Uniaxial compressive strength prediction method and system based on multi-source while-drilling information fusion

Abstract

The present invention relates to the technical field of rock strength prediction, and specifically discloses a uniaxial compressive strength prediction method and system based on multi-source downhole information fusion, the method comprising: obtaining rock physical parameters, mineralogical parameters and downhole parameters and corresponding uniaxial compressive strength in the historical drilling process; constructing a uniaxial compressive strength prediction index system; combining feature filtering algorithm and Pearson correlation coefficient to perform feature selection on the uniaxial compressive strength prediction index system to determine the optimal input feature; collecting the optimal input feature of the area to be tested, and using the trained uniaxial compressive strength prediction model to predict the uniaxial compressive strength of the rock in the area to be tested. The present invention integrates rock physical parameters, mineralogical parameters and downhole parameters to predict the uniaxial compressive strength of rocks, which can not only greatly improve the prediction accuracy, but also solve the problem that there are multiple solutions in the prior art that only downhole parameters are used to predict strength.

Description

Technical Field

[0001] The present invention relates to the technical field of rock strength prediction, and in particular to a uniaxial compressive strength prediction method and system based on multi-source while-drilling information fusion. Background Art

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

[0003] The uniaxial compressive strength is an important parameter reflecting the mechanical properties of rock mass. It is the basic basis for determining the surrounding rock quality classification, support structure design, and construction method selection. Therefore, it is of great significance to geotechnical engineering to quickly and accurately determine the uniaxial compressive strength of rock mass.

[0004] In the prior art, the detection of uniaxial compressive strength of rock mass requires first coring of the rock mass and then testing through indoor tests. The rock mass coring and laboratory testing process is cumbersome, time-consuming and costly.

[0005] At present, many studies have shown through a large number of field and indoor digital drilling test studies that there is a close quantitative relationship between rock mechanical parameters and drilling parameters. The prediction of rock uniaxial compressive strength using drilling parameters (such as torque, drilling speed, drilling pressure, rotation speed, etc.) has become a popular research direction; for example:

[0006] The prior art discloses a method for testing rock strength using a drilling monitoring technology. By studying the quantitative relationship between drill pipe axial pressure, drill pipe torque, rotation speed, drilling speed and rock strength, the purpose of quickly and accurately obtaining rock strength during drilling can be achieved; the prior art discloses a method for quickly and in real time grading surrounding rock of underground projects based on drilling parameters. The uniaxial compressive strength and rock integrity index in the depth direction of the drilling face are obtained through the drilling parameters measured during the drilling process, and then a relationship between the drilling parameters and the uniaxial compressive strength and the rock integrity index is established. In actual underground projects, the obtained drilling parameters are substituted into the above relationship to obtain the uniaxial compressive strength and rock integrity index.

[0007] The above schemes all establish a quantitative relationship between the drilling parameters and the uniaxial compressive strength of the rock mass, and obtain the uniaxial compressive strength data of the rock mass through the drilling parameters obtained in actual projects; however, the prediction of the uniaxial compressive strength of the rock using only the drilling parameters often has the following technical problems:

[0008] (1) The uniaxial compressive strength of rock is affected not only by the drilling parameters, but also by the properties of the rock itself (such as mineralogical parameters, rock physical parameters, etc.). These factors are not reflected in the drilling parameters, resulting in multiple solutions of the prediction model, that is, the same drilling parameters may correspond to different rock strengths, or different combinations of drilling parameters may predict the same rock strength.

[0009] (2) Using only the drilling parameters cannot fully reflect the factors affecting rock strength, resulting in low prediction accuracy. Summary of the invention

[0010] In order to solve the above problems, the present invention proposes a uniaxial compressive strength prediction method and system based on multi-source while-drilling information fusion, which fuses rock physical parameters, mineralogical parameters and while-drilling parameters to achieve accurate prediction of the uniaxial compressive strength of the rock mass.

