Implicit three-dimensional geological modeling method and device combined with machine learning and medium

Abstract

The invention provides an implicit three-dimensional geologic modeling method and device combined with machine learning and a medium, and relates to the field of three-dimensional geologic modeling, and the method comprises the steps: obtaining original drilling data of a research area; a preset number of virtual drill holes are generated through a distance weighting method, and virtual drill hole data are obtained; preprocessing the original drilling data and dividing the original drilling data into a training set and a test set; training and testing the random forest model through the training set and the test set; the lithology of the virtual drilling data is predicted through the trained random forest model, and a virtual lithology prediction result is obtained; adding the virtual lithology prediction result to the virtual drilling data to obtain first virtual drilling data; the original drilling data and the first virtual drilling data are imported into implicit modeling software for three-dimensional modeling, and a three-dimensional ore body model is obtained. A more accurate and reliable geologic model is constructed by adding the virtual drill hole and predicting the virtual drill hole by the random forest model.

Description

Technical Field

[0001] The present application relates to the field of three-dimensional geological modeling, and in particular to an implicit three-dimensional geological modeling method, device and medium combined with machine learning. Background Art

[0002] 3D geological modeling is a technology that uses computer technology to combine spatial information management, geological interpretation, spatial analysis and prediction, geostatistics, entity content analysis, and graphic visualization tools in a virtual 3D environment and applies it to geological analysis. After years of theoretical updates and technological development, 3D geological modeling has been widely used in mining production, road construction, urban design and other industries.

[0003] According to the different principles of model construction, 3D geological modeling methods are divided into explicit modeling and implicit modeling. Explicit modeling technology is relatively mature and is the most widely used method in early 3D geological software at home and abroad. The traditional solid modeling method uses triangle splicing to define the surface of complex shapes, which is called the "explicit" model of the entity. The explicit modeling method requires the exploration engineering data to be displayed through continuous sections, the contours of the complex 3D body to be digitized, and then these 2D polylines are connected with connecting lines to define the 3D connectivity between the polylines. Finally, the polylines are triangulated to produce a 3D solid triangulation model. When performing explicit modeling, the ore body contour line must first be delineated through manual interaction, and then the 2D contour line is reconstructed in 3D through the contour splicing method.

[0004] Implicit modeling is a mathematical model that can easily add exploration data of different scales to the model, which can not only greatly reduce manual interaction, but also improve the automation and efficiency of the modeling process, and can achieve fast dynamic update of the model. Implicit modeling is based on spatial sampling data, using the relationship between data to establish implicit functions, and using spatial interpolation algorithms to achieve three-dimensional visualization of the ore body.

[0005] However, current explicit modeling requires a large amount of human-computer interaction, resulting in low modeling efficiency, lack of support for dynamic updates, and is greatly influenced by the subjective consciousness of the modeler. The constructed model may be quite different from the actual situation. At the same time, the implicit functions used in current implicit modeling software are relatively simple, and the effect of implicit modeling depends to a large extent on the quality and distribution of the input data. If the data is sparse or there are errors, the model may be inaccurate. Summary of the invention

[0006] The purpose of the present invention is to provide an implicit three-dimensional geological modeling method, device and medium combined with machine learning in order to solve the problem of low modeling accuracy of existing three-dimensional geological modeling methods.

[0007] The above-mentioned purpose of the present application is achieved through the following technical solutions: S1: Obtain the original drilling data of the study area; S2: Generate a preset number of virtual boreholes according to a preset density by using a distance weighted method to obtain virtual borehole data; S3: preprocess the original drilling data and divide it into training set and test set; S4: Build a random forest model; train and test the random forest model through training sets and test sets; S5: predicting the lithology of the virtual borehole data by using the trained random forest model to obtain a virtual lithology prediction result; adding the virtual lithology prediction result to the virtual borehole data to obtain first virtual borehole data; S6: Importing the original drilling data and the first virtual drilling data into implicit modeling software for three-dimensional modeling to obtain a three-dimensional ore body model.

[0008] Optionally, step S1 includes: the original drilling data includes: drilling name, hole mouth coordinates, drilling depth, azimuth, inclination, lithology and element analysis data.

[0009] Optionally, step S3 includes: preprocessing includes: resampling and standardization.

