Rock grouting and stone body sensing system and method based on advanced drilling while drilling testing

Abstract

The application provides a surrounding rock grouting stone body sensing system and method based on advanced drilling while drilling testing, obtains real-time tunneling parameters and geological parameter data in the drilling process; processes the data, determines the weights of the tunneling parameters and the geological parameters in the final analysis model by using a function model, so as to construct the final analysis model; uses a particle swarm optimization algorithm to perform hyperparameter optimization on five submodels with different viewpoints, trains the model based on the known tunneling parameters and the geological parameters, inputs the tunneling parameters in the trained model, predicts unknown geological parameters, and senses the geological and mechanical characteristics of the grouting stone body; and fuses the characteristic results sensed by the five submodels after optimization by using D-S evidence theory, so as to obtain the final characteristics of the surrounding rock grouting stone body. The application can significantly improve the accuracy of surrounding rock grouting stone body sensing.

Description

Technical Field

[0001] This invention belongs to the field of directional drilling pre-grouting technology, specifically relating to a system and method for sensing grouting rock bodies in surrounding rock based on pre-drilling and drilling-while-drilling testing. Background Technology

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

[0003] In recent years, with the continuous advancement of infrastructure construction, horizontal directional drilling technology, which is widely used in municipal engineering, oil and gas pipeline laying and other fields, has become one of the research hotspots.

[0004] Existing research in long-distance directional drilling for lithological prediction focuses primarily on inverting uniaxial compressive strength from drilling parameters. However, research on the relationship between other rock mechanical parameters and drilling parameters is limited, and these studies suffer from drawbacks such as requiring the measurement of too many parameters, inherent subjectivity, and low accuracy. With the development of artificial intelligence, machine learning algorithms such as neural networks, support vector machines, and decision trees are increasingly being applied. However, as identification methods, these often rely on single models and suffer from low accuracy. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a system and method for perceiving grouting rock bodies based on advanced drilling and testing. This invention employs a function method to establish a rock body parameter perception model based on geological and tunneling parameters, uses deep learning mapping modeling to perceive the characteristics of grouting rock bodies, and fuses the identification results of multiple models through DS evidence theory to improve accuracy.

[0006] According to some embodiments, the present invention adopts the following technical solution: A rock mass sensing system based on advanced drilling and testing includes: The data acquisition module is used to acquire real-time tunneling and geological parameter data during the drilling process; The preprocessing module includes a data correction module and a data transformation module. The data correction module is used to remove outliers in the acquired data using a clustering algorithm, and the data transformation module is used to standardize the tunneling parameter data using the Z-Score method. The weight allocation module uses a function model to determine the weights of tunneling parameters and geological parameters in the final analysis model in order to construct the final analysis model. The feature screening module includes an extreme learning machine sub-model, a backpropagation artificial neural network sub-model, a radial basis function artificial neural network sub-model, a random forest sub-model, and a K-nearest neighbor algorithm sub-model. All five sub-models use the particle swarm optimization algorithm to optimize hyperparameters. The models are trained based on known tunneling parameters and geological parameters. After training, the tunneling parameters are input into the model to predict unknown geological parameters and perceive the geological and mechanical characteristics of the grouting stone body. The multimodal fusion module fuses the feature results perceived by the five sub-models after optimization using DS evidence theory to obtain the final features of the surrounding rock grouting stone body.

[0007] As an alternative implementation, the geological parameters include uniaxial compressive strength, initial ground stress state, groundwater aquifer state, and structural surface state; the tunneling parameters include thrust, penetration depth, torque, and tunneling speed.

[0008] As an alternative implementation, the data correction module uses the DBSCAN clustering algorithm to remove outliers from the acquired data.

[0009] As an alternative implementation, the data conversion module adopts the Z-Score normalization method to uniformly scale the data of each dimension to the range of (-1, 1).

[0010] As an alternative implementation, the weight allocation module fits the known tunneling and geological parameters to obtain functional forms relating the four geological parameters to the tunneling parameters. These four functional forms are then merged to establish the final functional model. This functional model is embedded into the recognition sub-model to enhance the perception of the geological and mechanical characteristics of the grouting rock mass.

[0011] As an alternative implementation, the five sub-models are respectively input with thrust and penetration depth, torque, and tunneling speed to obtain output features including the thickness of the rock mass, inclination angle, compressive strength, and permeability coefficient.

[0012] As an alternative implementation, when the five sub-models use the particle swarm optimization algorithm to optimize hyperparameters, they are trained using historical data, and after verification, the trained models are used to identify and predict the surrounding rock level.

