Rock mass integrity evaluation method and system based on fluctuation characteristics of drilling parameters

Abstract

The present invention belongs to the field of advanced drilling, and provides a method and system for evaluating rock integrity based on the fluctuation characteristics of drilling parameters, including obtaining relevant parameters for rock integrity evaluation - original drilling parameters, rock P-wave velocity and rock P-wave velocity, and calculating the rock integrity coefficient based on the rock P-wave velocity and rock P-wave velocity; preprocessing the original drilling parameters, extracting fluctuation characteristics based on the preprocessed drilling parameters, namely, time domain characteristics and frequency domain characteristics; performing feature selection on the time domain characteristics and frequency domain characteristics, and finally determining the optimal input characteristics for training a rock integrity evaluation model; using the optimal input characteristics as the model input and the corresponding rock integrity coefficient as the model output, the rock integrity evaluation model is trained. The present invention integrates the time domain characteristics and frequency domain characteristics of the drilling parameters to evaluate rock integrity, which can make the rock integrity evaluation more refined and accurate.

Description

Technical Field

[0001] The invention belongs to the technical field of advanced drilling, and in particular relates to a rock mass integrity evaluation method and system based on fluctuation characteristics of drilling parameters. 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] In the field of tunnel construction, natural rock masses often contain a large number of joints and fissures due to geological tectonic effects, weathering, erosion, etc., which can easily reduce the strength and stability of the rock mass and cause engineering disasters such as rock bursts, landslides, and sudden water and mud bursts. How to quickly and accurately obtain the integrity of the rock mass is an effective guarantee for the safe and efficient construction of geotechnical engineering.

[0004] Existing technologies mostly use geophysical exploration, coring and other methods to qualitatively identify the integrity of the surrounding rock, but such technologies are often time-consuming, labor-intensive, and complex, affecting the progress of on-site construction. Studies have found that there is a good correlation between drilling parameters and rock mass integrity. A large number of scholars have also conducted correlation analysis between drilling parameters and rock mass integrity, and proposed corresponding prediction models and schemes, but most of these schemes directly use drilling parameters to predict rock mass integrity, ignoring the fluctuation response characteristics of drilling parameters when passing through fractured rock mass and intact rock mass.

[0005] The inventor combined the existing drilling test research and found that when drilling intact rock and fractured rock, the drilling speed and torque will have response characteristics such as mutation and fluctuation. With the different nature of the fracture, its fluctuation response characteristics will also be different. Therefore, it is of great research significance to evaluate the integrity of the rock mass using the fluctuation characteristics of the drilling parameters. In summary, when predicting the integrity of the rock mass, the influence of the fluctuation characteristics of the drilling parameters on the integrity of the rock mass when drilling the complete rock mass and fractured rock mass is ignored. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes a rock integrity evaluation method and system based on the fluctuation characteristics of drilling parameters. The present invention comprehensively analyzes the time domain fluctuation characteristics and frequency domain fluctuation characteristics of the drilling parameters to establish a rock integrity evaluation model, which can evaluate the integrity of the unexcavated rock mass ahead in real time in combination with the drilling parameters, and make timely adjustments to the support and construction plans according to the rapid on-site evaluation results, thereby improving the safety of construction, reducing construction costs and improving construction efficiency.

[0007] According to some embodiments, a first solution of the present invention provides a rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters, using the following technical solutions:

[0008] The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters includes:

[0009] Obtain relevant parameters for rock mass integrity evaluation, including original while-drilling parameters, rock mass longitudinal wave velocity and rock longitudinal wave velocity, and calculate the rock mass integrity coefficient based on the rock mass longitudinal wave velocity and rock longitudinal wave velocity;

[0010] Preprocessing the original while-drilling parameters to obtain preprocessed while-drilling parameters;

[0011] Based on the pre-processed drilling parameters, the fluctuation characteristics closely related to the integrity of the rock mass are extracted, namely, the time domain characteristics and the frequency domain characteristics;

[0012] The feature filtering algorithm and Pearson correlation coefficient are combined to select the extracted time domain features and frequency domain features, and finally determine the optimal input features for training the rock mass integrity evaluation model;

[0013] The optimal input feature after optimization is used as the model input, and the corresponding rock mass integrity coefficient is used as the model output, and the rock mass integrity evaluation model is trained to obtain a trained rock mass integrity evaluation model;

[0014] The drilling parameters of the unexcavated section are collected, and the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model.

