Rock classification method based on sliding window feature extraction and random forest algorithm

Abstract

The application discloses a kind of TBM tunnel surrounding rock grade identification method based on sliding window feature extraction and weight self-adaptive random forest algorithm, steps are as follows: obtaining preprocessed data;The preprocessed data is intercepted using sliding window, and features are extracted from the intercepted data to obtain a feature dataset;Establish training set, validation set and test set;TBM tunnel surrounding rock grade identification model is constructed;Determine the optimal sliding window length and step, find the optimal model hyperparameter using search algorithm, to obtain the optimized TBM tunnel surrounding rock grade identification model;Obtain the weight adaptive TBM tunnel surrounding rock grade identification model;The surrounding rock grade is identified using the weight adaptive TBM tunnel surrounding rock grade identification model.The method of the application can extract surrounding rock grade related information from the excavation parameter change and can adaptively consider the influence of class weight on model performance, which can greatly improve the performance of TBM tunnel surrounding rock grade identification model, and has positive significance for guiding TBM construction.

Description

Technical Field

[0001] This invention relates to a method for identifying the surrounding rock grade of hard rock TBM tunnels, belonging to the field of tunnel surrounding rock grade identification. Background Technology

[0002] Compared to traditional drill-and-blast methods, hard rock tunnel boring machine (TBM) construction offers advantages such as faster excavation speed, higher construction quality, and less environmental impact. However, TBM performance is highly dependent on surrounding rock conditions, and it cannot effectively withdraw when encountering unfavorable geological conditions. Due to limitations in the TBM's own structural design, operators typically cannot directly observe the tunnel face, thus hindering the acquisition of information such as rock type and rock mass grade, compromising the safe and efficient operation of the TBM. Accurately perceiving the geological conditions of the tunnel face is fundamental for conducting important tasks such as surrounding rock stability analysis, TBM tunneling performance prediction, and early warning of machine jams and rockburst hazards. Based on data sources, indirect TBM tunnel surrounding rock condition perception methods can be categorized into perception based on TBM tunneling data, perception based on vibration data, perception based on rock debris images, and perception based on advanced geological prediction data. Among these, perception based on TBM tunneling data is currently the mainstream method.

[0003] However, current methods for sensing surrounding rock conditions based on TBM tunneling data generally use the selected tunneling parameters as model inputs directly. This fails to consider the geological information contained in the changes of tunneling parameters, and the corresponding machine learning and deep learning modeling methods rarely consider the impact of the weights of different categories of samples on model performance. Summary of the Invention

[0004] To address the shortcomings of current methods for sensing surrounding rock conditions based on TBM tunneling data, which fail to consider geological information contained in changes in tunneling parameters and rarely consider the impact of category weights on model performance, this invention provides a method for identifying the surrounding rock grade of TBM tunnels based on sliding window feature extraction and a weighted adaptive random forest algorithm.

[0005] To effectively solve the above problems, the technical solution provided by the present invention is as follows:

[0006] Firstly, a method for identifying surrounding rock grades based on sliding window feature extraction and random forest algorithm is provided, including the following steps:

[0007] Obtain the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data;

[0008] The preprocessed data is truncated using a sliding window, and features are extracted from the truncated data to obtain a feature dataset.

[0009] Using the feature dataset as input and the surrounding rock grade at the sliding window truncation time as output, a training set, a validation set, and a test set are established.

[0010] Construct a TBM tunnel surrounding rock grade identification model based on the random forest algorithm;

[0011] The optimal sliding window length and step size are determined, and the optimal model hyperparameters are found using a search algorithm to obtain an optimized TBM tunnel surrounding rock grade identification model.

[0012] The optimal weights for each category in the random forest model are found using a global optimization algorithm, resulting in a weight-adaptive TBM tunnel surrounding rock grade identification model.

[0013] A weighted adaptive TBM tunnel surrounding rock grade identification model is used to identify the surrounding rock grade.

[0014] In one possible implementation, the raw data collected by the TBM includes: tunneling parameters and corresponding surrounding rock grades; the tunneling parameters include thrust, cutterhead torque, cutterhead rotation speed, propulsion speed, guide roll angle, guide pitch angle, support shoe pump pressure, support pressure, main conveyor speed, and main conveyor pump pressure.

[0015] The preprocessing includes: screening tunneling parameters based on empirical and random forest algorithms, outlier detection based on isolated forest algorithm, and data normalization based on sensor range.