[0013] As an alternative implementation, when the five sub-models use the particle swarm optimization algorithm to optimize hyperparameters, the particles iteratively update their position and velocity by tracking their individual historical best and the global best of the group. The core parameters include the learning factor and the inertia weight.

[0014] The working method of the above-mentioned surrounding rock grouting stone body sensing system includes the following steps: The data acquisition module acquires real-time tunneling and geological parameter data during the drilling process; The data correction module uses a clustering algorithm to remove outliers from the acquired data, while the data transformation module uses the Z-Score method to standardize the tunneling parameter data. The weighting module uses a function model to determine the weights of tunneling parameters and geological parameters in the final analysis model. Using the Extreme Learning Machine sub-model, the Backpropagation Artificial Neural Network sub-model, the Radial Basis Function Artificial Neural Network sub-model, the Random Forest sub-model, and the K-Nearest Neighbors algorithm sub-model, all five sub-models employ the Particle Swarm Optimization algorithm for hyperparameter optimization, and based on the tunneling parameters, the mechanical characteristics of the grouting stone body are perceived. The multimodal fusion module integrates the feature results perceived by the five sub-models after optimization using DS evidence theory to obtain the final features of the surrounding rock grouting stone body.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention overcomes the shortcomings of the method in terms of stone body perception, such as too many parameters and strong subjectivity, and does not have the black box problem in artificial intelligence methods, making it easy for workers to understand and apply.

[0016] (2) This invention applies the DBSCAN clustering algorithm to outlier processing and uses the more advanced PSO optimization algorithm to optimize the selection of hyperparameters for the sub-model.

[0017] (3) This invention improves the traditional DS evidence theory in terms of surrounding rock level identification by adopting a multi-model fusion identification and prediction model. This system has good robustness and overcomes the shortcomings of low accuracy and one-sidedness of a single model.

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0020] Figure 1 This is a framework diagram of a rock mass sensing system based on advanced drilling and testing. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0023] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0024] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0025] Example 1 A rock mass sensing system based on advanced drilling and testing, such as Figure 1 As shown, it includes a data acquisition module, a preprocessing module, a weight allocation module, a feature screening module, and a multimodal fusion module.

[0026] in: The data acquisition module is used to acquire tunneling parameters such as rotational speed, torque, and pressure during real-time monitoring of the drilling process.

[0027] Geological parameters include uniaxial compressive strength, groundwater aquifer, initial geostress, and structural surface condition; tunneling parameters include thrust, penetration depth, torque, and tunneling speed.

[0028] The preprocessing module includes a data correction module and a data transformation module. The data correction module uses the DBSCAN clustering algorithm to remove outliers from the acquired data. Its principle is to find core objects (data points whose neighborhood contains at least...). n The data points (referred to as core objects) are connected to form different dense regions, and isolated data points not belonging to any cluster are considered outliers. The data transformation module uses the Z-Score method to uniformly scale all data to the range (-1, 1) to eliminate the influence of different dimensions between data points. The formula is:

[0029] in, The mean, The standard deviation is denoted as .

[0030] The weight allocation module constructs a function model for perceiving stone body parameters during drilling testing, as follows: First, the fitness value Y is defined:

[0031] X1, X2, X3, and X4 refer to thrust, penetration depth, torque, and tunneling speed, respectively. These three parameters combined can effectively represent the adaptability of directional drilling during the drilling-while-drilling testing process. These refer to the weights of the four parameters, which are determined using principal component analysis.

[0032] Principal component analysis (PCA) can be used to determine weights using SPSS software. The PCA function in SPSS generates tables of variance explained rates and loadings, from which weights can be calculated. First, calculate the linear combination coefficient matrix, which is the load coefficient divided by the corresponding square root.

[0033] Second, calculate the comprehensive score coefficient, which is calculated as the cumulative (linear combination coefficient * variance explained rate) / cumulative variance explained rate.

[0034] Third, determine the weights by summing and normalizing the comprehensive score coefficients to obtain the weight values ​​of each indicator.

[0035] The weight allocation module uses four parameters—Y1, Y2, Y3, and Y4—based on the fitness value Y to fit the uniaxial compressive strength, groundwater aquifer state, initial geostress state, and structural surface state, respectively, resulting in four functional forms. These four functional forms are then merged to establish the final functional model. The functional model is then embedded into the subsequent recognition sub-model.

[0036] The feature screening module includes five basic models: ELM, BP, RBF, RF, and KNN.

