CN114742937B - Three-dimensional geological analysis method and device - Google Patents

Abstract

The invention discloses a three-dimensional geological analysis method and device, belonging to the field of design and construction of underground space shield tunnels, the method comprising: S1: obtaining a data frame D corresponding to the subdivided geological exploration drilling data; D[:, 1], D[:, 2] and D[:, 3] are three-dimensional coordinate data of the geological exploration drilling; D[:, 4] is soil layer data; S2: using the three-dimensional coordinate data to perform drilling point cloud modeling, draw a geological shield section model, and mesh to obtain a mesh model M _dm '; calculating the minimum distance between the coordinates of non-repeated grid nodes in M _dm ' and the drilling point cloud model, analyzing the soil layer type and calculating the thickness of each soil layer to obtain the cross-section soil layer distribution; S3: analyzing the soil layer data D[:, 4] to obtain the soil layer type and assigning D[:, 5]; using the three-dimensional coordinate data and D[:, 5] as training samples to train the interpolation algorithm model; using the trained model to predict the soil layer parameter distribution of each grid node. The method provided by the invention can improve the comprehensiveness and accuracy of geological analysis.

Description

A three-dimensional geological analysis method and device

Technical Field

The present invention belongs to the field of design and construction of underground space shield tunnels, and more specifically, relates to a three-dimensional geological analysis method and device.

Background technique

Shield tunnels in underground spaces are distributed over long distances, usually passing through ultra-long soft and hard composite strata, including clay, sand, strong wind mudstone, gravel layers, unfavorable rock bursts, joint-dense zones, altered rock fracture zones, etc., so tunnel construction is prone to many unfavorable factors. For example, for high-strength (>100Mpa) rock and silt combined strata, shield machines have problems such as failure to dig, easy instability, and difficulty in controlling excavation parameters; when encountering unexpected hard rock, shield machines have difficulty breaking rocks, low excavation efficiency, short main bearing life, and severe tool wear. When shuttling through active plate areas or areas with high ground stress, large deformation of soft rocks, and strong rock bursts of hard rocks, tunnel soft rock deformation control and hard rock rock burst prevention need to be particularly strengthened. Therefore, when designing and constructing shield tunnels, it is usually necessary to fully understand the structure of the composite strata in the area in advance and take timely countermeasures.

Geological surveys usually use deterministic methods such as drilling and sampling to understand the geological environment, and use borehole samples and survey data as the only accurate basis. However, due to the limited density of boreholes and the variability of rock and soil during the construction process, geologists' analysis of geology is often not accurate enough, resulting in major mistakes in design and construction and even safety accidents. In fact, at this stage, many scholars at home and abroad have adopted prediction methods such as DSI, stochastic simulation, and geological statistics to calculate stratum boundaries and elevations, and proposed the Universal Kriging method to infer the uncertainty of geotechnical physical property indicators. Although the prediction results have improved the accuracy of geological assessment, the relevant surveyors still abstractly project the three-dimensional space of the geological body into a two-dimensional plane. People involved in shield design and construction can only understand the geological environment based on abstract two-dimensional plane maps.

Therefore, it is necessary to improve the comprehensiveness and accuracy of geological analysis and analyze the geological distribution that is unfavorable or unfavorable to shield engineering.

Summary of the invention

In view of the above defects or improvement needs of the prior art, the present invention provides a three-dimensional geological analysis method and device, the purpose of which is to analyze the three-dimensional geology based on geological exploration drilling data, so as to carry out shield tunnel design and construction, thereby solving the technical problems of single geological analysis data and low accuracy in the existing shield tunnel design and construction process.