[0015] Furthermore, the original drilling parameters are preprocessed to obtain the preprocessed drilling parameters, specifically:

[0016] The sampling frequency of the original LWD parameters is fixed by using the interpolation method;

[0017] Data elimination and data noise reduction are performed on the original drilling parameters, drilling parameters that have no interpretation value are eliminated, and signal noise reduction is performed on null values, short-term sudden change values, and drilling data with irregularity and roughness;

[0018] Taking into account the timeliness, operability and prediction accuracy of rock mass structure division, the interval of segment division is selected, and the drilling parameters after elimination and noise reduction are divided into segments to obtain the preprocessed drilling parameters.

[0019] Furthermore, the time domain characteristics reflect the variation law of the drilling signal in the time domain, including the maximum value, minimum value, peak-to-peak value, average value, root mean square value, standard deviation, and total signal value describing the intensity characteristics of the time domain drilling signal;

[0020] The kurtosis, skewness, peak factor, pulse factor, waveform index and margin factor that characterize the waveform and change characteristics of the time domain drilling signal.

[0021] Furthermore, the time-domain drilling signal is converted into a frequency-domain drilling signal using a time-frequency analysis method, and the frequency-domain feature extraction is performed after obtaining the corresponding spectrum diagram to obtain the frequency-domain feature of the drilling signal;

[0022] The frequency domain characteristics of the drilling signal include the frequency domain amplitude average value, the center of gravity frequency, the mean square frequency and the frequency variance.

[0023] Furthermore, the feature filtering algorithm and the Pearson correlation coefficient are combined to perform feature selection on the extracted time domain features and frequency domain features, and finally determine the optimal input features for training the rock mass integrity evaluation model, specifically:

[0024] Determine the initial input feature index and normalize the data set, then use the feature filtering algorithm to calculate and rank the impact of each input feature on the output feature, that is, the weight of the input feature index, determine the threshold and eliminate the input feature index below the threshold;

[0025] Calculate the Pearson correlation coefficient between the remaining input feature indexes. When the Pearson correlation coefficient is greater than the coefficient threshold, it indicates that there is a strong correlation between the input features.

[0026] One input feature with strong correlation is deleted, and finally the optimal input feature for training the rock mass integrity evaluation model is determined;

[0027] The optimal input features include optimal time domain features and optimal frequency domain features.

[0028] Furthermore, the rock mass integrity evaluation model includes two layers, the first layer is a bottom layer model of several different mathematical principles, and the second layer selects the top layer model;

[0029] Among them, the underlying model and the top model are both machine learning algorithms.

[0030] Furthermore, the optimal input feature after optimization is used as the model input, and the corresponding rock mass integrity coefficient is used as the model output to train the rock mass integrity evaluation model to obtain the trained rock mass integrity evaluation model, specifically:

[0031] The initial training set and test set are constructed using the optimal input features and rock mass integrity coefficients;

[0032] Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and new training sets and test sets are determined based on the combination of multiple prediction results;

[0033] 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;

[0034] According to the trained multiple bottom models and the trained top model, a trained rock mass integrity evaluation model is obtained;

[0035] A variety of evaluation indicators are selected to evaluate the performance of the trained rock integrity evaluation model.

[0036] Furthermore, the performance of the trained rock mass integrity evaluation model is evaluated by selecting a plurality of evaluation indicators, specifically:

[0037] The determination coefficient, root mean square error and mean absolute percentage error were used as evaluation indicators to compare the rock integrity predicted by the rock integrity evaluation model with the actual rock integrity, and then the performance of the trained rock integrity evaluation model was evaluated.