[0010] Optionally, the steps of resampling the original drilling data include: Assume that the No. 1 drilling hole is To Depth The lithology is A, and this record is converted into , ,……, The four-dimensional array form of; The calculation formula for the core coordinates of the kth section of borehole No. 1 during the resampling process is as follows:

[0011] in, is the coordinate of the hole opening of borehole No. 1, Indicates the resampling Section core; is the resampling interval; is the drilling inclination; is the drilling azimuth.

[0012] Optionally, the steps of standardizing the original drilling data and dividing it into a training set and a test set include: The original drilling data is standardized by Z-score standardization, as follows: The original drilling data was scaled to a range with a mean of 0 and a standard deviation of 1; The formula for Z-score standardization is as follows:

[0013] in, is the eigenvalue of the original drilling data, is the mean, is the standard deviation; The preprocessed raw drilling data is divided into a training set and a test set according to a preset ratio.

[0014] Optionally, step S4 includes: using the scikit-learn library of Python to train a random forest model.

[0015] Optionally, step S6 includes: S61: The implicit modeling software adopts Leapfrog; According to the original drilling data and the first virtual drilling data, a Leapfrog drilling database is constructed; the Leapfrog drilling database includes: a collar table, a survey table, a lithology table, and an assays table; S62: Establish the contact surface of the ore body through implicit modeling software combined with the drill hole database; S63: adjusting the parameters of the contact surface, the parameters including: trend surface and ellipsoid ratio; S64: Generate a three-dimensional ore body model through the contact surface after parameter adjustment.

[0016] An electronic device comprises a processor, a memory, a user interface and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes an implicit three-dimensional geological modeling method combined with machine learning.

[0017] A computer-readable storage medium stores instructions, which, when executed, perform an implicit three-dimensional geological modeling method combined with machine learning.

[0018] The beneficial effects of the technical solution provided by this application are: 1. Collect the original borehole data in the study area; generate virtual boreholes at a certain density by using the distance weighting method to achieve virtual encryption of the boreholes, and then preprocess the original borehole data; train the random forest model based on the training set of the preprocessed original borehole data, use the trained random forest model to predict the lithology data of the virtual borehole, and finally import the original borehole data and virtual borehole data into the implicit modeling software Leapfrog for 3D geological modeling. Based on the original borehole training machine learning model, the random forest model with the best prediction effect is selected to predict the lithology of the virtual borehole, achieve virtual encryption of the borehole, and meet the requirements of implicit modeling for data quality and quantity. The implicit modeling method overcomes the shortcomings of the traditional explicit modeling method, which is time-consuming and labor-intensive, has a huge workload and cannot be automatically updated, and also avoids the multiple solutions and uncertainty of the model caused by the subjectivity of the modeler; by combining virtual boreholes with machine learning, the requirements of implicit modeling for data quality and density are met, while ensuring a low cost, not only speeding up the modeling speed, but also improving the quality and refinement of the generated 3D geological model.

[0019] 2. Resample the boreholes. Since the geological significance of the same lithology between the upper and lower boundary points of different lithologies cannot be recognized in machine learning, it is difficult for the classifier to obtain ideal training results. Therefore, it is necessary to resample the boreholes so that the borehole sampling data becomes a series of four-dimensional array data sets with spatial positions and lithology attributes to improve the prediction accuracy of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present application will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 is a step diagram in an embodiment of the present application; Figure 2 is a comparison diagram of modeling with and without adding virtual grooves in an embodiment of the present application; Figure 3 It is a schematic diagram of a four-dimensional array in an embodiment of the present application; Figure 4 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to have a clearer understanding of the technical features, purposes and effects of the present application, the specific implementation methods of the present application are now described in detail with reference to the accompanying drawings.

[0022] An embodiment of the present application provides an implicit three-dimensional geological modeling method combined with machine learning.

[0023] Please refer to Figure 1 , Figure 1It is a step diagram of an implicit three-dimensional geological modeling method combined with machine learning in an embodiment of the present application, including: S1: Obtain the original drilling data of the study area; Step S1 includes: the original drilling data includes: drilling name, hole coordinates, drilling depth, azimuth, inclination, lithology and element analysis data.

[0024] S2: Generate a preset number of virtual boreholes according to a preset density by using a distance weighted method to obtain virtual borehole data; Specifically, in order to meet the requirements of refined modeling, virtual boreholes are generated according to a preset density through a distance weighted method to achieve virtual encryption of the boreholes.