[0016] One possible implementation method involves using a sliding window to truncate the preprocessed data and extracting features from the truncated data to obtain a feature dataset, as follows:

[0017] The parameters of the sliding window are set, and the preprocessed data is truncated using the sliding window. Features are then extracted from the truncated data to obtain a feature dataset. Specifically, the features extracted from the truncated data include the mean, standard deviation, maximum value, minimum value, quantiles of each tunneling parameter, and the Pearson correlation coefficient between the tunneling parameters.

[0018] One possible implementation method for determining the optimal sliding window length and step size is as follows: the optimal sliding window length and step size are determined by utilizing the performance of the TBM tunnel surrounding rock grade identification model based on the random forest algorithm on the validation set.

[0019] Furthermore, the model hyperparameters include: the number of weak learners, the maximum tree depth, and the maximum number of features.

[0020] One possible implementation method for finding the optimal weights for each category of the random forest model using a global optimization algorithm is as follows: using the performance index of the optimized TBM tunnel surrounding rock grade identification model on the validation set as the fitness function, the optimal weights for each category of the random forest model are found using a global optimization algorithm.

[0021] Furthermore, the performance metrics are one or a combination of several of Precision, Recall, F-measure, and Accuracy.

[0022] Secondly, a rock mass classification device based on sliding window feature extraction and random forest algorithm is provided, including:

[0023] The preprocessing data acquisition module is used to acquire the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data.

[0024] The sliding window extraction module is used to truncate preprocessed data using a sliding window and extract features from the truncated data to obtain a feature dataset.

[0025] The dataset construction module is used to build training, validation and test sets by taking the feature dataset as input and the surrounding rock grade at the sliding window truncation time as output.

[0026] The primary model building module is used to build a TBM tunnel surrounding rock grade identification model based on the random forest algorithm;

[0027] The model optimization module is used to determine the optimal sliding window length and step size, and to use a search algorithm to find the optimal model hyperparameters to obtain an optimized TBM tunnel surrounding rock grade identification model.

[0028] The final model acquisition module is used to find the optimal weights for each category of the random forest model using a global optimization algorithm, and obtain a weight-adaptive TBM tunnel surrounding rock grade identification model.

[0029] The identification module is used to identify the surrounding rock grade using a weighted adaptive TBM tunnel surrounding rock grade identification model.

[0030] Thirdly, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the aforementioned method for identifying surrounding rock grades based on sliding window feature extraction and random forest algorithm.

[0031] Fourthly, a non-transitory computer-readable storage medium is provided, on which a computer program is stored, wherein when the computer program is executed by a processor, it implements the aforementioned method for identifying surrounding rock grades based on sliding window feature extraction and random forest algorithm.

[0032] Fifthly, a computer program product is provided, including a computer program that, when executed by a processor, implements the aforementioned method for identifying surrounding rock grades based on sliding window feature extraction and random forest algorithm.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The method provided by this invention extracts features related to the surrounding rock grade from the changes in TBM tunneling parameters, and can adaptively adjust and obtain the optimal category weights, significantly improving the performance of the TBM tunnel surrounding rock grade identification model, and ultimately ensuring safe and efficient TBM construction. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0036] Figure 1 The flowchart of the TBM tunnel surrounding rock grade identification method based on sliding window feature extraction and weighted adaptive random forest algorithm provided by the present invention is shown below.

[0037] Figure 2 This is a schematic diagram illustrating the performance of the random forest model on the test set by directly using the selected tunneling parameters as input and extracting features using a sliding window as input, according to an embodiment of the present invention.

[0038] Figure 3 This is a schematic diagram showing the performance comparison of various modeling methods in the embodiments of the present invention on the validation set and the test set.

[0039] Figure 4 This is a schematic diagram of the surrounding rock grade identification device based on sliding window feature extraction and random forest algorithm provided by the present invention;

[0040] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0042] refer to Figure 1 A method for identifying the surrounding rock grade of TBM tunnels based on sliding window feature extraction and weighted adaptive random forest algorithm includes the following steps:

[0043] Step 100: Obtain the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data.

[0044] As a specific embodiment, the raw data includes tunneling parameters and the corresponding surrounding rock grade.

[0045] Specifically, the tunneling parameters include thrust, cutterhead torque, cutterhead rotation speed, propulsion speed, guide roll angle, guide pitch angle, support shoe pump pressure, support pressure, main conveyor speed, and main conveyor pump pressure.