[0037] Extreme Learning Machine (ELM) is a single-hidden-layer feedforward neural network architecture whose core feature lies in its random initialization mechanism for hidden layer parameters. This network consists of an input layer, a non-linear hidden layer, and an output layer. Its breakthrough design lies in the fact that the weight vectors and biases of the hidden layer neurons are generated randomly, eliminating the need for iterative optimization via backpropagation. The output layer weights are directly calculated analytically (e.g., least squares). This non-iterative training method significantly improves learning speed, but model performance is sensitive to the distribution of initial random parameters; inappropriate initialization can lead to differences in prediction stability.

[0038] Backpropagation (BP) networks are a fundamental architecture in deep learning, employing a multilayer perceptron structure comprising an input layer, multiple hidden layers, and an output layer. Its core learning mechanism is based on the chain rule, calculating the predicted output through forward propagation, propagating the error gradient through backpropagation, and using gradient descent to adjust the connection weights and biases between neurons layer by layer. Despite requiring the optimization of a large number of parameters, BP networks and their variants (such as CNNs and RNNs) have become standard tools in fields such as image recognition and speech processing due to their powerful nonlinear fitting capabilities, driving the deep learning revolution.

[0039] Random Forest (RF) is a representative algorithm in the field of ensemble learning, which improves the generalization ability of a model by constructing a combination of multiple decision trees. This algorithm introduces double randomness: at the sample level, multiple training subsets are generated through bootstrap sampling; at the feature level, a subset of features is randomly selected for evaluation when each node splits. The final prediction results are ensembled through either voting (classification) or averaging (regression). Key hyperparameters include the number of decision trees (n_estimators) and the number of features per node split (max_features), which directly affect model complexity and resistance to overfitting.

[0040] The Radial Basis Function (RBF) neural network is a three-layer feedforward network whose core design lies in using radial basis functions (such as Gaussian functions) as activation functions in the hidden layers. The input signal undergoes a nonlinear mapping from low-dimensional to high-dimensional space through the hidden layer nodes, while the output layer performs a linearly weighted combination of the high-dimensional features. This network possesses global approximation properties and can theoretically approximate continuous functions with arbitrary precision. Key parameters include the center of the radial basis functions (usually determined through clustering), the width parameter (σ value), and the output layer weights. These parameters collectively determine the network's function approximation ability and classification boundary complexity.

[0041] K-Nearest Neighbors (KNN) is an instance-based supervised learning algorithm suitable for classification and regression tasks. Its core idea is "grouping by type": for a new sample, calculate its distance (e.g., Euclidean distance) to all samples in the training set, select the K nearest neighbors, and make a prediction using majority voting (classification) or the average (regression). The value of K needs to be manually set; smaller values ​​are sensitive to noise, while larger values ​​may blur class boundaries. The algorithm does not require explicit training, but it has high computational cost during prediction, making it suitable for small-scale data. It is also sensitive to feature scale and requires normalization. Its advantages lie in its simplicity and intuitiveness, while its disadvantages include distance becoming ineffective and low efficiency in high-dimensional data.

[0042] The hyperparameter settings of the five basic models mentioned above have a significant impact on performance, thus requiring optimization algorithms to find the optimal hyperparameters. PSO (Particle Swarm Optimization) is a swarm intelligence optimization method inspired by bird flocks. Particles iteratively update their position and velocity by tracking their individual historical optima and the global optima of the group. The core parameters include learning factors (c1 / c2) and inertia weights (w). It has both global search capabilities and fast convergence, making it suitable for high-dimensional continuous problems such as function optimization and parameter tuning. However, in high-dimensional scenarios, it may get stuck in local optima and is sensitive to parameters.

[0043] Therefore, five initial recognition sub-models were obtained: PSO-ELM, PSO-BP, PSO-RBF, PSO-RF, and PSO-KNN.

[0044] The feature screening module, based on known tunneling and geological parameters, trains five sub-models. Thrust, penetration depth, torque, and tunneling speed are input into each of the five trained sub-models, and the output indicators are the predicted grouting rock mass feature parameters (rock mass thickness, inclination angle, compressive strength, and permeability coefficient). A global optimization algorithm is used to find the optimal hyperparameters. 80% of the historical data is used for training, and 20% is used for verification. The trained model enables the identification and prediction of the surrounding rock level.

[0045] The traditional DS evidence theory method in the multimodal fusion module has the following characteristics: The DS evidence theory mainly consists of an identification framework, a basic probability distribution function, and rules for combining evidence.

[0046] First, identify the frame Θ={ θ 1, θ 2,..., θ n}, θ j It is a subset of the recognition frame Θ. It is its power set, if the mapping The following conditions must be met:

[0047] Where Ø is the empty set, m yes θ The basic probability function (BPA). m(A i ) It is a subset of certain evidence within the identification framework. A i The probability of support, if m(A i ) >0 means it is considered A It is a focal element of BPA.