[0025] In an exemplary embodiment, Figure 2 As shown, Figure 2 (Figure 3) Comparison of modeling with and without virtual grooves added) The requirements of implicit modeling on data quality and quantity are met, and the hole coordinates, azimuth, inclination, and drilling depth of the virtual borehole are obtained; the lithology of the generated virtual borehole is unknown, and subsequent prediction is required using a machine learning model trained based on the original borehole data.

[0026] S3: preprocess the original drilling data and divide it into training set and test set; Step S3 includes: preprocessing including: resampling and standardization.

[0027] The steps to resample the original drill hole data include: In an exemplary embodiment, since the geological significance of the same lithology between the upper and lower boundary points of different lithologies cannot be recognized in machine learning, it is difficult for the classifier to obtain an ideal training effect. Therefore, it is necessary to resample the boreholes so that the borehole sampling data becomes a series of four-dimensional array data sets with spatial positions and lithology attributes, such as Figure 3 shown.

[0028] Assume that the No. 1 drilling hole is To Depth The lithology is A, and this record is converted into , ,……, The four-dimensional array form of; The calculation formula for the core coordinates of the kth section of borehole No. 1 during the resampling process is as follows:

[0029] in, is the coordinate of the hole opening of borehole No. 1, Indicates the resampling Section core; is the resampling interval; is the drilling inclination; is the drilling azimuth.

[0030] The steps of standardizing the raw drilling data and dividing it into training and test sets include: The original drilling data is standardized by Z-score standardization, as follows: The original drilling data was scaled to a range with a mean of 0 and a standard deviation of 1; The formula for Z-score standardization is as follows:

[0031] in, is the eigenvalue of the original drilling data, is the mean, is the standard deviation; Specifically, directly using the original data will cause the algorithm to require many iterations during the convergence process. However, through standardization, the numerical range of each feature can be made similar, thereby accelerating the convergence of the algorithm, while also improving the accuracy of the model and preventing the model from overfitting.

[0032] The preprocessed raw drilling data is divided into a training set and a test set according to a preset ratio.

[0033] In an exemplary embodiment, the data set is divided: the data is randomly divided into a training set and a test set, where the ratio of training set to test set is 8:2, the training set is used to train the machine learning model, and the test set is used to comprehensively evaluate the predictive performance of the predictive model.

[0034] S4: Build a random forest model; train and test the random forest model through training sets and test sets; Step S4 includes: using the scikit-learn library of Python to train the random forest model.

[0035] Specifically, three machine learning models, namely, KNN, random forest, and gradient boosting machine, are identified in this application. The main hyperparameters selected by the KNN model are k, weight, and p, the main parameters selected by the random forest model are max_depth, min_samples_splits, and n_estimators, and the main parameters selected by the gradient boosting model are n_estimators, learning_rate, max_depth, min_samples_split, and min_samples_leaf. Bayesian optimization and 5-fold cross-validation methods are used to tune parameters and find the best hyperparameter combination. The machine learning model is trained using the data of the training set under the condition of the best hyperparameter combination; the overall prediction effect of the model is evaluated by the accuracy of the machine learning model under the best hyperparameter combination; the prediction performance of the model for different types of lithology is demonstrated by the confusion matrix and the receiver operating characteristic curve (i.e., ROC curve); the machine learning model with the best prediction performance is the random forest model based on the comprehensive accuracy, recall rate, confusion matrix, and ROC curve.

[0036] S5: predicting the lithology of the virtual borehole data by using the trained random forest model to obtain a virtual lithology prediction result; adding the virtual lithology prediction result to the virtual borehole data to obtain first virtual borehole data; S6: Import the original drilling data and the virtual drilling data into the implicit modeling software for three-dimensional modeling to obtain a three-dimensional ore body model.

[0037] Step S6 includes: S61: The implicit modeling software adopts Leapfrog; According to the original drilling data and the first virtual drilling data, a Leapfrog drilling database is constructed; the Leapfrog drilling database includes: a collar table, a survey table, a lithology table, and an assays table; Specifically, the commonly used borehole database mainly has four data files, namely: wellhead file (collar table), inclinometer file (survey table), lithology file (lithology table), sample analysis file (assays table), among which the collar table contains the borehole name (Holeid), borehole coordinates X, Y, Z and borehole depth (Max depth); the survey table contains the borehole name (Holeid), borehole depth (Max depth), azimuth (Amizuth) and dip (Dip); the lithology table contains the borehole name (Holeid), from (From), to (To) and lithology (lithology); the assays table contains the borehole name (Holeid), from (From), to (To) and elemental analysis data.

