Rock-soil boundary surface division method and device, equipment and storage medium
Through the combination of neural network training and prediction point data, the problem of insufficient accuracy of traditional geotechnical boundary division methods is solved, and a more accurate construction of geotechnical boundary model is achieved.
Patent Information
- Application Number
- CN202510041145.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-10
- Publication Date
- 2025-03-11
AI Technical Summary
The traditional method of geotechnical boundary division cannot fully utilize all measured data in the modeling area, resulting in insufficient construction accuracy of geotechnical boundary.
By obtaining the geotechnical data of drilling samples in the engineering area, constructing matrix A and normalizing attribute data, assigning soil types, randomly dividing the data set to train neural networks, generating prediction point data, fitting the geotechnical boundary and calculating volume difference until the preset threshold is reached.
A more accurate construction of the geotechnical boundary model is achieved, all measured data are fully utilized, and the accuracy of stratigraphic boundary division is improved.
Smart Images
Figure CN119600222A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of rock and soil boundary division, and in particular to a rock and soil boundary surface division method, device, equipment and storage medium. Background Art
[0002] In geotechnical engineering investigation, the division of strata is one of its main purposes. With the continuous advancement of science and technology, geological modeling technology has become a key tool for characterizing and analyzing geological conditions.
[0003] Geological models are usually divided into three categories: wireframe models, surface models and solid models. Among them, surface models play a core role in dividing stratum boundaries, and their accuracy directly affects the establishment of solid models and the statistics of stratum volumes.
[0004] In order to construct a surface model for dividing the geotechnical boundary, it is usually necessary to obtain the measured boundary depth of the hole-laying points and the predicted values of the non-hole-laying points. The traditional method of obtaining the predicted values of non-hole-laying points relies on the derivation of several measured values closest to the predicted points, but this method often cannot make full use of all the measured data in the modeling area, thus limiting the accuracy of geotechnical boundary surface construction. Summary of the invention
[0005] The purpose of the present application is to overcome the above existing technologies and provide a rock-soil boundary surface division method, device, equipment and storage medium.
[0006] The present application provides a method for dividing a rock-soil boundary surface, comprising:
[0007] Obtaining geotechnical data of borehole sampling in the project area and determining the spatial boundaries of the geotechnical data;
[0008] Extracting attribute data and soil type from the geotechnical data;
[0009] Construct a matrix A according to the geotechnical data, normalize the attribute data in the matrix A, and assign values to the soil types: assign a value of 1 to rock and 0 to other types, thereby obtaining a matrix B;
[0010] The matrix B is randomly divided into a training data set and a validation data set, and the soil type is used as output data and the attribute data is used as input data to train a neural network to obtain a rock and soil boundary estimation model;
[0011] Setting a rock and soil depth and a vertical expansion interval, and randomly generating prediction point data within the spatial boundary according to the rock and soil depth and the vertical expansion interval;
[0012] The prediction point data, the total data of the geotechnical data, and the geotechnical data are respectively input into the geotechnical boundary estimation model as input data to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary;
[0013] According to the first eigenvalue and the second eigenvalue, respectively fitting a first rock-soil boundary surface and a second rock-soil boundary surface, and calculating a volume difference of a corresponding solid model based on the first rock-soil boundary surface and the second rock-soil boundary surface;
[0014] Determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, merge the predicted point data into the geotechnical data and repeat the above steps; if the volume difference is less than the preset threshold, output the second geotechnical boundary surface.
[0015] Optionally, obtain geotechnical data from borehole sampling in the project area, including:
[0016] Traverse the data.* files in the specified address, where the data.* file types include: .xsl, .csv.
[0017] Optionally, the attribute data include: spatial x-coordinate, spatial y-coordinate, orifice elevation, and representative bottom depth.
[0018] Optionally, the predicted point data is 1 to 2 times the data in the database.