To achieve the above object, according to one aspect of the present invention, a three-dimensional geological analysis method is provided for shield tunnel design and construction, comprising:

S1: Obtain the data frame D corresponding to the subdivided geological exploration drilling data; the first column D[:,0] of D is the sequence number of the geological exploration drilling hole; the second column D[:,1], the third column D[:,2] and the fourth column D[:,3] of D are the three-dimensional coordinate data of the geological exploration drilling hole; the fifth column D[:,4] of D is the soil layer data of the geological exploration drilling hole, which is used to analyze the type of soil layer and assign geotechnical parameters to each soil layer to obtain the sixth column D[:,5] of D;

S2: Use the three-dimensional coordinate data {D[:,1], D[:,2], D[:,3]} of the geological exploration borehole to perform borehole point cloud modeling to obtain a borehole point cloud model P; draw a geological and shield section model _Mdm based on the borehole point cloud model P, and mesh _Mdm to obtain a new mesh model _Mdm '; extract the coordinate data frame V of all non-repeated mesh nodes in _Mdm ', the i-th row of V is the vertex coordinate (x _i y _i z _i ) of the i-th mesh node, and calculate the minimum distance between the coordinates of each mesh node and the borehole point cloud model P; use the minimum distance as an indicator to analyze the maximum likelihood soil layer type SoilType to which each mesh node belongs to generate a soil layer type information data frame V'=(V SoilType); use the soil layer type information data frame V' to calculate the average thickness of each soil layer included in the geological and shield section model _Mdm to obtain the cross-sectional soil layer distribution;

S3: The three-dimensional coordinate data of the geological exploration borehole {D[:,1], D[:,2], D[:,3]} and the geotechnical parameters D[:,5] of each soil layer are used as training samples to train the interpolation algorithm model to obtain a target model; the target model is used to predict the soil layer parameter distribution of each grid node.

In one embodiment, the soil layer data D[:,4] of the geological exploration borehole is analyzed in S1 to obtain the type of soil layer, including:

The soil layer data D[:,4] of the geological exploration borehole are divided into soil layers according to a division scale to obtain soil layer types; wherein the division scale is smaller than the elevation difference of the interface between the upper soil layer and the lower soil layer; and the average thickness of each soil layer is the average of three or more thickness values corresponding to the same soil layer.

In one of the embodiments, the cross-sectional soil layer distribution in S2 is used for segment design of a shield tunnel.

In one embodiment, the geological and shield section model _Mdm in S2 includes a solid object Solid and a boundary object Brep.

In one embodiment, the S2 uses the minimum distance as an indicator to analyze the maximum likelihood soil type SoilType of each grid node to generate a soil type information data frame V'=(V SoilType), including:

Taking the minimum distance of each grid node as the evaluation index, the KNN algorithm is used to analyze the maximum likelihood soil type SoilType of each grid node, and generate the soil type information data frame V'.

In one of the embodiments, the soil layer parameter distribution in S3 is used to predict the geotechnical parameter distribution of the tunnel face during shield construction.

In one embodiment, the interpolation algorithm model in S3 is a three-dimensional Universal Kriging interpolation algorithm model.

In one embodiment, the method further comprises:

S4: Visualizing the cross-section soil layer distribution and the soil layer parameter distribution.

In one of the embodiments, the S4 includes: performing vertex shading and fuzzy rendering in sequence to perform three-dimensional visualization processing on the cross-section soil layer distribution and the geotechnical parameter distribution.

According to another aspect of the present invention, a three-dimensional geological analysis device is provided, which is applied to shield tunnel design and construction, and the three-dimensional geological analysis device is used to execute the three-dimensional geological analysis method.

In general, the above technical scheme conceived by the present invention has the following beneficial effects compared with the prior art: 1. Based on the mileage parameters, the stratum distribution of the tunnel surrounding rock can be automatically analyzed by using the program command flow; 2. Based on the mileage parameters, the soil parameter distribution of the excavation face can be automatically analyzed by using the program command flow; 3. The above results can be visualized by using the graphic rendering method, and a more intuitive data display effect can be brought to the design and construction personnel.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a three-dimensional geological analysis method according to an embodiment of the present invention;

FIG2 is a horizontal distribution diagram of geological exploration boreholes according to an embodiment of the present invention;

FIG3 is a schematic diagram of a geological exploration borehole point cloud model according to an embodiment of the present invention;

FIG4 is a schematic diagram of a shield section model after meshing in one embodiment of the present invention;

FIG5 is a schematic diagram of a geological section model after meshing in one embodiment of the present invention;

FIG6 is a geological cross-section stratum distribution diagram after analysis in one embodiment of the present invention;