[0038] Furthermore, the drilling parameters of the unexcavated section are collected, and the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model, specifically:

[0039] Collect the drilling parameters of the unexcavated section and pre-process them;

[0040] Extracting optimal input features as input parameters based on the preprocessed while-drilling parameters, wherein the optimal input features include optimal time domain features and optimal frequency domain features;

[0041] Based on the input parameters, the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model.

[0042] According to some embodiments, a second solution of the present invention provides a rock mass integrity evaluation system based on the fluctuation characteristics of drilling parameters, which adopts the following technical solutions:

[0043] The rock mass integrity evaluation system based on the fluctuation characteristics of drilling parameters includes:

[0044] A parameter acquisition module is configured to acquire relevant parameters for rock mass integrity evaluation, including original while-drilling parameters, rock mass longitudinal wave velocity and rock longitudinal wave velocity, and calculate a rock mass integrity coefficient based on the rock mass longitudinal wave velocity and rock longitudinal wave velocity;

[0045] The drilling parameter preprocessing module is configured to preprocess the original drilling parameters to obtain the preprocessed drilling parameters;

[0046] A feature extraction module is configured to extract fluctuation features closely related to rock mass integrity based on the preprocessed while-drilling parameters, namely, time domain features and frequency domain features;

[0047] A feature optimization module is configured to select the extracted time domain features and frequency domain features by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally determine the optimal input features for training the rock mass integrity evaluation model;

[0048] The model training module is configured to use the optimal input feature after optimization as the model input and the corresponding rock mass integrity coefficient as the model output, train the rock mass integrity evaluation model, and obtain the trained rock mass integrity evaluation model;

[0049] The rock mass integrity evaluation module collects the drilling parameters of the unexcavated section and uses the trained rock mass integrity evaluation model to evaluate the integrity of the unexcavated rock mass.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The present invention integrates the time domain characteristics and frequency domain characteristics of the drilling parameters to evaluate the rock mass integrity, which can make the rock mass integrity evaluation more refined and accurate compared to directly using the drilling parameters to evaluate the rock mass integrity.

[0052] The present invention uses an ensemble learning algorithm instead of a single machine learning algorithm, combines the advantages of multiple underlying models, improves prediction performance and stability, reduces the risk of overfitting, and has greater flexibility and adaptability.

[0053] The method of the invention can evaluate the rock integrity in real time during the drilling process. Compared with obtaining the rock integrity through indoor tests by coring holes, it has significant economy and timeliness, broad application prospects and significant engineering value. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0055] Figure 1 It is a flow chart of a rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters in an embodiment of the present invention;

[0056] Figure 2 is a training flow chart of a rock mass integrity evaluation method based on fluctuation characteristics of drilling parameters in an embodiment of the present invention;

[0057] Figure 3 It is a feature selection flow chart in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0059] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0060] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. 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 "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0061] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0062] Embodiment 1

[0063] like Figure 1 As shown, this embodiment provides a rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters, and the method includes the following steps:

[0064] Obtain relevant parameters for rock mass integrity evaluation, including original while-drilling parameters, rock mass longitudinal wave velocity and rock longitudinal wave velocity, and calculate the rock mass integrity coefficient based on the rock mass longitudinal wave velocity and rock longitudinal wave velocity;

[0065] Preprocessing the original while-drilling parameters to obtain preprocessed while-drilling parameters;

[0066] Based on the pre-processed drilling parameters, the fluctuation characteristics closely related to the integrity of the rock mass are extracted, namely, the time domain characteristics and the frequency domain characteristics;

[0067] The feature filtering algorithm and Pearson correlation coefficient are combined to select the extracted time domain features and frequency domain features, and finally determine the optimal input features for training the rock mass integrity evaluation model;

[0068] The optimal input feature after optimization is used as the model input, and the corresponding rock mass integrity coefficient is used as the model output, and the rock mass integrity evaluation model is trained to obtain a trained rock mass integrity evaluation model;

[0069] The drilling parameters of the unexcavated section are collected, and the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model.