[0011] In some embodiments, the following technical solutions are adopted:

[0012] A uniaxial compressive strength prediction method based on multi-source while-drilling information fusion, comprising:

[0013] Obtain rock physical parameters, mineralogical parameters and drilling parameters and corresponding uniaxial compressive strength during historical drilling;

[0014] A uniaxial compressive strength prediction index system is constructed using rock physical parameters, mineralogical parameters and drilling parameters;

[0015] The feature filtering algorithm and Pearson correlation coefficient were combined to select the features of the uniaxial compressive strength prediction index system, and finally the optimal input features for training the uniaxial compressive strength prediction model were determined;

[0016] Taking the optimal input feature as input and the corresponding uniaxial compressive strength as output, the uniaxial compressive strength prediction model is trained to obtain a trained uniaxial compressive strength prediction model;

[0017] The optimal input features of the area to be tested are collected, and the uniaxial compressive strength of the rock in the area to be tested is predicted using the trained uniaxial compressive strength prediction model.

[0018] Wherein, the rock physical parameters include point load intensity index, P wave velocity, rebound value and rock density;

[0019] The mineralogical parameters include: mineral type, mineral content, mineral grain size and mineral morphology;

[0020] The drilling parameters include: drilling response parameters such as drilling speed and torque, and drilling operation parameters such as drilling pressure and rotation speed set by the drilling rig operator.

[0021] As a further solution, the feature filtering algorithm and Pearson correlation coefficient are combined to perform feature selection on the uniaxial compressive strength prediction index system, and finally determine the optimal input features for training the uniaxial compressive strength prediction model. Specifically, the feature data in the uniaxial compressive strength prediction index system are normalized, the weight of each feature in the index system is calculated, the weights of all features are sorted, and the features with weights less than the set threshold are eliminated;

[0022] Calculate the Pearson correlation coefficient between the remaining features. When the correlation coefficient is greater than the coefficient threshold, it means that there is a strong correlation between the two features. For two features with a strong correlation, delete one of them.

[0023] The remaining features are used as the optimal input features for training the uniaxial compressive strength prediction model.

[0024] As a further solution, the uniaxial compressive strength prediction model includes a two-layer structure, the first layer is a plurality of different bottom layer models, and the second layer is a top layer model;

[0025] The input data passes through each bottom-level model to obtain the corresponding uniaxial compressive strength prediction results; the uniaxial compressive strength prediction results of all bottom-level models are combined into new features and input into the top-level model to obtain the final uniaxial compressive strength prediction results;

[0026] Among them, the underlying model and the top model are both machine learning models.

[0027] As a further solution, the training process of the uniaxial compressive strength prediction model is as follows:

[0028] Construct the initial training set and test set using the optimal input features and the corresponding uniaxial compressive strength;

[0029] Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and the prediction results of all underlying models are combined to determine the new training set and test set;

[0030] Use the new training set to train the top-level model, and use the new test set to test it to obtain the trained top-level model;

[0031] According to the trained multiple bottom-level models and the trained top-level model, a trained uniaxial compressive strength prediction model is obtained.

[0032] As a further solution, it also includes: during the training process of the uniaxial compressive strength prediction model, the determination coefficient, root mean square error and mean absolute percentage error are used as evaluation indicators to compare the uniaxial compressive strength of rock predicted by the uniaxial compressive strength prediction model with the measured value.

[0033] In other embodiments, the following technical solutions are adopted:

[0034] A uniaxial compressive strength prediction system based on multi-source while-drilling information fusion, comprising:

[0035] Data acquisition module, used to obtain rock physical parameters, mineralogical parameters and drilling parameters and corresponding uniaxial compressive strength during historical drilling;

[0036] An index system construction module is used to construct a uniaxial compressive strength prediction index system using rock physical parameters, mineralogical parameters and drilling parameters;

[0037] The optimal input feature determination module is used to select features of the uniaxial compressive strength prediction index system by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally determine the optimal input features for training the uniaxial compressive strength prediction model;

[0038] A model training module is used to train a uniaxial compressive strength prediction model using the optimal input feature as input and the corresponding uniaxial compressive strength as output, thereby obtaining a trained uniaxial compressive strength prediction model;

[0039] The uniaxial compressive strength prediction module is used to collect the optimal input features of the area to be tested, and use the trained uniaxial compressive strength prediction model to predict the uniaxial compressive strength of the rock in the area to be tested.