FIG7 is a spatial distribution diagram of geotechnical parameters of a shield section after analysis in one embodiment of the present invention;

FIG8 is a probability distribution diagram of geotechnical parameters of a shield section after analysis in one embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

The present invention provides a three-dimensional geological analysis method for shield tunnel design and construction, comprising:

S1: Get the data frame D corresponding to the subdivided geological exploration drilling data; the first column D[:,0] of D is the sequence number of the geological exploration drilling hole; the second column D[:,1], the third column D[:,2] and the fourth column D[:,3] of D are the three-dimensional coordinate data of the geological exploration drilling hole; the fifth column D[:,4] of D is the soil layer data of the geological exploration drilling hole, which is used to analyze the type of soil layer and assign geotechnical parameters to each soil layer to obtain the sixth column D[:,5] of D;

S2: Use the three-dimensional coordinate data {D[:,1], D[:,2], D[:,3]} of the geological exploration borehole to perform borehole point cloud modeling to obtain a borehole point cloud model P; draw the geological and shield section model _Mdm based on the borehole point cloud model P, and mesh _Mdm to obtain a new mesh model _Mdm '; extract the coordinate data frame V of all non-repeated mesh nodes in _Mdm ', the i-th row of V is the vertex coordinate (x _i y _i z _i ) of the i-th mesh node, and calculate the minimum distance between the coordinates of each mesh node and the borehole point cloud model P; use the minimum distance as an indicator to analyze the maximum likelihood soil layer type SoilType to which each mesh node belongs to generate a soil layer type information data frame V'=(V SoilType); use the soil layer type information data frame V' to calculate the average thickness of each soil layer included in the geological and shield section model _Mdm to obtain the cross-sectional soil layer distribution;

S3: The three-dimensional coordinate data of the geological exploration borehole {D[:,1], D[:,2], D[:,3]} and the geotechnical parameters D[:,5] of each soil layer are used as training samples to train the interpolation algorithm model to obtain the target model; the target model is used to predict the distribution of soil layer parameters of each grid node.

As shown in FIG1 , the three-dimensional geological analysis method of the present invention is specifically as follows:

1) Subdivide the geological exploration drilling data according to the elevation and equidistance and form a data frame D after information collation;

2) Extract the subdivided geological exploration borehole three-dimensional coordinate points (D[:,1], D[:,2], D[:,3]) and import them into the three-dimensional modeling software for drilling point cloud modeling P;

3) Extract the subdivided geological survey borehole soil layer data D[:,4], analyze the total number of soil layer types, and set labels for each soil layer;

4) According to the design reference value of soil layer parameters (e.g. natural gravity γ) in the geological survey report, each soil layer is assigned corresponding parameters as D[:,5];

5) Draw the corresponding geological and shield section model _Mdm on the existing drilling point cloud model P according to the requirements;

6) Meshing the established model M _dm to obtain a new mesh model M _dm ';

7) extracting all non-repeated grid node coordinates V of the grid model M _dm ';

8) Calculate the minimum distance between each grid node V and the borehole point cloud model P in turn, infer the maximum likelihood soil layer type of each grid node through the KNN algorithm, assign corresponding labels, and generate V' with soil layer information labels;

9) If it is used for shield tunnel segment design, it is necessary to calculate the elevation range of the Z coordinate V’[:,3] under each soil layer in V’, and finally output the average thickness and arrangement order of each soil layer under the section;

10) If it is used to predict the distribution of geotechnical parameters of the tunnel face in shield construction, the drilling data (D[:,1], D[:,2], D[:,3], D[:,5]) are extracted as training samples and then imported into the UniversalKriging interpolation algorithm model, and the trained model is used to predict the distribution of geotechnical mechanical parameters of the section;

11) The cross-section soil layers and physical property parameter distribution are displayed in three dimensions using visualization methods.