[0019] Optionally, the total data of the prediction point data and the geotechnical data and the geotechnical data are respectively input into the geotechnical boundary estimation model as input data to obtain the first eigenvalue and the second eigenvalue of the geotechnical boundary, including:
[0020] The first eigenvalue and the second eigenvalue of the rock and soil boundary are assigned a value of 0 if they are less than 0.5, indicating that the rock and soil type is other, and are assigned a value of 1 if they are greater than or equal to 0.5, indicating that the rock and soil type is rock;
[0021] The number k of data links with boundary characteristic values of 0 at the same horizontal position is counted, and the geotechnical boundary value at the horizontal prediction point is calculated according to hi=dk+0.1, where d is the vertical expansion interval.
[0022] Optionally, the threshold is 5000m 3 .
[0023] Optionally, a python algorithm is used to perform the various steps.
[0024] The present application also provides a rock-soil boundary surface division device, comprising:
[0025] An acquisition module, used to acquire geotechnical data sampled by drilling in the project area and determine the spatial boundary of the geotechnical data;
[0026] An extraction module, used for extracting attribute data and soil type from the geotechnical data;
[0027] A matrix module is used to construct a matrix A according to the geotechnical data, normalize the attribute data in the matrix A, and assign values to the soil types: rock is assigned a value of 1, and other values are assigned a value of 0, thereby obtaining a matrix B;
[0028] A model module, used for randomly dividing the matrix B into a training data set and a verification data set, taking the soil type as output data and the attribute data as input data, and training a neural network to obtain a rock and soil boundary estimation model;
[0029] A generation module, used to set the rock and soil depth and the vertical expansion interval, and randomly generate prediction point data within the spatial boundary according to the rock and soil depth and the vertical expansion interval;
[0030] A prediction module, used to input the prediction point data, the total data of the geotechnical data, and the geotechnical data as input data into the geotechnical boundary estimation model to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary;
[0031] A comparison module, which fits a first rock-soil boundary surface and a second rock-soil boundary surface respectively according to the first eigenvalue and the second eigenvalue, and calculates a volume difference of a corresponding solid model based on the first rock-soil boundary surface and the second rock-soil boundary surface;
[0032] An output module is used to determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, the predicted point data is merged into the geotechnical data to repeat the above steps; if the volume difference is less than the preset threshold, the second geotechnical boundary surface is output.
[0033] The present application provides a rock-soil boundary surface division device, comprising:
[0034] A memory for storing a computer executable program of the above-mentioned rock-soil boundary surface division method;
[0035] The processor is used to call the computer executable steps and execute: obtaining geotechnical data sampled by drilling in the project area and determining the spatial boundary of the geotechnical data; extracting attribute data and soil type in the geotechnical data; constructing a matrix A according to the geotechnical data, normalizing the attribute data in the matrix A, and assigning values to the soil type: rock is assigned a value of 1, and other values are assigned a value of 0, so as to obtain a matrix B; randomly dividing the matrix B into a training data set and a verification data set, taking the soil type as output data and the attribute data as input data, and training a neural network to obtain a geotechnical boundary prediction model; setting the geotechnical depth and the vertical expansion interval, and within the spatial boundary, according to the geotechnical depth and the vertical expansion interval, Generate prediction point data at random intervals; input the prediction point data and the total data of the geotechnical data and the geotechnical data as input data into the geotechnical boundary estimation model to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary; fit a first geotechnical boundary surface and a second geotechnical boundary surface according to the first eigenvalue and the second eigenvalue, and calculate the volume difference of the corresponding entity model based on the first geotechnical boundary surface and the second geotechnical boundary surface; determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, merge the prediction point data into the geotechnical data and repeat the above steps; if the volume difference is less than the preset threshold, output the second geotechnical boundary surface.
[0036] The present application also provides a storage medium, comprising a computer executable program stored therein, wherein the computer executable program is used to be called by a processor to execute the steps of the above-mentioned rock-soil boundary surface division method.