[0070] like Figure 2 As shown, this embodiment provides a training process of a rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters, including the following steps:

[0071] Step 1: Pre-processing of drilling parameters:

[0072] Firstly, the sampling frequency of the original LWD parameters is fixed by the interpolation method, so as to facilitate the subsequent time domain and frequency domain analysis and feature extraction of the fluctuation signal characteristics of each LWD parameter; then data elimination and data noise reduction are performed to improve the data quality of the LWD parameters; finally, segmentation is performed.

[0073] The data elimination refers to eliminating drilling data that has no interpretation value, such as data in the initial stage of drilling. At this time, due to the influence of factors such as idling of the drilling rig, opening holes and passing through the initial concrete spraying section, the collected drilling data has no interpretation value;

[0074] The data denoising is to perform signal denoising on null values, transient sudden change values, and drilling data with irregularity and roughness, so as to improve the data quality of drilling parameters.

[0075] The segment division is to comprehensively consider the timeliness, operability and prediction accuracy of the rock mass structure division, select the interval of the segment division, and divide the drilling data into equal intervals so as to study the drilling data characteristics and the corresponding surrounding rock integrity characteristics in sections.

[0076] It should be noted that, during the initial data collection, in addition to the original downhole parameters, it is also necessary to obtain the rock mass longitudinal wave velocity and the rock longitudinal wave velocity, and use the rock mass longitudinal wave velocity and the rock longitudinal wave velocity to calculate the rock mass integrity coefficient as the output parameter of the rock mass integrity evaluation model.

[0077] Step 2: Extract the fluctuation characteristics closely related to the rock mass integrity evaluation:

[0078] Comprehensively considering the influence of time domain characteristics and frequency domain characteristics of drilling signals on the evaluation of rock integrity, the fluctuation characteristics closely related to the rock integrity, namely, time domain characteristics and frequency domain characteristics, are extracted based on the preprocessed while-drilling parameters, and the time domain characteristics of drilling signals and the frequency domain characteristics of drilling signals that are closely related to the rock integrity are selected.

[0079] The time domain characteristics reflect the variation law of the drilling signal in the time domain, including the maximum value, minimum value, peak-to-peak value, average value, root mean square value, standard deviation, and total signal value that describe the intensity characteristics of the drilling signal; the kurtosis, skewness, peak factor, pulse factor, waveform index, and margin factor that characterize the waveform and variation characteristics of the drilling signal;

[0080] The technical solution for extracting frequency domain features is to use a time-frequency analysis method to convert the time domain drilling signal into the frequency domain, obtain the corresponding spectrum diagram, and then perform frequency domain feature extraction to obtain the frequency domain features of the drilling signal;

[0081] The frequency domain characteristics of the drilling signal include the frequency domain amplitude average value, the center of gravity frequency, the mean square frequency and the frequency variance.

[0082] Extracting fluctuation characteristics is to extract the initial input characteristics of the rock mass integrity evaluation model; however, whether these characteristics have an important impact on the rock mass integrity, or whether there is multicollinearity between the characteristics, these need to be selected through feature selection to determine the final input parameters.

[0083] Step 3: Feature selection:

[0084] like Figure 3 As shown in the figure, the feature filtering algorithm is used to test and select the validity of the initial input features, eliminate the input features with little influence on the uniaxial compressive strength of rock, and delete the input features with strong correlation, so as to determine the optimal input features for the training of the rock integrity evaluation model.

[0085] The feature filtering algorithm and the Pearson correlation coefficient are combined to perform feature selection on the extracted time domain features and frequency domain features, and finally determine the optimal input features for training the rock mass integrity evaluation model, specifically:

[0086] Determine the initial input feature index and normalize the data set, then use the feature filtering algorithm to calculate and rank the impact of each input feature on the output feature, that is, the weight of the input feature index, determine the threshold and eliminate the input feature index below the threshold, that is, generally delete the input features with little impact on the output feature;

[0087] Calculate the Pearson correlation coefficient between the remaining input feature indexes. When the Pearson correlation coefficient is greater than the coefficient threshold, it indicates that there is a strong correlation between the input features.