In step 1), the acquisition format of the original geological survey data is as follows:

Table 1 Format of geological exploration drilling data

Among them, _Dzk [:,0] represents the data of the first column of all rows in _Dzk , and _Dzk [:,n] represents the data of the n-1th column of all rows in _Dzk , and the same applies below. For _Dzk , the same geological survey data borehole has different soil layer distributions. The elevation of the upper and lower interfaces of each soil layer under each borehole and the soil layer code are filled in from top to bottom. After the parameters in Table 1 are finely divided into units, they finally form a data frame (DataFrame) D with n rows and 5 columns, namely:

Among them, the subdivision of the original geological exploration borehole data D _zk is to generate points with more regular and dense distribution and soil layer codes in the borehole to finally form new geological exploration borehole data D. The difference between D and D _zk is mainly in the number of rows, k<<n. The specific algorithm is shown in Table 2:

Table 2

In step 2), D[:,1], D[:,2], D[:,3] are the 2nd, 3rd, and 4th columns of all rows in the data frame D, that is, each set of x, y, and z coordinates. The 3D modeling software is any 3D modeling software with command flow drawing or programming drawing functions. The point cloud modeling P is a point cloud model automatically generated by directly inputting x, y, and z coordinates, where x, y, and z are stored in the model in the format of n rows and 3 columns of the data frame, that is:

In step 3), the soil layer number analysis and soil layer type calibration algorithm are shown in Table 3:

table 3

In step 4), the algorithm for assigning the corresponding soil layer parameter design reference value to each soil layer type is shown in Table 4:

Table 4

In step 5), the geological and shield section model _Mdm is a three-dimensional digital model object represented by the Brep boundary in the modeling software.

In step 6), the mesh model M _dm ' is a mesh model object represented by Mesh in the modeling software.

In step 7), the mesh node coordinates V come from the coordinates of the mesh vertex Vertex in the Mesh object.

In step 8), the label V' is an n-row 4-column data frame with soil layer type information added after the grid node V, that is:

In step 8), the KNN algorithm is used to analyze the maximum likelihood soil layer type to which each grid node belongs. The specific description is shown in Table 5:

table 5

In step 9), the algorithm for determining the thickness of the soil layer in the cross section is shown in Table 6:

Table 6

In step 10), the training of the three-dimensional Universal Kriging interpolation algorithm model requires the input of parameters D[:,1], D[:,2], D[:,3], D[:,5], variogram_model = "hole-effect", drift_terms = ["regional_linear"]. After training the above model, the distribution of geotechnical parameters in the section is predicted by the algorithm shown in Table 7:

Table 7

In step 11), the visualization means is to use a graphics rendering software with command stream operation for visualization.

In one of the embodiments, the cross-sectional soil layer distribution in S2 is used for segment design of a shield tunnel.

In one embodiment, the geological and shield section model _Mdm in S2 includes a solid object Solid and a boundary object Brep.

In one embodiment, in S2, the maximum likelihood soil type SoilType to which each grid node belongs is analyzed using the minimum distance as an indicator to generate a soil type information data frame V'=(V SoilType), including:

Taking the minimum distance of each grid node as an indicator, the KNN algorithm is used to analyze the maximum likelihood soil type SoilType of each grid node, and generate the soil type information data frame V'.

In one of the embodiments, the soil layer parameter distribution in S3 is used to predict the geotechnical parameter distribution of the tunnel face during shield construction.

In one embodiment, the soil layer data D[:,4] of the geological exploration borehole is analyzed in S1 to obtain the types of soil layers, including:

The soil layer data D[:,4] of the geological exploration borehole are divided into soil layers according to the division scale to obtain the soil layer types; wherein the division scale is smaller than the elevation difference of the interface between the upper soil layer and the lower soil layer; the average thickness of each soil layer is the average value of more than three thickness values corresponding to the same soil layer.

In one embodiment, the interpolation algorithm model in S3 is a three-dimensional Universal Kriging interpolation algorithm model.

In one embodiment, the method further comprises:

S4: Visualize the cross-section soil layer distribution and soil layer parameter distribution.

In one embodiment, S4 includes: performing vertex shading and fuzzy rendering in sequence to perform three-dimensional visualization processing on the cross-section soil layer distribution and geotechnical parameter distribution.