[0037] The beneficial effects of this application are:
[0038] The present application provides a method for dividing a geotechnical boundary surface, including: obtaining geotechnical data sampled by drilling in an engineering area, and determining the spatial boundary of the geotechnical data; extracting attribute data and soil type from the geotechnical data; constructing a matrix A according to the geotechnical data, normalizing the attribute data in the matrix A, and assigning values to the soil type: assigning a value of 1 to rock and 0 to other values, thereby obtaining a matrix B; randomly dividing the matrix B into a training data set and a verification data set, using the soil type as output data and the attribute data as input data, and training a neural network to obtain a geotechnical boundary estimation model; setting a geotechnical depth and a vertical expansion interval, and within the spatial boundary, according to the geotechnical depth and the vertical expansion interval, Randomly generate prediction point data; input the prediction point data and the total data of the geotechnical data and the geotechnical data as input data into the geotechnical boundary estimation model to obtain the first eigenvalue of the geotechnical boundary and the second eigenvalue of the geotechnical boundary; fit the first geotechnical boundary surface and the second geotechnical boundary surface according to the first eigenvalue and the second eigenvalue, respectively, and calculate the volume difference of the corresponding entity model based on the first geotechnical boundary surface and the second geotechnical boundary surface; determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, merge the prediction point data into the geotechnical data and repeat the above steps; if the volume difference is less than the preset threshold, output the second geotechnical boundary surface. The present application combines the prediction point position with the actual drilling data to more accurately construct a geotechnical boundary surface model. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 This is a schematic diagram of the geotechnical boundary surface division process in this application;
[0040] Figure 2 It is a schematic diagram of the surface fitting comparison diagram of the rock-soil boundary surface in this application;
[0041] Figure 3 Schematic diagram of boundary surface evaluation results of an embodiment of the present application. DETAILED DESCRIPTION
[0042] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0043] Please refer to Figure 1 As shown, the present application provides a method for dividing a rock-soil boundary surface, comprising:
[0044] S101, obtaining geotechnical data of borehole sampling in the project area, and determining the spatial boundary of the geotechnical data.
[0045] This application uses Python to perform high-precision division of geotechnical boundary surfaces.
[0046] Automatically identify and read geotechnical data in the engineering area drilling sampling database by traversing the data.* files in the specified address. File types include .xsl, .csv, etc. The function automatically identifies the file type and reads the database based on the corresponding method.
[0047] The spatial boundary of the geotechnical data is determined according to the determination of the geotechnical data, and the spatial boundary refers to the spatial range of the geotechnical data.
[0048] S102: extracting attribute data and soil type from the geotechnical data.
[0049] The geotechnical data includes attribute data and soil type.
[0050] Attribute data: spatial x-coordinate, spatial y-coordinate, orifice elevation, representative bottom depth;
[0051] Soil types: silt, silty clay, silty silty clay, clay, silty clay, sand, rock.
[0052] S103, constructing a matrix A according to the geotechnical data, normalizing the attribute data in the matrix A, and assigning values to the soil types: assigning a value of 1 to rock and 0 to others, thereby obtaining a matrix B.
[0053] The geotechnical data include: data number, x-coordinate, y-coordinate, elevation z, representative bottom depth and soil type.
[0054] According to the geotechnical data, an m×n matrix A is constructed based on the numpy module of python. The elements of the matrix A are a ij , where the row coordinate i represents the data number, the maximum value m is the total number of data links, and the column coordinate j represents the data type required to build the algorithm model, and the maximum value n is the total number of types.
[0055] The 2nd to (n-1)th columns are attribute data, which are the input columns of the algorithm model. The nth column is the soil type, which is the output column of the algorithm model.
[0056] Assign values to the nth column, set "rock" = 1, and other soil types = 0; for the 2nd to (n-1) columns, use formula e ij =(a ij -a jmin ) / (a jmax -a jmin ) is normalized and the processed data is also stored in the form of a matrix as matrix B.