[0040] In other embodiments, the following technical solutions are adopted:

[0041] A terminal device comprises a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store a plurality of instructions, wherein the instructions are suitable for being loaded by the processor and executing the above-mentioned uniaxial compressive strength prediction method based on multi-source while-drilling information fusion.

[0042] In other embodiments, the following technical solutions are adopted:

[0043] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded and executed by a processor of a terminal device. Compared with the prior art, the invention has the following beneficial effects:

[0044] (1) The present invention integrates rock physical parameters, mineralogical parameters and drilling parameters to predict the uniaxial compressive strength of rocks. Compared with the prediction model established only with drilling parameters, the present invention can not only greatly improve the prediction accuracy, but also solve the problem of multiple solutions in the prior art that only drilling parameters are used to predict the strength.

[0045] (2) The method of the present invention can acquire and process data in real time during the drilling process to predict the uniaxial compressive strength of rock. Compared with the traditional indoor uniaxial compression test, this method significantly improves the timeliness and saves a lot of time and cost.

[0046] (3) In the process of model construction, the present invention first combines the feature filtering algorithm and the Pearson correlation coefficient to perform feature optimization, thereby reducing the complexity of the model; at the same time, a model architecture combining multiple underlying models and a top-level model is adopted to improve the prediction accuracy and reduce the risk of overfitting.

[0047] Other features and advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a method for predicting uniaxial compressive strength based on multi-source while-drilling information fusion in an embodiment of the present invention;

[0049] Figure 2 Schematic diagram of the data processing process of the uniaxial compressive strength prediction model in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0051] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0052] Embodiment 1

[0053] In one or more embodiments, a method for predicting uniaxial compressive strength based on multi-source while-drilling information fusion is disclosed, combined with Figure 1 , specifically including the following processes:

[0054] Step (1): Obtain rock physical parameters, mineralogical parameters, and drilling parameters during historical drilling and the corresponding uniaxial compressive strength; and construct a uniaxial compressive strength prediction index system using the rock physical parameters, mineralogical parameters, and drilling parameters.

[0055] In this embodiment, the influence of physical parameters, mineralogical parameters and drilling parameters on the uniaxial compressive strength of rock is comprehensively considered, and an intelligent prediction index system for the uniaxial compressive strength of rock based on multi-source drilling information fusion is constructed.

[0056] Among them, rock physical parameters must be easy to obtain and closely related to the uniaxial compressive strength of rocks. They are mainly physical parameters that can reflect the initial defects inside rocks, including the point load intensity index that reflects the overall strength characteristics of rocks, the P-wave velocity that reflects the density and elastic properties of rocks, and the rebound value and rock density that reflect the mineral characteristics and weathering degree of rocks.

[0057] Mineralogical parameters include mineral type, mineral content, mineral particle size, mineral morphology, etc.

[0058] The drilling parameters include: drilling response parameters such as drilling speed and torque, and drilling operation parameters such as drilling pressure and speed set by the drilling rig operator.

[0059] Step (2): Combine the feature filtering algorithm and the Pearson correlation coefficient to perform feature selection on the uniaxial compressive strength prediction index system, and finally determine the optimal input feature for training the uniaxial compressive strength prediction model; use the optimal input feature as input and the corresponding uniaxial compressive strength as output to train the uniaxial compressive strength prediction model to obtain a trained uniaxial compressive strength prediction model.

