly extracted. The distance between the discrete points and the best fitting plane is fitted using a normal curve, and the standard deviation σ of the normal distance from the discrete points to the best fitting plane is calculated as the dispersion.

[0072] The best fitting plane can be realized through MATLAB programming, specifically: when all discrete points to a certain plane are obtained by using the least squares method and all are minimum values, then the plane is the best fitting plane.

[0073] Because the borehole itself has a vertex angle ε and an azimuth angle κ, the coordinate system of the digital representation of each borehole fracture is different. Therefore, it is necessary to regularize the information through unified coordinates and scales to unify the structural plane coordinates into an absolute coordinate system. The specific coordinate conversion method is implemented using existing technology.

[0074] Finally, the digital attributes of each fracture surface are represented as follows: the center coordinates (x n ,y n , z n ), dip angles γ′ and α′, plane equation related parameters A′, B′, C′, D′, standard deviation σ of the normal distance from discrete points to the best fitting plane; these are all data after coordinate transformation, representing multiple fracture surfaces forming a discontinuous data set.

[0075] S103: performing fracture association analysis on the fracture surfaces and three-dimensional attitude information on the tunnel face surface and the attribute representation of the discontinuous fracture surfaces inside the rock mass to determine whether different fracture surfaces belong to the same coplanar data set; thereby obtaining a first data set of multiple coplanar discontinuous fracture surfaces; the first data set includes multiple groups of coplanar discontinuous fracture surfaces, the attitude information of all fracture surfaces in each group, and the attitude information of each group of convex polygonal fracture surfaces obtained by fitting.

[0076] The continuous discontinuity of the fracture surface may be affected by faults, thus reducing the lateral continuity; the internal fractures of the tunnel face are generally exposed by borehole television, but the fracture information obtained by borehole television is incomplete, especially the fracture size data of the fracture group is always unknown, but it is very important. The fractures exposed by different boreholes have spatial correlation, such as Figure 4As shown in the figure, two unconnected borehole fracture surface areas may be compatible with the same plane equation, but from a physical point of view, they should be classified as different planes. In order to balance the fragmentation of the plane and the accuracy of the surface approximation, this embodiment analyzes the continuous discontinuity surface that has a greater impact on tunnel safety.

[0077] Among them, combined Figure 3 , the process of judging whether different fracture surfaces belong to the same coplanarity dataset is as follows:

[0078] (1) Cluster all fracture surfaces according to the plane normal vector of each fracture surface to obtain a clustered data set;

[0079] Specifically, if the plane normal vectors N′ of two fracture surfaces are the same or similar and they can be considered to belong to the same cluster, they are divided into one category; however, the fracture surfaces in the same category are non-coplanar because the constant parameter D representing the distance to the origin in the fracture plane equation is different.

[0080] Each fracture surface is labeled with its category to obtain a clustered data set; each fracture surface in the clustered data set has the following digital attributes: dip angle (θ′α′), location coordinate information (x n ,y n ,z n ), the corresponding fracture surface equation parameters (A′, B′, C′, D′), the standard deviation σ of the normal distance from the discrete points to the best fitting plane, and the category.

[0081] (2) For two fracture surfaces in the same cluster category, calculate the absolute value of the difference in the equation parameters of the best fitting plane of the two fracture surfaces |D1-D2|, as well as the product of the sum of the standard deviations of the normal distances of all discrete points of the two fracture surfaces to the best fitting plane and the set sensitivity parameter k×(σ1+σ2); if it satisfies:

[0082] k×(σ1+σ2)≥|D1-D2|;

[0083] It is determined that the two fracture surfaces are coplanar and belong to the same coplanarity data set.

[0084] The coplanar fracture surfaces of different precisions can be divided by adjusting the k value; if k is set to 0, all clusters of the same discontinuity set are regarded as different planes; using a larger threshold will result in only the largest plane being detected, and planes with smaller differences in dip angle and dip direction cannot be distinguished; conversely, using a smaller threshold may result in the extraction of smaller planes and there is a risk of dividing the rock surface into too many parts; the sensitivity parameter k value is generally obtained based on actual engineering tests.

[0085] Next, we extract the occurrence parameters of the coplanar structural surface. The specific process is as follows:

[0086] (1) Fit all circle center coordinates in the same coplanar data set into a convex polygonal ring as the edge of the discontinuity surface.

[0087] The center coordinates of the fracture surface are used to determine its location. After obtaining the digital attributes of the fracture surface, such as Figure 5 As shown in Figure 3, by fitting the center coordinates of all crack surfaces in the coplanar data set into a convex polygon, the margins of the delineated discontinuity surface can be obtained.