[0057] S104, randomly dividing the matrix B into a training data set and a verification data set, taking the soil type as output data and the attribute data as input data, training a neural network to obtain a rock and soil boundary estimation model.
[0058] The geotechnical boundary estimation model is established based on the algorithm module of Python through the matrix B. Preferably, the matrix is randomly divided into 70% training data and 30% verification data to obtain a training data set and a verification data set. A neural network is trained based on the training data set and the verification data set, such as XGBoost, random forest, gradient boosting method, to establish a geotechnical boundary estimation model.
[0059] S105, setting the rock and soil depth and the vertical expansion interval, and randomly generating prediction point data within the spatial boundary according to the rock and soil depth and the vertical expansion interval.
[0060] The vertical expansion interval is the vertical interval of the sampling points, which is related to the boundary surface accuracy and can be manually entered by the technician. The smaller the interval value, the higher the boundary surface accuracy, which is used to expand the calculation data required by the algorithm module. In this application, the horizontal position of the prediction point is determined by random value of normal distribution, and the number of initial prediction points generated is 1 to 2 times the actual number of holes drilled.
[0061] Specifically, within the spatial boundary of the geotechnical data, the horizontal position coordinates are randomly generated according to the input geotechnical depth and vertical expansion interval. Then, the elevation value under the horizontal position coordinates is obtained by spatial interpolation, and the representative bottom depth value is selected according to the vertical expansion interval, and finally the prediction point input data is formed.
[0062] For example, if the depth threshold is set to h and the vertical expansion interval is set to d, the number of prediction point data at the same horizontal position is x=h / d+1.
[0063] S106, respectively inputting the prediction point data, the total data of the geotechnical data, and the geotechnical data as input data into the geotechnical boundary estimation model to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary.
[0064] After normalizing the prediction point data, two calculations are performed using the total data of the prediction point data and geotechnical data as input data, and the geotechnical data as input data.
[0065] Input the input data into the geotechnical boundary prediction model, and use the geotechnical boundary prediction model to calculate the geotechnical boundary characteristic values of the geotechnical boundary, which are used to distinguish between rocks and other types in the geotechnical types. If the geotechnical boundary characteristic value obtained by calculating any input data is less than 0.5, assign 0; if it is greater than or equal to 0.5, assign 1.
[0066] Statistically count the number of data chains with geotechnical boundary characteristic values of 0 at the same horizontal position, and finally obtain the geotechnical boundary value at this horizontal prediction point after calculation. Specifically, when the number of data chains with boundary characteristic values of 0 at the horizontal position is a (a < x), the geotechnical boundary value h1 = a × d + d / 2, and the meaning of its value is the representative bottom depth.
[0067] The output data corresponding to the geotechnical data is the first characteristic value, and the total data of the prediction point data and the geotechnical data as the input data corresponds to the second characteristic value.
[0068] S107. According to the first characteristic value and the second characteristic value, respectively fit the first geotechnical interface and the second geotechnical interface, and calculate the volume difference of the corresponding solid model based on the first geotechnical interface and the second geotechnical interface.
[0069] Use the geotechnical data as the input data and input it into the geotechnical boundary prediction model to obtain the geotechnical boundary value. According to the geotechnical boundary value, perform geotechnical interface fitting based on the matplotlib module of python to obtain the first geotechnical interface.
[0070] Use the prediction point and the geotechnical data together as the input data and input it into the geotechnical boundary prediction model to obtain the geotechnical boundary value. According to the geotechnical boundary value, perform geotechnical interface fitting based on the matplotlib module of python to obtain the second geotechnical interface.
[0071] Based on the first geotechnical interface and the second geotechnical interface, combined with the vertical depth, calculate the volumes of the solid models corresponding to the first geotechnical interface and the second geotechnical interface respectively, and calculate the volume difference.