[0046] As a specific embodiment, the preprocessing includes: screening tunneling parameters based on experience and random forest algorithms, outlier detection based on isolated forest algorithms, and data normalization based on sensor range.

[0047] For example, tunneling parameters are screened based on experience and the random forest algorithm. The screened parameters are thrust, cutterhead torque, cutterhead speed, advance speed, guide roll angle, guide pitch angle, support shoe pump pressure, support pressure, main conveyor speed, and main conveyor pump pressure. Due to electromagnetic interference, transmission distortion, sensor quality, etc., outliers may appear in each collected tunneling parameter. Outliers are detected and deleted using the isolated forest algorithm. Since streaming data cannot obtain accurate maximum and minimum values, all tunneling parameters are normalized using the sensor range. The normalization formula is as follows: Where x i_norm Let x be the normalized value of the i-th tunneling parameter. i Let x be the monitoring value of the i-th tunneling parameter. i_max Let x be the upper limit of the sensor range corresponding to the i-th tunneling parameter. i_min This represents the lower limit of the sensor range corresponding to the i-th tunneling parameter.

[0048] Step 200: Use a sliding window to truncate the preprocessed data and extract features from the truncated data to obtain a feature dataset.

[0049] As a specific implementation, the method for using a sliding window to truncate preprocessed data and extracting features from the truncated data to obtain a feature dataset is as follows:

[0050] Set the parameters of the sliding window (window length, step size, etc.), use the sliding window to truncate the preprocessed data obtained by S100, and then extract features from the truncated data to obtain the feature dataset.

[0051] Specifically, the features extracted from the intercepted data include the mean, standard deviation, maximum value, minimum value, quantiles (20%, 40%, 60%, and 80% quantiles, used to describe the data distribution characteristics) of each tunneling parameter, as well as the Pearson correlation coefficient between tunneling parameters.

[0052] Understandably, the data captured by the sliding window is a series of two-dimensional arrays. The rows of the arrays represent data sampled at different times, and the columns represent various tunneling parameters. Step 300 uses the feature dataset as input and the surrounding rock grade at the sliding window capture time as output to establish training, validation, and test sets.

[0053] In step 300, it is assumed that all sliding windows within the same tunneling ring correspond to the same surrounding rock grade.

[0054] Step 400: Construct a TBM tunnel surrounding rock grade identification model based on the random forest algorithm.

[0055] Step 500: Determine the optimal sliding window length and step size, and use a search algorithm to find the optimal model hyperparameters to obtain the optimized TBM tunnel surrounding rock grade identification model.

[0056] As a specific implementation, the method for determining the optimal sliding window length and step size is as follows: the optimal sliding window length and step size are determined by using the performance of the TBM tunnel surrounding rock grade identification model based on the random forest algorithm on the validation set.

[0057] Understandably, performance on the validation set refers to the F measurement performance index of the surrounding rock grade identification model. The combination of window length and step size that maximizes the F measurement index is the optimal window length and step size.

[0058] For example, in this embodiment, the optimal window length is 32 and the optimal step size is 20.

[0059] As a specific implementation, the optimal model hyperparameters are found using a validation set and a grid search algorithm.

[0060] Specifically, the model hyperparameters include: number of weak learners, maximum tree depth, and maximum number of features.

[0061] Step 600: Use a global optimization algorithm to find the optimal weights for each category in the random forest model to obtain a weight-adaptive TBM tunnel surrounding rock grade identification model.

[0062] As a specific embodiment, the method for finding the optimal weights for each category of the random forest model using a global optimization algorithm is as follows: using the performance index of the optimized TBM tunnel surrounding rock grade identification model on the validation set as the fitness function, the optimal weights for each category of the random forest model are found using a global optimization algorithm.

[0063] Specifically, the performance metrics are one or a combination of Precision, Recall, F-measure, and Accuracy.

[0064] For example, in this embodiment, the model's classification index F-measure on the validation set is used as the fitness function. The particle swarm optimization algorithm (PSO) is used to find the optimal weights for each category of the random forest model. The optimal category weights for Class II to V surrounding rock are 1.05, 1.21, 1.83 and 0.87, respectively, to achieve adaptive adjustment of category weights and establish a TBM tunnel surrounding rock grade identification model based on the weight adaptive random forest algorithm.