[0060] In this embodiment, the method for performing feature selection on the uniaxial compressive strength prediction index system by combining the feature filtering algorithm and the Pearson correlation coefficient is specifically as follows:

[0061] (2-1) Normalize the feature data in the uniaxial compressive strength prediction index system and calculate the weight of each feature in the index system; sort the weights of all features and eliminate the features whose weights are less than the set threshold;

[0062] (2-2) Calculate the Pearson correlation coefficient between the remaining features. When the correlation coefficient is greater than the set coefficient threshold, it means that there is a strong correlation between the two features. For two features with a strong correlation, delete one of them.

[0063] (2-3) The remaining features are used as the optimal input features for training the uniaxial compressive strength prediction model.

[0064] For (2-1) of the above process, existing feature filtering algorithms (such as chi-square test, variance filtering, F test, etc.) can be used to implement it; among them, the normalization of feature data refers to converting the numerical range of the data into a unified standard range to ensure that all features have the same scale to avoid affecting the training effect of the model due to different feature dimensions. For example, Min-Max can be used for standardization.

[0065] As an optional example, this embodiment calculates the weight of each feature in the index system, and the specific process is as follows:

[0066] A sample is randomly selected and recorded as the sample; the sample includes various characteristics in the uniaxial compressive strength prediction index system;

[0067] Calculate the Euclidean distance between the sample and other samples:

[0068] ;

[0069] in, and Represents two samples, and Represents the eigenvalue of each sample, the number of features in n samples;

[0070] Select the sample closest to the sample based on the Euclidean distance and record it as the closest sample;

[0071] Taking the calculation of the weight of feature A in the sample as an example, the following are calculated based on the Bayesian criterion:

[0072] ; ; ;

[0073] in, Represents the probability of difference between feature A and the nearest sample; represents the probability of the prediction difference from the most recent sample; It represents the probability of difference in uniaxial compressive strength prediction under the premise that the nearest sample with difference has difference in feature A; , represents the numerical difference with the nearest sample on feature A; , represents the numerical difference from the most recent sample in the predicted uniaxial compressive strength; and represents the value of feature A between the sample and the most recent sample, and represents the predicted uniaxial compressive strength values ​​predicted using the sample and the most recent sample respectively; Represents the most recent sample.

[0074] The predicted value of the uniaxial compressive strength in this embodiment refers to the predicted value of the uniaxial compressive strength obtained by inputting the sample into the uniaxial compressive strength prediction model; of course, it may also be a predicted value obtained by other methods.

[0075] Then the weight W(A) of feature A is expressed as:

[0076] .

[0077] Set a weight threshold and remove features whose weight values ​​are less than the threshold.

[0078] For (2-2) of the above process, in order to further remove redundant features / input parameters, avoid multicollinearity, and improve the generalization ability of the model, this embodiment uses the Pearson correlation coefficient method for the remaining features after removal, and calculates the correlation coefficients between the remaining features. When the correlation coefficient is greater than the set threshold (for example, 0.8), it indicates that there is a strong correlation between the variables. The input features with strong correlation are deleted to determine the input features finally used for predictive model training.

[0079] Among them, the specific formula for calculating the correlation coefficient by the Pearson correlation coefficient method is as follows:

[0080] ;

[0081] ;

[0082] ;

[0083] Where r is the correlation coefficient and n is the total number of features; for different values ​​of a characteristic (such as torque), Represents the mean of all values ​​of this feature; represents different values ​​of another feature (such as rebound value), Represents the mean of all values ​​of this feature.

[0084] For (2-3) of the above process, after feature selection, the retained rock physical parameters and mineralogical parameters are all parameters that are closely related to the uniaxial compressive strength of rock, have low correlation with each other, and are easy to obtain on site; finally, the remaining features after feature selection in this embodiment are torque, drilling speed, P-wave velocity, rebound value, mineral type and mineral content, and these features are used as the optimal input features for training the uniaxial compressive strength prediction model.