[0072] S108. Judge whether the volume difference is less than the preset threshold: if the volume difference is not less than the preset threshold, merge the prediction point data into the geotechnical data and repeat the above steps; if the volume difference is less than the preset threshold, output the second geotechnical interface.
[0073] Compare the volume difference with the preset threshold. If it is less than the preset threshold, the division of the geotechnical interface is completed; if it is greater than or equal to the preset threshold, use the prediction point data and the geotechnical data used in this fitting as the new geotechnical data, and repeat the above steps until the geotechnical interface is less than the threshold requirement.
[0074] The specific application effects are as follows:
[0075] Based on the database, the modeling area is determined to be 400m long, 200m wide, with an elevation range of 72 to 8m. There are 89 measured boreholes with a sampling interval of 2m.
[0076] Input vertical expansion interval d = 0.2m, use volume error for boundary surface evaluation, and its threshold is 5000m 3 The number of initial prediction points generated is 50, which are randomly generated using normal distribution.
[0077] The data number, X coordinate, Y coordinate, elevation Z, representative bottom depth h and soil type of 1800 data chains based on 89 measured boreholes in the initial database were extracted and expanded into 18000 data chains at a vertical expansion interval of d = 0.2m. A 18000 × 6 matrix A was constructed based on the numpy module of python.
[0078] The first column of the matrix A is the data number, and the second to sixth columns are the X coordinate, Y coordinate, elevation Z, representative bottom depth h and soil type, respectively.
[0079] The 2nd to 5th columns are used as the input columns of the algorithm model, and the 6th column is used as the output column of the algorithm model. Let "rock" = 1 in the 6th column, and the others = 0. Then, the 2nd to 5th columns are calculated by formula e ij =(a ij -a jmin ) / (a jmax -a jmin ) is normalized.
[0080] Finally, the processed data is also stored in the form of a matrix as matrix B, and the matrix is randomly divided into 70% for training data and 30% for verification data.
[0081] Through matrix B, a geotechnical boundary estimation model is established based on the BP neural network algorithm in python.
[0082] Within the spatial boundary conditions of the initial database, 50 point horizontal coordinates (x coordinate, y coordinate) are randomly generated. The elevation value under the horizontal position coordinate is obtained based on the measured borehole elevation value z through spatial interpolation method, and the representative bottom depth value is determined from top to bottom by the elevation according to the vertical expansion interval d = 0.2m, finally forming 4 columns of prediction point input data.
[0083] After normalizing the input data of the prediction points, the second eigenvalue of the geotechnical boundary is calculated using the geotechnical boundary estimation model. The second eigenvalue is assigned a value of 0 if it is less than 0.5, and a value of 1 if it is greater than or equal to 0.5.
[0084] The number of data links k with boundary characteristic values of 0 at the same horizontal position is counted, and the geotechnical boundary value (representing the bottom depth) at the horizontal prediction point is calculated according to h1=a×d+d / 2.
[0085] like Figure 2 As shown in the figure, the predicted points and the actual drilling sampling points are used as data points, and the geotechnical boundary surface fitting is performed based on the python matplotlib module ( Figure 2 The actual drilling sampling points are used as data points, and the geotechnical boundary surface fitting is performed based on the Python matplotlib module ( Figure 2 right figure).
[0086] Set the vertical depth of the solid model to 20m and calculate the volume difference between the two solid models. If it is less than 5000m 3 , the rock-soil boundary surface division is completed; if it is not less than the requirement, all the data points used in this fitting are taken as measured points, and the above steps are repeated until it is less than the threshold.
[0087] like Figure 3 As shown in the evaluation results of this embodiment, after 4 fittings, the rock-soil boundary surface meets the requirements.