[0065] Model reliability testing:

[0066] The reliability of the final TBM tunnel surrounding rock grade identification model was tested on the test set. The specific test steps are as follows:

[0067] (1) Use the final TBM tunnel surrounding rock grade identification model to predict the surrounding rock grade corresponding to all sliding windows in the tunneling ring;

[0068] (2) Statistical analysis of the percentage of each predicted surrounding rock grade within the tunneling ring;

[0069] (3) The final predicted surrounding rock grade is the surrounding rock grade with the largest proportion.

[0070] Results and analysis:

[0071] Figure 2 This is a schematic diagram illustrating the performance of the present invention on a test set, where selected tunneling parameters are directly used as input to a random forest model, and features extracted using a sliding window are used as input to the random forest model. Figure 2 It can be seen that after using the sliding window to extract features, the model's performance on the test set improved by 12.1%.

[0072] Figure 3 This diagram illustrates the performance comparison of the method provided in this embodiment of the invention and various comparative methods on the validation and test sets, where RF represents the Random Forest model and WA-RF represents the model described in this embodiment. Figure 3 As can be seen, the model described in this embodiment performs best on the test set, with an F-measure of 0.841, which is 1.9%, 4.9%, 4.3%, and 3.0% higher than the traditional random forest model, AdaBoost model, XGBoost model, and LightGBM, respectively.

[0073] This invention proposes a method for identifying the surrounding rock grade of TBM tunnels based on sliding window feature extraction and a weighted adaptive random forest algorithm. This method utilizes a sliding window to extract data segments and extracts geologically relevant features from data variations. Furthermore, the proposed weighted adaptive random forest algorithm adaptively considers the impact of class weights on model performance, significantly improving the accuracy of the surrounding rock grade identification model.

[0074] The rock grade identification device based on sliding window feature extraction and random forest algorithm provided by the present invention will be described below. The rock grade identification device based on sliding window feature extraction and random forest algorithm described below can be referred to in correspondence with the rock grade identification method based on sliding window feature extraction and random forest algorithm described above.

[0075] Figure 4 This is a schematic diagram of the surrounding rock grade identification device based on sliding window feature extraction and random forest algorithm provided in an embodiment of the present invention, as shown below. Figure 4 As shown, it includes: a preprocessing data acquisition module 41, a sliding window extraction module 42, a dataset construction module 43, a primary model construction module 44, a model optimization module 45, a final model acquisition module 46, and a recognition module 47, wherein:

[0076] The preprocessing data acquisition module 41 is used to acquire the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data.

[0077] The sliding window extraction module 42 is used to extract features from the preprocessed data using a sliding window and to obtain a feature dataset.

[0078] The dataset construction module 43 is used to build training, validation and test sets by taking the feature dataset as input and the surrounding rock grade at the sliding window truncation time as output.

[0079] The primary model building module 44 is used to build a TBM tunnel surrounding rock grade identification model based on the random forest algorithm;

[0080] The model optimization module 45 is used to determine the optimal sliding window length and step size, and to use a search algorithm to find the optimal model hyperparameters to obtain an optimized TBM tunnel surrounding rock grade identification model.

[0081] The final model acquisition module 46 is used to find the optimal weights for each category of the random forest model using a global optimization algorithm, and obtain a weight-adaptive TBM tunnel surrounding rock grade identification model.

[0082] The identification module 47 is used to identify the surrounding rock grade using a weighted adaptive TBM tunnel surrounding rock grade identification model.

[0083] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call logical instructions in the memory 530 to execute a rock mass classification method based on sliding window feature extraction and random forest algorithm, the method including:

[0084] Obtain the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data;

[0085] The preprocessed data is truncated using a sliding window, and features are extracted from the truncated data to obtain a feature dataset.

[0086] Using the feature dataset as input and the surrounding rock grade at the sliding window truncation time as output, a training set, a validation set, and a test set are established.

[0087] Construct a TBM tunnel surrounding rock grade identification model based on the random forest algorithm;

[0088] The optimal sliding window length and step size are determined, and the optimal model hyperparameters are found using a search algorithm to obtain an optimized TBM tunnel surrounding rock grade identification model.

[0089] The optimal weights for each category in the random forest model are found using a global optimization algorithm, resulting in a weight-adaptive TBM tunnel surrounding rock grade identification model.

[0090] A weighted adaptive TBM tunnel surrounding rock grade identification model is used to identify the surrounding rock grade.

[0091] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0092] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the rock grade identification method based on sliding window feature extraction and random forest algorithm provided by the above methods.