For example, the geological survey data comes from the geological survey report of a city's underground space shield project. The data of 62 boreholes in the shield section between the Provincial Museum starting well and the Qinyuan East Road receiving well are selected as shown in Table 8, and the horizontal arrangement of the boreholes is shown in Figure 2.

Table 8 Original geological exploration drilling data (unit: meter) After the original geological exploration drilling data is divided into units: (assuming that the size of the division unit is 0.5)

Table 9: Geological exploration drilling data after unit division (unit: meter) The parameters in the above table eventually form an n-row 5-column data frame D, namely:

Step 2: Import the data columns D[:,1], D[:,2], and D[:,3] of the data frame D into the modeling software and automatically create the drilling point cloud model P through the command stream, as shown in Figure 3.

Step 3: After reading the 62 geological exploration drilling data of the divided units, the total types of strata that appear are analyzed and the soil layer labels are calibrated as shown in the following table:

Table 10: Geological exploration drilling data after unit division (unit: meter)

Step 4: According to the information in the geological survey report, assign geotechnical parameter design reference values to each soil layer in the above table. This embodiment takes natural gravity γ (kN/m ³ ) as an example, and the specific values are shown in the following table:

Table 11: Design reference values of natural gravity γ for each soil layer

Steps 5-6: Based on the drilling point cloud model, draw the corresponding geological and shield section models according to the specific location requirements and perform grid division. The present invention takes 100 meters of the planned shield mileage as an example, and the specific section model M _dm ' is shown in Figures 4 and 5.

Steps 7-8: Extract all non-repeated node coordinates in each grid model M _dm ' and analyze the maximum likelihood soil layer type to which each grid node belongs through the KNN algorithm, and then assign corresponding labels to generate V' with soil layer labels. In this embodiment, after repeated tests and comparisons, the parameter k of the KNN algorithm is 7. The V' of the shield section is as follows:

The data frame V' of the geological section is as follows:

Step 9: Calculate the elevation range of V’[:,2] under each soil layer label according to the V’ data frame of the geological section, and calculate the average thickness and distribution of each soil layer under the geological section from high to low. The results are shown in the following table:

Table 12 Distribution of soil layers in geological sections at 100m of shield mileage

Step 10: Use the trained Universal Kriging interpolation algorithm model to predict the geotechnical parameters of each node in the shield section (taking natural gravity γ as an example). The prediction results are shown in the following table:

Table 13 Soil parameters at each node in the shield section (natural gravity γ)

Step 11: Use VertexColor to visualize the mesh section models of geology and shield. Since the geological section needs to distinguish different types of soil layers, grayscale rendering is used; for the shield section used for construction, the threshold of geotechnical parameters needs to be displayed, so the grayscale parameters during rendering are linearly related to the geotechnical parameters. The visualization effects of the two are shown in Figures 6 to 8.

Table 12 and Figure 6 clearly show the geological distribution at 100m of the shield mileage. 1 to 5 in Figure 6 represent the five soil layers 1-1, 1-2, 10-1, 10-4, and 20a-2 in Table 5. This result can provide feedback on the soil depth parameters of the section, which can be used for the design and calculation of the segment structure; Table 13 and Figures 7 and 8 show the spatial distribution and probability distribution of geotechnical parameters at the working face at 100m of the shield mileage, which helps to take corresponding construction management decisions in advance during excavation construction.

The present invention subdivides the drilling data in the geological survey report and performs drilling point cloud modeling based on the spatial data. Then, the grid nodes are divided into units by establishing a grid model of the section. The present invention uses the KNN algorithm based on the point cloud drilling data to predict the soil layer distribution of the section grid; and uses the Universal Kriging interpolation algorithm based on the point cloud drilling data to predict the distribution of geotechnical parameters. The present invention finally visualizes the results by color rendering. The present invention is executed by a pre-coded program, which overcomes the low efficiency of repeated manual operations and manual calculations, and the whole process can be automatically repeated. The present invention is based on the computer to automatically execute the script, and the feedback numerical results are more accurate and faster, and the results can also be more intuitively displayed from a three-dimensional perspective. The present invention helps designers and construction personnel to grasp the geological conditions more accurately and timely.