[0088] The present application also provides a rock-soil boundary surface division device, comprising:
[0089] An acquisition module is used to acquire geotechnical data sampled by drilling in the project area and determine the spatial boundary of the geotechnical data;
[0090] An extraction module, used for extracting attribute data and soil type from the geotechnical data;
[0091] A matrix module is used to construct a matrix A according to the geotechnical data, normalize the attribute data in the matrix A, and assign values to the soil types: rock is assigned a value of 1, and other values are assigned a value of 0, thereby obtaining a matrix B;
[0092] A model module, used for randomly dividing the matrix B into a training data set and a verification data set, taking the soil type as output data and the attribute data as input data, and training a neural network to obtain a rock and soil boundary estimation model;
[0093] A generation module, used to set the rock and soil depth and the vertical expansion interval, and randomly generate prediction point data within the spatial boundary according to the rock and soil depth and the vertical expansion interval;
[0094] A prediction module, used to input the prediction point data, the total data of the geotechnical data, and the geotechnical data as input data into the geotechnical boundary estimation model to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary;
[0095] A comparison module, which fits a first rock-soil boundary surface and a second rock-soil boundary surface respectively according to the first eigenvalue and the second eigenvalue, and calculates a volume difference of a corresponding solid model based on the first rock-soil boundary surface and the second rock-soil boundary surface;
[0096] An output module is used to determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, the predicted point data is merged into the geotechnical data to repeat the above steps; if the volume difference is less than the preset threshold, the second geotechnical boundary surface is output.
[0097] The above embodiments are only used to illustrate the technical ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. The patent scope of the present invention cannot be limited only by these embodiments. For ordinary technicians in the technical field to which the present invention belongs, they can make several simple deductions and substitutions without departing from the concept of the present invention, which should be regarded as belonging to the protection scope of the present invention.
Claims
1. A method for dividing a rock-soil boundary surface, characterized in that: include: Obtaining geotechnical data of borehole sampling in the project area and determining the spatial boundaries of the geotechnical data; Extracting attribute data and soil type from the geotechnical data; Construct a matrix A according to the geotechnical data, normalize the attribute data in the matrix A, and assign values to the soil types: assign a value of 1 to rock and 0 to other types, thereby obtaining a matrix B; The matrix B is randomly divided into a training data set and a validation data set, and the soil type is used as output data and the attribute data is used as input data to train a neural network to obtain a rock and soil boundary estimation model; Setting a rock and soil depth and a vertical expansion interval, and randomly generating prediction point data within the spatial boundary according to the rock and soil depth and the vertical expansion interval; The prediction point data, the total data of the geotechnical data, and the geotechnical data are respectively input into the geotechnical boundary estimation model as input data to obtain a first eigenvalue and a second eigenvalue of the geotechnical boundary; According to the first eigenvalue and the second eigenvalue, respectively fitting a first rock-soil boundary surface and a second rock-soil boundary surface, and calculating a volume difference of a corresponding solid model based on the first rock-soil boundary surface and the second rock-soil boundary surface; Determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, merge the predicted point data into the geotechnical data and repeat the above steps; if the volume difference is less than the preset threshold, output the second geotechnical boundary surface.
2. The method for dividing the rock-soil boundary surface according to claim 1, characterized in that: Obtain geotechnical data from borehole sampling in the project area, including: Traverse the data.* files in the specified address, where the data.* file types include: .xsl, .csv.
3. The rock-soil boundary surface division method according to claim 1, characterized in that: The attribute data include: spatial x-coordinate, spatial y-coordinate, hole elevation, and representative bottom depth.
4. The rock-soil boundary surface division method according to claim 1, characterized in that: The predicted point data is 1 to 2 times the data in the database.
5. The rock-soil boundary surface division method according to claim 1, characterized in that: The prediction point data, the total data of the geotechnical data, and the geotechnical data are respectively input into the geotechnical boundary estimation model as input data to obtain the first eigenvalue and the second eigenvalue of the geotechnical boundary, including: The first eigenvalue and the second eigenvalue of the rock and soil boundary are assigned a value of 0 if they are less than 0.5, indicating that the rock and soil type is other, and are assigned a value of 1 if they are greater than or equal to 0.5, indicating that the rock and soil type is rock; The number k of data links with boundary characteristic values of 0 at the same horizontal position is counted, and the geotechnical boundary value at the horizontal prediction point is calculated according to hi=dk+0.1, where d is the vertical expansion interval.