[0088] This step can be implemented using the convex hull algorithm. The obtained convex hull is the largest convex polygon containing a given point set. The point set here is the point set composed of the center coordinates of all crack surfaces. Each center coordinate is an endpoint in the point set. The main steps are as follows:

[0089] ① Select the point with the smallest y coordinate in the point set as the starting point P0;

[0090] ② Sort the remaining points in ascending order according to their polar angles with P0;

[0091] ③ Traverse the sorted points one by one, and for each point, check the direction of the two line segments formed by it and the two top points of the stack; if it is a left turn (counterclockwise), push it into the stack; if it is a right turn (clockwise), pop it out of the stack until a left turn is formed or a convex hull is formed.

[0092] ④The points in the final stack are the points on the convex hull, and these points form the minimum convex polygon.

[0093] (2) Extract the occurrence data of the convex polygonal discontinuity surface.

[0094] After obtaining the convex hull, the discontinuity trace length is defined as the distance between the two endpoints with the largest Euler distance, and the distance between the two endpoints is continuously searched until the maximum value is obtained. The dip angle is the dip angle corresponding to any fracture surface in the absolute coordinate system, which is expressed by the formula:

[0095] Length = maxlength(Con(X ij )) (6)

[0096] dip direction=θ′=κ+θ (7)

[0097] dip angle = α′ = ε + α (8)

[0098] Among them, ε and κ represent the top angle and azimuth of the borehole, respectively, θ represents the dip, and α represents the inclination.

[0099] Length is the trace length of the coplanar crack surface, length(Con(X ij)) is the maximum value between the two endpoints of the convex hull, dip direction is the inclination, and dip angle is the inclination.

[0100] Use the cross product method to obtain the area of ​​a convex polygon. Assume that the vertices of the convex polygon are given in clockwise order, namely (x1, y1), (x2, y2), (x3, y3)…(x n ,y n ); Calculate the cross product of each edge, for each pair of adjacent vertices, assuming the current vertex is (x i ,y i ), the next vertex is (x i+1 ,y i+1 ), calculate the cross product:

[0101] cross product=(x i *y i+1 )-(x i+1 *y i ) (9)

[0102] Adding all the cross products and taking the absolute value gives the convex hull area, which can be expressed as:

[0103]

[0104] S104: Building a tunnel entity model according to the fracture data and drilling data actually collected on site, and obtaining a second data set of multiple coplanar discontinuous fracture surfaces through the tunnel entity model;

[0105] Specifically, the tunnel entity model of this embodiment includes a tunnel internal model, a tunnel surrounding rock part, and a simulated drilling part. In actual tunnel engineering excavation, the drilling and blasting method is mostly used. A new face will be obtained after each drilling and blasting. However, it is difficult to simulate the drilling and blasting process with existing materials. Therefore, this embodiment chooses to use segmented model splicing to realize the tunnel body production, which can realize the process of continuous exposure of the face as the excavation progresses; as shown in Figure 6(a). The model drilling distribution is set according to the drilling position of the actual tunnel, and a small diameter PVC round pipe is used to simulate the actual drilling process. The drilling hole can reflect the position of all cracks passing through the drilling hole.

[0106] Referring to Figure 6(b), the tunnel excavation process is simulated. According to the intersection relationship between the overall crack distribution and each tunnel face, the three-dimensional shape of each tunnel face and various cracks are produced. Each tunnel face has multiple intersection lines with the crack distribution model, thereby forming surface cracks and trace cracks on the tunnel face; referring to Figure 6(c), the model drilling distribution is set according to the actual tunnel drilling position to realize the visualization process of the intersection of the drilling and each crack.

[0107] After the model is made, the 3D shape and color information of the face is collected using a camera and a 3D laser scanner. To simulate the real tunnel excavation scene, after collecting one face, the model of this section of the face can be taken out of the overall model, and the next face is simulated to simulate the falling of rocks and the exposure of cracks, and the scanning is continued. This cycle is repeated until all faces are scanned, which means that the excavation of this tunnel is completed, and all the face data of a complete tunnel can be obtained.

[0108] This embodiment constructs a tunnel entity model, simulates the actual drilling process, and obtains the intersection of the borehole and each fracture, so as to obtain a second data set of multiple coplanar discontinuous fracture surfaces. The second data set also includes multiple groups of coplanar discontinuous fracture surfaces, the occurrence information of all fracture surfaces in each group, and the occurrence information of each group of fracture surfaces.

[0109] S105: Training a crack correlation prediction model based on the first data set and the second data set.

[0110] The fracture correlation prediction model can adopt structures such as convolutional neural networks or deep neural networks; the input of the fracture correlation prediction model is: the three-dimensional occurrence information of the fracture surface fitting on the tunnel face and the fracture surface information extracted from the borehole, including the inclination, dip angle, trace length, and center coordinates of the fracture surface; the output is: the number of the coplanar fracture surface. In the training stage, the label result is input, that is, the number of the coplanar fracture surface obtained by the coplanarity theory.