[0088] One input feature with strong correlation is deleted, and finally the optimal input feature for training the rock mass integrity evaluation model is determined;

[0089] The optimal input features include optimal time domain features and optimal frequency domain features.

[0090] For example, it includes but is not limited to using a feature optimization algorithm to calculate and rank the weight of each input feature index, then determining a threshold and eliminating input feature indexes below the threshold;

[0091] Calculate the correlation coefficients between the remaining input feature indexes. When the correlation coefficient between the features is greater than the coefficient threshold, it means that there is a strong correlation between the features. To reduce the complexity of the model, delete one of the two input features with strong correlation, determine the best feature index data set, and finally use it to predict the input features of the model training.

[0092] Step 4: Construction of rock mass integrity evaluation model:

[0093] The rock mass integrity evaluation model includes two layers, the first layer is a bottom model based on several different mathematical principles, and the second layer uses a top model; wherein both the bottom model and the top model are machine learning algorithms;

[0094] The optimal input features after optimization are used as model input, and the corresponding rock integrity coefficient is used as model output to train the rock integrity evaluation model. Specifically:

[0095] The initial training set and test set are constructed using the optimal input features and rock mass integrity coefficients;

[0096] Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and new training sets and test sets are determined based on the combination of multiple prediction results;

[0097] 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;

[0098] According to the trained multiple bottom models and the trained top model, a trained rock mass integrity evaluation model is obtained;

[0099] A variety of evaluation indicators are selected to evaluate the performance of the trained rock integrity evaluation model.

[0100] A rock mass integrity evaluation model is built using machine learning algorithms based on a variety of different mathematical principles. The model consists of two layers. The first layer is a bottom model based on several different mathematical principles. In order to avoid overfitting, the second layer is a simple linear top model. By training multiple bottom models to obtain multiple prediction results, these prediction results are combined into new features, and then these new features are used as inputs for training the top model, so that a more accurate and stable prediction result is finally obtained.

[0101] This step includes optimizing the hyperparameters of the underlying model, training the underlying model and obtaining new training sets and test sets, and training the meta-learner.

[0102] Optimizing the hyperparameters of the underlying model includes using an optimization algorithm to optimize the hyperparameters of the underlying model to find the best hyperparameters of the underlying model;

[0103] The technical solution for training the underlying models is to use k-fold cross validation to train each underlying model and obtain new training sets and test sets, finally completing the training of all underlying models and finally obtaining new feature matrices required for multiple groups of top-level models.

[0104] The training of the top-level model mainly involves training the top-level model based on the newly constructed training set combination, predicting the new test set combination, and finally outputting the final result.

[0105] Step 5: Model evaluation:

[0106] A variety of evaluation indicators are selected to evaluate the performance of the trained rock integrity evaluation model.

[0107] Specifically, the coefficient of determination (R 2 ), root mean square error (RMSE) and mean absolute percentage error (MAPE) were used as evaluation indicators to compare the rock integrity predicted by the rock integrity evaluation model with the actual rock integrity.

[0108] The R 2 Indicates the accuracy of the model fitting the data. The closer it is to 1, the better the fitting effect.

[0109] The 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.

[0110] The MAPE represents 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.

[0111] Step 6: Rock mass integrity assessment:

[0112] The evaluation model after training and evaluation is applied to the actual project to carry out the integrity evaluation of the unexcavated rock mass ahead, including:

[0113] Collect the drilling parameters of the unexcavated section and pre-process them;

[0114] Extracting optimal input features as input parameters based on the preprocessed while-drilling parameters, wherein the optimal input features include optimal time domain features and optimal frequency domain features;

[0115] Based on the input parameters, the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model.

[0116] In actual engineering, the drilling parameters are obtained, and the time domain and frequency domain features are extracted. The integrity evaluation of the unexcavated rock mass ahead is carried out using the rock integrity evaluation model after training and evaluation.