[0085] After feature selection, the selected features such as torque, drilling speed, P-wave velocity, rebound value, mineral type and mineral content are used to train the uniaxial compressive strength prediction model.

[0086] In this embodiment, combined with Figure 2 ,The uniaxial compressive strength prediction model includes a two-layer structure, the first layer is a number of different bottom-layer models, and the second layer is the top-layer model;

[0087] The original data set (features after feature selection) passes through each bottom-level model to obtain the corresponding uniaxial compressive strength prediction results; the uniaxial compressive strength prediction results of all bottom-level models are combined into a new data set (new features), which is input into the top-level model to obtain the final uniaxial compressive strength prediction results;

[0088] Among them, both the bottom model and the top model are machine learning models, which can be selected according to actual needs; for example: the bottom model can choose random forest, support vector machine and neural network, etc., which are mainly heterogeneous bottom models, that is, bottom models with different principles are selected to fully extract the relationship between input parameters (drilling parameters, rock physical parameters and rock mineralogy parameters) and output parameters (uniaxial compressive strength); the top model usually selects a relatively simple model with good generalization ability, such as linear regression.

[0089] The training process of the uniaxial compressive strength prediction model in this embodiment is specifically as follows:

[0090] Based on the determined optimal input features, historical torque, drilling speed, P-wave velocity, rebound value, mineral type and mineral content, and the corresponding uniaxial compressive strength of the rock mass are obtained; initial training sets and test sets are constructed;

[0091] Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and the prediction results of all underlying models are combined to determine the new training set and test set;

[0092] Use the new training set to train the top-level model, and use the new test set to test it to obtain the trained top-level model;

[0093] According to the trained multiple bottom-level models and the trained top-level model, a trained uniaxial compressive strength prediction model is obtained.

[0094] In this embodiment, the hyperparameters of each underlying model are optimized. The specific process includes: using an optimization algorithm to optimize the hyperparameters of the underlying model, finding the best hyperparameters of the underlying model, and using the Early-stopping strategy and regularization in the hyperparameter optimization process to avoid overfitting.

[0095] By optimizing the model's hyperparameters, the underlying model can perform better on the training data, thereby improving the predictive power of the entire model. At the same time, optimizing hyperparameters can improve the stability and robustness of the model. In practical applications, the data set may contain noise or imbalance. By adjusting the hyperparameters, the model's robustness to noise and outliers can be enhanced, making its performance on different data sets more stable. The better the performance of the underlying model, the more significant the integration effect of the final model. If each underlying model is optimized and has stronger predictive power, the prediction model composed of these underlying models will have better predictive performance.

[0096] The trained uniaxial compressive strength prediction model was evaluated to determine the coefficient (R 2 ), root mean square error (RMSE) and mean absolute percentage error (MAPE) are used as evaluation indicators to compare the rock uniaxial compressive strength predicted by the uniaxial compressive strength prediction model with the measured value. 2 It indicates the accuracy of model fitting data. The closer it is to 1, the better the fitting effect. RMSE indicates the degree of deviation between the model prediction value and the measured value. The smaller the value, the better the prediction performance of the model. MAPE indicates the average value of the absolute error between the model prediction value and the measured value. The smaller the value, the better the prediction performance of the model.

[0097] Step (3): Use the uniaxial compressive strength prediction model to predict the uniaxial compressive strength. Specifically, obtain the optimal input characteristic data during the drilling process of the test area. In this embodiment, the torque, drilling speed, P-wave velocity, rebound value, mineral type and mineral content data are obtained; wherein the torque and drilling speed are measured by a downhole sensor; the P-wave velocity is detected by ultrasonic wave, and the rebound value is measured by a Schmidt rebound hammer; the mineral type and mineral content are obtained by portable X-ray diffractometer (XRD) analysis.