[0038] S62: Establish the contact surface of the ore body through implicit modeling software combined with the drill hole database; S63: adjusting the parameters of the contact surface, the parameters including: trend surface and ellipsoid ratio; Specifically, the parameters of the contact surface are adjusted to make the ore body model more realistic.

[0039] S64: Generate a three-dimensional ore body model through the contact surface after parameter adjustment.

[0040] The present application also discloses an electronic device. Figure 4 , Figure 4 The electronic device 500 may include: at least one processor 501 , at least one network interface 504 , a user interface 503 , a memory 505 , and at least one communication bus 502 .

[0041] The communication bus 502 is used to realize the connection and communication between these components.

[0042] The user interface 503 may include a display screen, and the optional user interface 503 may also include a standard wired interface or a wireless interface.

[0043] The network interface 504 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0044] The present application also discloses a computer-readable storage medium storing a plurality of instructions suitable for loading by a processor to execute the above-mentioned implicit three-dimensional geological modeling method combined with machine learning.

[0045] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0046] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. An implicit three-dimensional geological modeling method combined with machine learning, characterized in that: The method comprises the following steps: S1: Obtain the original drilling data of the study area; S2: Generate a preset number of virtual boreholes according to a preset density by using a distance weighted method to obtain virtual borehole data; S3: preprocess the original drilling data and divide it into training set and test set; S4: Build a random forest model; train and test the random forest model through training sets and test sets; S5: predicting the lithology of the virtual borehole data by using the trained random forest model to obtain a virtual lithology prediction result; adding the virtual lithology prediction result to the virtual borehole data to obtain first virtual borehole data; S6: Importing the original drilling data and the first virtual drilling data into implicit modeling software for three-dimensional modeling to obtain a three-dimensional ore body model.

2. The implicit three-dimensional geological modeling method combined with machine learning according to claim 1, characterized in that: Step S1 includes: the original drilling data includes: drilling name, hole coordinates, drilling depth, azimuth, inclination, lithology and element analysis data.

3. The implicit three-dimensional geological modeling method combined with machine learning according to claim 1, characterized in that: Step S3 includes: preprocessing including: resampling and standardization.

4. The implicit three-dimensional geological modeling method combined with machine learning according to claim 3, characterized in that: The steps to resample the original drill hole data include: Assume that the No. 1 drilling hole is To Depth The lithology is A, and this record is converted into , ,……, The four-dimensional array form of; The calculation formula for the core coordinates of the kth section of borehole No. 1 during the resampling process is as follows: in, is the hole coordinate of borehole No. 1, Indicates the resampling Section core; is the resampling interval; is the drilling inclination; is the drilling azimuth.

5. The implicit three-dimensional geological modeling method combined with machine learning according to claim 3, characterized in that: The steps of standardizing the raw drilling data and dividing it into training and test sets include: The original drilling data is standardized by Z-score standardization, as follows: The original drilling data was scaled to a range with a mean of 0 and a standard deviation of 1; The formula for Z-score standardization is as follows: in, is the eigenvalue of the original drilling data, is the mean, is the standard deviation; The preprocessed raw drilling data is divided into a training set and a test set according to a preset ratio.

6. The implicit three-dimensional geological modeling method combined with machine learning according to claim 1, characterized in that: Step S4 includes: using the scikit-learn library of Python to train the random forest model.

7. The implicit three-dimensional geological modeling method combined with machine learning according to claim 1, characterized in that: Step S6 includes: S61: The implicit modeling software adopts Leapfrog; According to the original drilling data and the first virtual drilling data, a Leapfrog drilling database is constructed; the Leapfrog drilling database includes: a collar table, a survey table, a lithology table, and an assays table; S62: Establish the contact surface of the ore body through implicit modeling software combined with the drill hole database; S63: adjusting the parameters of the contact surface, the parameters including: trend surface and ellipsoid ratio; S64: Generate a three-dimensional ore body model through the contact surface after parameter adjustment.

8. An electronic device, characterized in that: The electronic device comprises a processor (501), a memory (505), a user interface (503) and a network interface (504), wherein the memory (505) is used to store instructions, the user interface (503) and the network interface (504) are used to communicate with other devices, and the processor (501) is used to execute the instructions stored in the memory (505) so that the electronic device executes the method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed by a computer, the method according to any one of claims 1 to 7 is executed.

Images