[0117] Embodiment 2

[0118] This embodiment provides a rock mass integrity evaluation system based on the fluctuation characteristics of drilling parameters, including:

[0119] A parameter acquisition module is configured to acquire relevant parameters for rock mass integrity evaluation, including original while-drilling parameters, rock mass longitudinal wave velocity and rock longitudinal wave velocity, and calculate a rock mass integrity coefficient based on the rock mass longitudinal wave velocity and rock longitudinal wave velocity;

[0120] The drilling parameter preprocessing module is configured to preprocess the original drilling parameters to obtain the preprocessed drilling parameters;

[0121] A feature extraction module is configured to extract fluctuation features closely related to rock mass integrity based on the preprocessed while-drilling parameters, namely, time domain features and frequency domain features;

[0122] A feature optimization module is configured to select the extracted time domain features and frequency domain features by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally determine the optimal input features for training the rock mass integrity evaluation model;

[0123] The model training module is configured to use the optimal input feature after optimization as the model input and the corresponding rock mass integrity coefficient as the model output, train the rock mass integrity evaluation model, and obtain the trained rock mass integrity evaluation model;

[0124] The rock mass integrity evaluation module collects the drilling parameters of the unexcavated section and uses the trained rock mass integrity evaluation model to evaluate the integrity of the unexcavated rock mass.

[0125] The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment 1. It should be noted that the above modules as part of the system can be executed in a computer system such as a set of computer executable instructions.

[0126] The description of each embodiment in the above embodiments has different emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0127] The proposed system can be implemented in other ways. For example, the system embodiment described above is only illustrative, and the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0128] 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 rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters, characterized in that: include: Obtain relevant parameters for rock mass integrity evaluation, including original while-drilling parameters, rock mass longitudinal wave velocity and rock longitudinal wave velocity, and calculate the rock mass integrity coefficient based on the rock mass longitudinal wave velocity and rock longitudinal wave velocity; Preprocessing the original while-drilling parameters to obtain preprocessed while-drilling parameters; Based on the pre-processed drilling parameters, the fluctuation characteristics closely related to the integrity of the rock mass are extracted, namely, the time domain characteristics and the frequency domain characteristics; The feature filtering algorithm and the Pearson correlation coefficient are combined to perform feature selection on the extracted time domain features and frequency domain features, and finally determine the optimal input features for training the rock mass integrity evaluation model; the rock mass integrity evaluation model includes two layers, the first layer is a bottom model of several different mathematical principles, and the second layer selects a simple model as the top model; wherein the bottom model and the top model are both machine learning algorithms; The optimal input features after optimization are used as model input, and the corresponding rock integrity coefficient is used as model output. The rock integrity evaluation model is trained to obtain the trained rock integrity evaluation model, which is specifically: The initial training set and test set are constructed using the optimal input features and rock mass integrity coefficients; Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and new training sets and test sets are determined based on the combination of multiple prediction results; 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 models and the trained top model, a trained rock mass integrity evaluation model is obtained; Select a variety of evaluation indicators to evaluate the performance of the trained rock mass integrity evaluation model; The drilling parameters of the unexcavated section are collected, and the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model.

2. The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters according to claim 1, characterized in that: The original drilling parameters are preprocessed to obtain the preprocessed drilling parameters, specifically: The sampling frequency of the original LWD parameters is fixed by using the interpolation method; Data elimination and data noise reduction are performed on the original drilling parameters, drilling parameters that have no interpretation value are eliminated, and signal noise reduction is performed on null values, short-term sudden change values, and drilling data with irregularity and roughness; Taking into account the timeliness, operability and prediction accuracy of rock mass structure division, the interval of segment division is selected, and the drilling parameters after elimination and noise reduction are divided into segments to obtain the preprocessed drilling parameters.

3. The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters according to claim 1, characterized in that: The time domain characteristics reflect the variation law of the drilling signal in the time domain, including the maximum value, minimum value, peak-to-peak value, average value, root mean square value, standard deviation, and total signal value that describe the intensity characteristics of the time domain drilling signal; The kurtosis, skewness, peak factor, pulse factor, waveform index and margin factor that characterize the waveform and change characteristics of the time domain drilling signal.