[0111] Embodiment 2

[0112] A rock mass crack correlation analysis system in a tunnel, comprising:

[0113] The data acquisition module is used to respectively acquire the three-dimensional occurrence information of the fracture surface of the rock face to be predicted and the occurrence information of the fractures in the borehole;

[0114] A coplanar prediction module is used to input the acquired information into the trained fracture correlation prediction model to obtain the coplanar fracture surface numbers;

[0115] The training process of the crack correlation prediction model is as follows:

[0116] Based on the tunnel face surface image and point cloud information, the tunnel face surface crack surface and three-dimensional occurrence information are obtained;

[0117] Based on the on-site borehole image, cracks in the borehole image are identified, and the occurrence information of the cracks is calculated, and the attribute representation of the discontinuous crack surface inside the rock body is obtained based on the occurrence information;

[0118] Perform fracture correlation analysis on the fracture surfaces and three-dimensional occurrence information on the tunnel face surface, as well as the attribute representation of the discontinuous fracture surfaces inside the rock mass, to determine whether different fracture surfaces belong to the same coplanarity data set; thereby obtaining a first data set of multiple coplanar discontinuous fracture surfaces;

[0119] Building a tunnel entity model according to the fracture data and drilling data actually collected on site, and obtaining a second data set of multiple coplanar discontinuous fracture surfaces through the tunnel entity model;

[0120] The first data set and the second data set are used as training data sets to train the fracture correlation prediction model.

[0121] The specific implementation of the above modules is the same as that in Example 1 and will not be described in detail.

[0122] Embodiment 3

[0123] In one or more embodiments, a terminal device is disclosed, including a server, the server including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the rock mass fracture correlation analysis method in the tunnel in Example 1 when executing the program. For the sake of brevity, it will not be described here.

[0124] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0125] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0126] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0127] Embodiment 4

[0128] In one or more embodiments, a computer-readable storage medium is disclosed, in which a plurality of instructions are stored, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the method for analyzing rock mass fracture association in a tunnel described in Example 1.

[0129] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A method for analyzing rock mass crack correlation in a tunnel, characterized in that: include: The three-dimensional occurrence information of the fracture surface on the face of the rock mass to be predicted and the occurrence information of the fractures in the borehole are obtained respectively; the obtained information is input into the trained fracture correlation prediction model to obtain the coplanar fracture surface numbers; The training process of the crack correlation prediction model is as follows: Based on the face surface image and point cloud information, the face surface crack surface and three-dimensional occurrence information are obtained, specifically: the face surface image and point cloud information are obtained, the trace cracks are identified by using the trained machine vision semantic segmentation model, and the trace cracks are fitted to obtain the face surface crack surface and three-dimensional occurrence information; the three-dimensional occurrence information at least includes: crack surface inclination, inclination angle, trace length, center coordinates and face number; Based on the on-site borehole image, cracks in the borehole image are identified, and the occurrence information of the cracks is calculated, and the attribute representation of the discontinuous crack surface inside the rock body is obtained based on the occurrence information; Perform fracture correlation analysis on the fracture surfaces and three-dimensional occurrence information on the tunnel face surface, as well as the attribute representation of the discontinuous fracture surfaces inside the rock mass, to determine whether different fracture surfaces belong to the same coplanarity data set; thereby obtaining a first data set of multiple coplanar discontinuous fracture surfaces; Building a tunnel entity model according to the fracture data and drilling data actually collected on site, and obtaining a second data set of multiple coplanar discontinuous fracture surfaces through the tunnel entity model; The first data set and the second data set are used as training data sets to train the fracture correlation prediction model.

2. A method for analyzing rock mass crack correlation in a tunnel according to claim 1, characterized in that: Calculate the occurrence information of the fracture, specifically: Based on the location of the borehole where the fracture is located and the radius of the borehole, the inclination and dip angle of the fracture are calculated; According to the inclination angle of the crack, the plane normal vector of the plane equation corresponding to the crack surface is established; The depth of the borehole where the fracture is located is taken as the depth of the fracture surface; The fracture surface in the borehole fracture image is expanded into a sine curve along the edge of the borehole, and discrete points in the borehole fracture image are randomly extracted. The distance between the discrete points and the best fitting plane is fitted using a normal curve, and the standard deviation of the normal distance from the discrete points to the best fitting plane is calculated as the dispersion.

3. A method for analyzing rock mass crack correlation in a tunnel according to claim 2, characterized in that: Specifically, the best fitting plane is: when the distances from all discrete points to a certain plane calculated by the least square method are all the smallest, then the plane is the best fitting plane.