According to another aspect of the present invention, a three-dimensional geological analysis device is provided, which is applied to shield tunnel design and construction, and the three-dimensional geological analysis device is used to execute a three-dimensional geological analysis method.

It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. The three-dimensional geological analysis method is characterized by being used for shield tunnel design and construction and comprising the following steps:

S1: acquiring a data frame D corresponding to the subdivided geological survey drilling data; the first column D of D is that 0 is the serial number of the geological survey drilling hole; the second column D of D is 1, the three columns D is 2 and the four columns D is 3 are three-dimensional coordinate data of the geological survey borehole; the fifth column D [: 4] of the D is soil layer data of the geological survey drill hole, is used for analyzing the types of soil layers, and assigns values for rock-soil parameters of each soil layer to obtain a sixth column D [: 5];

S2: carrying out drilling point cloud modeling by utilizing the three-dimensional coordinate data { D [: 1], D [: 2], D [: 3] } of the geological survey drilling hole to obtain a drilling point cloud model P; drawing a geological and shield section model M _dm based on the drilling point cloud model P, and carrying out grid division on the model M _dm to obtain a new grid model M _dm'; extracting the vertex coordinates (x _i y_i z_i) of the ith grid node of the ith behavior of the ith grid node in the coordinate data frames V and V of all non-repeated grid nodes in the M _dm', and calculating the minimum distance between the coordinates of each grid node and the drilling point cloud model P; analyzing the maximum likelihood soil layer types SoilType of each grid node by taking the minimum distance as an index to generate a soil layer type information data frame V' = (V SoilType); calculating the average thickness of each soil layer contained in the geological and shield section model M _dm by utilizing the soil layer type information data frame V' so as to obtain section soil layer distribution;

s3: training the interpolation algorithm model by taking the three-dimensional coordinate data { D [: 1, D [: 2, D [: 3] } of the geological survey drill hole and the rock-soil parameters D [: 5] of each soil layer as training samples to obtain a target model; and predicting soil layer parameter distribution of each grid node by using the target model.

2. The three-dimensional geological analysis method of claim 1, wherein the analyzing of the soil layer data D [: 4] of the geological borehole in S1 to obtain the kind of soil layer comprises:

Carrying out soil layer division on soil layer data D [: 4] of the geological survey drilling hole according to a division scale to obtain soil layer types; the dividing scale is smaller than the elevation difference of the interface between the upper soil layer and the lower soil layer; the average thickness of each soil layer is the average value of more than three thickness values corresponding to the same soil layer.

3. The three-dimensional geological analysis method according to claim 1, wherein the profile soil layer distribution in the step S2 is used for segment design of a shield tunnel.

4. The three-dimensional geologic analysis method of claim 1, wherein the geologic and shield cross-sectional model M _dm in S2 comprises Solid objects Solid and boundary objects Brep.

5. The three-dimensional geological analysis method according to claim 1, wherein the step S2 of analyzing the maximum likelihood soil layer type SoilType to which each grid node belongs by using the minimum distance as an index to generate the soil layer type information data frame V' = (V SoilType) includes:

And analyzing the maximum likelihood soil layer types SoilType of each grid node by using a KNN algorithm by taking the minimum distance of each grid node as an index, and generating a soil layer type information data frame V'.

6. The three-dimensional geological analysis method according to claim 1, wherein the soil layer parameter distribution in the step S3 is used for predicting the rock-soil parameter distribution of the face in the shield construction.

7. The three-dimensional geologic analysis method of claim 1, wherein the interpolation algorithm model in S3 is a three-dimensional Universal Kriging interpolation algorithm model.

8. The three-dimensional geologic resolution method of claim 1, further comprising:

S4: and carrying out visual treatment on the profile soil layer distribution and the soil layer parameter distribution.

9. The three-dimensional geologic resolution method of claim 8, wherein S4 comprises: and sequentially carrying out vertex coloring and fuzzy rendering to carry out three-dimensional visualization processing on the profile soil layer distribution and the rock-soil parameter distribution.

10. A three-dimensional geological analysis device, applied to shield tunnel design and construction, for performing the method of any one of claims 1-9.