6. The rock-soil boundary surface division method according to claim 1, characterized in that: The threshold is 5000m 3 .
7. The method for dividing a rock-soil boundary surface according to any one of claims 1 to 6, characterized in that: The python algorithm is used to perform each step.
8. A rock-soil boundary surface division device, characterized in that: include: An acquisition module, used to acquire geotechnical data sampled by drilling in the project area and determine the spatial boundary of the geotechnical data; An extraction module, used for extracting attribute data and soil type from the geotechnical data; A matrix module is used to construct a matrix A according to the geotechnical data, normalize the attribute data in the matrix A, and assign values to the soil types: rock is assigned a value of 1, and other values are assigned a value of 0, thereby obtaining a matrix B; A model module, used for randomly dividing the matrix B into a training data set and a verification data set, taking the soil type as output data and the attribute data as input data, and training a neural network to obtain a rock and soil boundary estimation model; A generation module, used to set the rock and soil depth and the vertical expansion interval, and randomly generate prediction point data within the spatial boundary according to the rock and soil depth and the vertical expansion interval; A prediction module, used to input the prediction point data, the total data of the geotechnical data, and the geotechnical data as input data into the geotechnical boundary estimation model to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary; A comparison module, which fits a first rock-soil boundary surface and a second rock-soil boundary surface respectively according to the first eigenvalue and the second eigenvalue, and calculates a volume difference of a corresponding solid model based on the first rock-soil boundary surface and the second rock-soil boundary surface; An output module is used to determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, the predicted point data is merged into the geotechnical data to repeat the above steps; if the volume difference is less than the preset threshold, the second geotechnical boundary surface is output.
9. A rock-soil boundary surface division device, characterized in that: include: A memory for storing a computer executable program of a rock-soil boundary surface division method according to any one of claims 1 to 6; A processor is used to call the computer executable steps to execute: obtaining geotechnical data of drilling sampling in the engineering area, determining the spatial boundary of the geotechnical data; extracting attribute data and soil type in the geotechnical data; constructing a matrix A according to the geotechnical data, normalizing the attribute data in the matrix A, and assigning values to the soil type: assigning a value of 1 to rock and 0 to other values, thereby obtaining a matrix B; The matrix B is randomly divided into a training data set and a validation data set, and the soil type is used as output data and the attribute data is used as input data to train a neural network to obtain a geotechnical boundary prediction model; the geotechnical depth and the vertical expansion interval are set, and within the spatial boundary, prediction point data are randomly generated according to the geotechnical depth and the vertical expansion interval; the prediction point data and the total data of the geotechnical data and the geotechnical data are respectively input into the geotechnical boundary prediction model as input data to obtain a first eigenvalue of the geotechnical boundary and a second eigenvalue of the geotechnical boundary; According to the first eigenvalue and the second eigenvalue, respectively fitting a first rock-soil boundary surface and a second rock-soil boundary surface, and calculating a volume difference of a corresponding solid model based on the first rock-soil boundary surface and the second rock-soil boundary surface; Determine whether the volume difference is less than a preset threshold: if the volume difference is not less than the preset threshold, merge the predicted point data into the geotechnical data and repeat the above steps; if the volume difference is less than the preset threshold, output the second geotechnical boundary surface.
10. A storage medium, characterized in that: The invention comprises a computer executable program stored therein, wherein the computer executable program is used to be called by a processor to execute the steps of a rock-soil boundary surface division method as described in any one of claims 1 to 6.
Citation Information
Cited By
Terrestrial clastic rock reservoir stratum division method, device, equipment and medium
CN120296473A