4. A method for analyzing rock mass crack correlation in a tunnel according to claim 1, characterized in that: The occurrence information of the fractures is converted into a unified absolute coordinate system to obtain the property representation of the discontinuous fracture surface inside the rock mass, including: the coordinates of the center of the circle, the dip angle, the related parameters of the plane equation, and the standard deviation of the normal distance from the discrete points to the best fitting plane in the unified absolute coordinate system.

5. The method for analyzing rock mass crack correlation in a tunnel according to claim 1, characterized in that: Determine whether different fracture surfaces belong to the same coplanarity dataset, specifically: According to the plane normal vector of each fracture surface, all fracture surfaces are clustered to obtain a clustered data set; For two fracture surfaces in the same cluster category, calculate the absolute value of the difference in the equation parameters of the best fitting plane of the two fracture surfaces |D1-D2|, as well as the product of the sum of the standard deviations of the normal distances of all discrete points of the two fracture surfaces to the best fitting plane and the set sensitivity parameter k×(σ1+σ2); if it satisfies: k×(σ1+σ2)≥|D1-D2| It is determined that the two fracture surfaces are coplanar and belong to the same coplanarity data set.

6. A method for analyzing rock mass crack correlation in a tunnel according to claim 1, characterized in that: After obtaining a first data set of a plurality of coplanar discontinuous fracture surfaces, the method further includes: For all coplanar discontinuous crack surfaces in each first data set, their center coordinates are fitted into convex polygonal rings as the edges of the discontinuous surfaces.

7. A method for analyzing rock mass crack correlation in a tunnel according to claim 6, characterized in that: The process of fitting into a convex polygonal ring is specifically as follows: The coordinates of the centers of all fracture surfaces in each first data set form a point set; Select the point with the smallest y coordinate in the point set as the starting point P0, and sort the remaining points in ascending order according to their polar angles with P0; Traverse the sorted points one by one, and for each point, check the direction of the two line segments formed by it and the two points at the top of the stack; If it is counterclockwise, push it into the stack; if it is clockwise, pop it out of the stack until a left turn is formed or a convex hull is formed; The points in the final stack are the points on the convex hull. These points form the minimum convex polygon, and the cross product method is used to obtain the area of ​​the convex polygon.

8. A method for analyzing rock mass crack correlation in a tunnel according to claim 1, characterized in that: The tunnel entity model is obtained by splicing multiple segmented tunnel faces. According to the overall crack distribution and the intersection relationship between the cracks and each tunnel face, corresponding surface cracks and trace cracks are made on each segmented tunnel face, and the drilling distribution is simulated.

9. A method for analyzing rock mass crack correlation in a tunnel according to claim 1, characterized in that: The fracture correlation prediction model adopts a convolutional neural network or a deep neural network structure; the input of the fracture correlation prediction model is the three-dimensional occurrence information of the fracture surface on the surface of the rock mass face to be predicted, and the occurrence information of the fractures in the borehole, and the output is the coplanar fracture surface numbers.

10. A rock mass crack correlation analysis system in a tunnel, characterized in that: include: The data acquisition module is used to respectively acquire the three-dimensional occurrence information of the fracture surface of the rock face to be predicted and the occurrence information of the fractures in the borehole; A coplanar prediction module is used to input the acquired information into the trained fracture correlation prediction model to obtain the coplanar fracture surface numbers; The training process of the crack correlation prediction model is as follows: Based on the face surface image and point cloud information, the face surface crack surface and three-dimensional occurrence information are obtained, specifically: the face surface image and point cloud information are obtained, the trace cracks are identified by using the trained machine vision semantic segmentation model, and the trace cracks are fitted to obtain the face surface crack surface and three-dimensional occurrence information; the three-dimensional occurrence information at least includes: crack surface inclination, inclination angle, trace length, center coordinates and face number; Based on the on-site borehole image, cracks in the borehole image are identified, and the occurrence information of the cracks is calculated, and the attribute representation of the discontinuous crack surface inside the rock body is obtained based on the occurrence information; Perform fracture correlation analysis on the fracture surfaces and three-dimensional occurrence information on the tunnel face surface, as well as the attribute representation of the discontinuous fracture surfaces inside the rock mass, to determine whether different fracture surfaces belong to the same coplanarity data set; thereby obtaining a first data set of multiple coplanar discontinuous fracture surfaces; Building a tunnel entity model according to the fracture data and drilling data actually collected on site, and obtaining a second data set of multiple coplanar discontinuous fracture surfaces through the tunnel entity model; The first data set and the second data set are used as training data sets to train the fracture correlation prediction model.

11. A terminal device, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the method for correlation analysis of rock fractures in a tunnel as described in any one of claims 1-10.

12. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instructions are suitable for being loaded by a processor of a terminal device and executing the method for analyzing rock mass fracture association in a tunnel as described in any one of claims 1 to 9.

Images