[0093] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the rock mass classification identification method based on sliding window feature extraction and random forest algorithm provided by the above methods.

[0094] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0095] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying the grade of surrounding rock based on sliding window feature extraction and random forest algorithm, characterized in that, Includes the following steps: Obtain the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data; The preprocessed data is truncated using a sliding window, and features are extracted from the truncated data to obtain a feature dataset. Using the feature dataset as input and the surrounding rock grade at the sliding window truncation time as output, a training set, a validation set, and a test set are established. Construct a TBM tunnel surrounding rock grade identification model based on the random forest algorithm; The optimal sliding window length and step size are determined, and the optimal model hyperparameters are found using a search algorithm to obtain an optimized TBM tunnel surrounding rock grade identification model. The optimal weights for each category in the random forest model are found using a global optimization algorithm, resulting in a weight-adaptive TBM tunnel surrounding rock grade identification model. A weighted adaptive TBM tunnel surrounding rock grade identification model is used to identify the surrounding rock grade.

2. The method for identifying surrounding rock grade based on sliding window feature extraction and random forest algorithm according to claim 1, characterized in that, The raw data collected by the TBM includes: tunneling parameters and corresponding surrounding rock grades; the tunneling parameters include thrust, cutterhead torque, cutterhead rotation speed, propulsion speed, guide roll angle, guide pitch angle, support shoe pump pressure, support pressure, main conveyor speed, and main conveyor pump pressure. The preprocessing includes: screening tunneling parameters based on experience and random forest algorithms, outlier detection based on isolated forest algorithms, and data normalization based on sensor range.

3. The method for identifying surrounding rock grade based on sliding window feature extraction and random forest algorithm according to claim 1, characterized in that: The method for using a sliding window to truncate preprocessed data and extracting features from the truncated data to obtain a feature dataset is as follows: Set the parameters of the sliding window, use the sliding window to truncate the preprocessed data, and then extract features from the truncated data to obtain a feature dataset; The features extracted from the intercepted data include the mean, standard deviation, maximum value, minimum value, quantiles of each tunneling parameter, and the Pearson correlation coefficient between the tunneling parameters.

4. The method for identifying surrounding rock grade based on sliding window feature extraction and random forest algorithm according to claim 1, characterized in that, The optimal sliding window length and step size are determined as follows: the optimal sliding window length and step size are determined by using the performance of the TBM tunnel surrounding rock grade identification model based on the random forest algorithm on the validation set; The model hyperparameters include: number of weak learners, maximum tree depth, and maximum number of features.

5. The method for identifying surrounding rock grade based on sliding window feature extraction and random forest algorithm according to claim 1, characterized in that, The method for finding the optimal weights for each category of the random forest model using a global optimization algorithm is as follows: using the performance index of the optimized TBM tunnel surrounding rock grade identification model on the validation set as the fitness function, the optimal weights for each category of the random forest model are found using a global optimization algorithm.

6. The method for identifying surrounding rock grade based on sliding window feature extraction and random forest algorithm according to claim 5, characterized in that, The performance metrics are one or a combination of Precision, Recall, F-measure, and Accuracy.

7. A rock mass classification device based on sliding window feature extraction and random forest algorithm, characterized in that, include: The preprocessing data acquisition module is used to acquire the raw data collected by the TBM, preprocess it according to the tunneling ring, and obtain preprocessed data. The sliding window extraction module is used to truncate preprocessed data using a sliding window and extract features from the truncated data to obtain a feature dataset. The dataset construction module is used to build training, validation and test sets by taking the feature dataset as input and the surrounding rock grade at the sliding window truncation time as output. The primary model building module is used to build a TBM tunnel surrounding rock grade identification model based on the random forest algorithm; The model optimization module is used to determine the optimal sliding window length and step size, and uses a search algorithm to find the optimal model hyperparameters to obtain an optimized TBM tunnel surrounding rock grade identification model. The final model acquisition module is used to find the optimal weights for each category of the random forest model using a global optimization algorithm, and obtain a weight-adaptive TBM tunnel surrounding rock grade identification model. The identification module is used to identify the surrounding rock grade using a weighted adaptive TBM tunnel surrounding rock grade identification model.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the rock mass classification method based on sliding window feature extraction and random forest algorithm as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the rock mass classification method based on sliding window feature extraction and random forest algorithm as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the rock mass classification method based on sliding window feature extraction and random forest algorithm as described in any one of claims 1 to 6.

Images