[0098] The obtained optimal input feature data is input into the trained uniaxial compressive strength prediction model to obtain the prediction result of the uniaxial compressive strength of the rock mass.

[0099] The method of this embodiment integrates rock physical parameters, mineralogical parameters and drilling parameters to predict the uniaxial compressive strength of rock. It has higher universality and can be applied in various geotechnical engineering projects. It provides an important basis for surrounding rock quality classification, support structure design and construction method selection. It has broad application prospects and significant engineering value.

[0100] Embodiment 2

[0101] In one or more embodiments, a uniaxial compressive strength prediction system based on multi-source while-drilling information fusion is disclosed, comprising:

[0102] Data acquisition module, used to obtain rock physical parameters, mineralogical parameters and drilling parameters and corresponding uniaxial compressive strength during the historical drilling process;

[0103] An index system construction module is used to construct a uniaxial compressive strength prediction index system using rock physical parameters, mineralogical parameters and drilling parameters;

[0104] The optimal input feature determination module is used to select features of the uniaxial compressive strength prediction index system by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally determine the optimal input features for training the uniaxial compressive strength prediction model;

[0105] A model training module is used to train a uniaxial compressive strength prediction model using the optimal input feature as input and the corresponding uniaxial compressive strength as output, thereby obtaining a trained uniaxial compressive strength prediction model;

[0106] The uniaxial compressive strength prediction module is used to collect the optimal input features of the area to be tested, and use the trained uniaxial compressive strength prediction model to predict the uniaxial compressive strength of the rock in the area to be tested.

[0107] The specific implementation of the above modules is the same as that in Example 1 and will not be described in detail.

[0108] Embodiment 3

[0109] In one or more embodiments, a terminal device is disclosed, including a server, the server including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the uniaxial compressive strength prediction method based on multi-source while-drilling information fusion in Example 1 when executing the program. For the sake of brevity, it will not be described in detail here.

[0110] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0111] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0112] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0113] Embodiment 4

[0114] In one or more embodiments, a computer-readable storage medium is disclosed, in which a plurality of instructions are stored, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the uniaxial compressive strength prediction method based on multi-source while-drilling information fusion described in Example 1.

[0115] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A method for predicting uniaxial compressive strength based on multi-source while-drilling information fusion, characterized in that: include: Obtain rock physical parameters, mineralogical parameters and drilling parameters and corresponding uniaxial compressive strength during historical drilling; A uniaxial compressive strength prediction index system is constructed using rock physical parameters, mineralogical parameters and drilling parameters; The feature filtering algorithm and Pearson correlation coefficient were combined to select the features of the uniaxial compressive strength prediction index system, and finally the optimal input features for training the uniaxial compressive strength prediction model were determined; Taking the optimal input feature as input and the corresponding uniaxial compressive strength as output, the uniaxial compressive strength prediction model is trained to obtain a trained uniaxial compressive strength prediction model; Collect the optimal input features of the area to be tested, and use the trained uniaxial compressive strength prediction model to predict the uniaxial compressive strength of the rock in the area to be tested; Among them, the feature selection of the uniaxial compressive strength prediction index system is carried out by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally the optimal input features for training the uniaxial compressive strength prediction model are determined, specifically: the feature data in the uniaxial compressive strength prediction index system is normalized, the weight of each feature in the index system is calculated, the weights of all features are sorted, and the features with weights less than the set threshold are eliminated; the Pearson correlation coefficient between the remaining features is calculated, and when the correlation coefficient is greater than the coefficient threshold, it indicates that there is a strong correlation between the two features; for two features with strong correlation, one of them is deleted; and the remaining features are finally used as the optimal input features for training the uniaxial compressive strength prediction model; The uniaxial compressive strength prediction model includes a two-layer structure, the first layer is a plurality of different bottom-layer models, and the second layer is a top-layer model; the input data passes through each bottom-layer model to obtain the corresponding uniaxial compressive strength prediction results; the uniaxial compressive strength prediction results of all bottom-layer models are combined into new features, which are input into the top-layer model to obtain the final uniaxial compressive strength prediction results; wherein, both the bottom-layer model and the top-layer model are machine learning models.