4. The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters according to claim 3, characterized in that: The time-domain drilling signal is converted into the frequency-domain drilling signal using the time-frequency analysis method, and the frequency-domain feature extraction is performed after obtaining the corresponding spectrum diagram to obtain the frequency-domain feature of the drilling signal; The frequency domain characteristics of the drilling signal include the frequency domain amplitude average value, the center of gravity frequency, the mean square frequency and the frequency variance.

5. The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters according to claim 1, characterized in that: The feature filtering algorithm and the Pearson correlation coefficient are combined to perform feature selection on the extracted time domain features and frequency domain features, and finally determine the optimal input features for training the rock mass integrity evaluation model, specifically: Determine the initial input feature index and normalize the data set, then use the feature filtering algorithm to calculate and rank the impact of each input feature on the output feature, that is, the weight of the input feature index, determine the threshold and eliminate the input feature index below the threshold; Calculate the Pearson correlation coefficient between the remaining input feature indexes. When the Pearson correlation coefficient is greater than the coefficient threshold, it indicates that there is a strong correlation between the input features. One input feature with strong correlation is deleted, and finally the optimal input feature for training the rock mass integrity evaluation model is determined; Among them, the optimal input features include optimal time domain features and optimal frequency domain features.

6. The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters according to claim 1, characterized in that: The performance of the trained rock mass integrity evaluation model is evaluated by selecting a plurality of evaluation indicators, specifically: The determination coefficient, root mean square error and mean absolute percentage error were used as evaluation indicators to compare the rock integrity predicted by the rock integrity evaluation model with the actual rock integrity, and then the performance of the trained rock integrity evaluation model was evaluated.

7. The rock mass integrity evaluation method based on the fluctuation characteristics of drilling parameters according to claim 1, characterized in that: The drilling parameters of the unexcavated section are collected, and the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model, specifically: Collect the drilling parameters of the unexcavated section and pre-process them; Extracting optimal input features as input parameters based on the preprocessed while-drilling parameters, wherein the optimal input features include optimal time domain features and optimal frequency domain features; Based on the input parameters, the integrity of the unexcavated rock mass is evaluated using the trained rock mass integrity evaluation model.

8. A rock mass integrity evaluation system based on the fluctuation characteristics of drilling parameters is characterized by: include: A parameter acquisition module is configured to acquire relevant parameters for rock mass integrity evaluation, including original while-drilling parameters, rock mass longitudinal wave velocity and rock longitudinal wave velocity, and calculate a rock mass integrity coefficient based on the rock mass longitudinal wave velocity and rock longitudinal wave velocity; A drilling parameter preprocessing module is configured to preprocess the original drilling parameters to obtain preprocessed drilling parameters; A feature extraction module is configured to extract fluctuation features closely related to rock mass integrity based on the preprocessed while-drilling parameters, namely, time domain features and frequency domain features; The feature optimization module is configured to select the extracted time domain features and frequency domain features by combining the feature filtering algorithm and the Pearson correlation coefficient, and finally determine the optimal input features for training the rock mass integrity evaluation model; the rock mass integrity evaluation model includes two layers, the first layer is a bottom layer model of several different mathematical principles, and the second layer selects a simple model as a top layer model; wherein the bottom layer model and the top layer model are both machine learning algorithms; The model training module is configured to use the optimal input features after optimization as model input and the corresponding rock mass integrity coefficient as model output to train the rock mass integrity evaluation model and obtain the trained rock mass integrity evaluation model, specifically: The initial training set and test set are constructed using the optimal input features and rock mass integrity coefficients; Based on the initial training set and test set, k-fold cross validation is used to train each underlying model, and new training sets and test sets are determined based on the combination of multiple prediction results; 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 models and the trained top model, a trained rock mass integrity evaluation model is obtained; Select a variety of evaluation indicators to evaluate the performance of the trained rock mass integrity evaluation model; The rock mass integrity evaluation module collects the drilling parameters of the unexcavated section and uses the trained rock mass integrity evaluation model to evaluate the integrity of the unexcavated rock mass.

Images