2. The uniaxial compressive strength prediction method based on multi-source while-drilling information fusion according to claim 1, characterized in that: The rock physical parameters include point load intensity index, P wave velocity, rebound value and rock density; The mineralogical parameters include: mineral type, mineral content, mineral grain size and mineral morphology; The drilling parameters include: drilling response parameters such as drilling speed and torque, and drilling operation parameters such as drilling pressure and rotation speed set by the drilling rig operator.

3. The uniaxial compressive strength prediction method based on multi-source while-drilling information fusion according to claim 1, characterized in that: The training process of the uniaxial compressive strength prediction model is as follows: Construct the initial training set and test set using the optimal input features and the corresponding uniaxial compressive strength; Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and the prediction results of all underlying models are combined to determine the new training set and test set; Use the new training set to train the top-level model, and use the new test set to test it to obtain the trained top-level model; According to the trained multiple bottom-level models and the trained top-level model, a trained uniaxial compressive strength prediction model is obtained.

4. The uniaxial compressive strength prediction method based on multi-source while-drilling information fusion according to claim 3 is characterized in that: Also includes: During the training process of the uniaxial compressive strength prediction model, the determination coefficient, root mean square error and mean absolute percentage error are used as evaluation indicators to compare the uniaxial compressive strength of rock predicted by the uniaxial compressive strength prediction model with the measured value.

5. A uniaxial compressive strength prediction system based on multi-source while-drilling information fusion, characterized in that: include: Data acquisition module, used to obtain rock physical parameters, mineralogical parameters and drilling parameters and corresponding uniaxial compressive strength during the historical drilling process; An index system construction module is used to construct a uniaxial compressive strength prediction index system using rock physical parameters, mineralogical parameters and drilling parameters; The optimal input feature determination module is used to select features of the uniaxial compressive strength prediction index system by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally determine the optimal input features for training the uniaxial compressive strength prediction model; A model training module is used to train a uniaxial compressive strength prediction model using the optimal input feature as input and the corresponding uniaxial compressive strength as output, thereby obtaining a trained uniaxial compressive strength prediction model; The uniaxial compressive strength prediction module is used to collect the optimal input features of the area to be tested, and use the trained uniaxial compressive strength prediction model to predict the uniaxial compressive strength of the rock in the area to be tested; Among them, the feature selection of the uniaxial compressive strength prediction index system is carried out by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally the optimal input features for training the uniaxial compressive strength prediction model are determined, specifically: the feature data in the uniaxial compressive strength prediction index system is normalized, the weight of each feature in the index system is calculated, the weights of all features are sorted, and the features with weights less than the set threshold are eliminated; the Pearson correlation coefficient between the remaining features is calculated, and when the correlation coefficient is greater than the coefficient threshold, it indicates that there is a strong correlation between the two features; for two features with strong correlation, one of them is deleted; and the remaining features are finally used as the optimal input features for training the uniaxial compressive strength prediction model; The uniaxial compressive strength prediction model includes a two-layer structure, the first layer is a plurality of different bottom-layer models, and the second layer is a top-layer model; the input data passes through each bottom-layer model to obtain the corresponding uniaxial compressive strength prediction results; the uniaxial compressive strength prediction results of all bottom-layer models are combined into new features, which are input into the top-layer model to obtain the final uniaxial compressive strength prediction results; wherein, both the bottom-layer model and the top-layer model are machine learning models.

6. A terminal device, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the uniaxial compressive strength prediction method based on multi-source while-drilling information fusion as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instructions are suitable for being loaded by a processor of a terminal device and executing the uniaxial compressive strength prediction method based on multi-source drilling information fusion as described in any one of claims 1 to 4.

Images