[0048] The confidence function for BPA is:

[0049] Bel(A) This represents the sum of the probabilities of all subsets of an event and the probability that supports the event.

[0050] The likelihood function of BPA is:

[0051] Pl(A) This represents the sum of probabilities that the intersection with the event is not empty.

[0052] Pl(A)>Bel(A) , representing the upper and lower limits of support for the event, respectively.

[0053] For two independent pieces of evidence within the same forensic framework, their BPAs are respectively m 1 and m 2. The focus elements are respectively A 1, ... A k and B 1, ... B k The combination rules are as follows:

[0054] Preferably, the multimodal fusion module defines the thickness, tilt angle, compressive strength, and permeability coefficient of the grouting rock body as Θ = (I, II, III, and IV), and selects PSO-ELM, PSO-BP, PSO-RBF, PSO-RF, and PSO-KNN models as fusion evidence. In Θ, the support probability set for each piece of evidence for the rock characteristic parameters is a basic probability assignment function (BPA). m i (i = 1, 2, ..., 5). The prediction of each feature parameter sample yields the DS evidence theory probability matrix, as shown in Table 1.

[0055] Table 1. Probability matrix of DS evidence theory

[0056] Then, the weights between the evidence are calculated according to the formula, and the recognition results are fused according to the fusion rule formula.

[0057] Weight calculation formula: Evidence distance (Jousselme distance): First construct the compatibility matrix Its elements:

[0058] Any two pieces of evidence Distance to Jousselme:

[0059] Similarity:

[0060] Support:

[0061] Normalized to weights:

[0062] The working method of the above system includes the following steps: A. Parameters required for advanced drilling and testing while drilling; B. Remove abnormal data using the data correction module; C. Compress all data to (-1,1) using the data conversion module; D. Input the geological parameters and tunneling parameters into the weight allocation model to obtain the final function model; E. During the advanced drilling and testing phase, the tunneling parameters are input into the feature screening module, and the grouting stone body characteristics are initially obtained through five sub-models. F. Based on the output of the feature screening module, the recognition results of the five sub-models are fused through the multimodal fusion module to obtain a prediction result with high accuracy.

[0063] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of one or more computer-usable storage media (including, but not limited to, disk storage, etc.) containing computer-usable program code. CD - ROM It takes the form of a computer program product implemented on (such as optical memory, etc.).

[0064] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0065] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0066] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0067] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A rock mass sensing system for grouting based on advanced drilling and testing, characterized in that, include: The data acquisition module is used to acquire real-time tunneling parameters and geological parameters during the drilling process. The geological parameters include uniaxial compressive strength, initial ground stress state, groundwater aquifer state, and structural surface state. The tunneling parameters include thrust, penetration depth, torque, and tunneling speed. The preprocessing module includes a data correction module and a data transformation module. The data correction module is used to remove outliers in the acquired data using a clustering algorithm, and the data transformation module is used to standardize the tunneling parameter data using the Z-Score method. The weight allocation module uses a function model to determine the weights of tunneling parameters and geological parameters in the final analysis model in order to construct the final analysis model. The feature screening module includes an extreme learning machine sub-model, a backpropagation artificial neural network sub-model, a radial basis function artificial neural network sub-model, a random forest sub-model, and a K-nearest neighbor algorithm sub-model. All five sub-models use particle swarm optimization algorithm for hyperparameter optimization. The models are trained based on known tunneling parameters and geological parameters. The tunneling parameters are input into the trained models to predict unknown geological parameters and perceive the geological and mechanical characteristics of the grouting stone body. The five sub-models are respectively input with thrust and penetration, torque, and tunneling speed to obtain output features including stone body thickness, inclination angle, compressive strength, and permeability coefficient. The multimodal fusion module fuses the feature results perceived by the five sub-models after optimization using DS evidence theory to obtain the final features of the surrounding rock grouting stone body.

2. The rock mass sensing system based on advanced drilling and testing as described in claim 1, characterized in that, The data correction module uses the DBSCAN clustering algorithm to remove outliers from the acquired data.

3. The rock mass sensing system based on advanced drilling and testing as described in claim 1, characterized in that, The data transformation module uses the Z-Score normalization method to uniformly scale the data of each dimension to the range of (-1, 1).

4. The rock mass sensing system based on advanced drilling and testing as described in claim 1, characterized in that, The weight allocation module fits the known tunneling parameters and geological parameters to obtain functional forms of the relationship between the four geological parameters and the tunneling parameters. The four functional forms are then merged to establish the final functional model, which is then embedded into the recognition sub-model to enhance the perception of the geological and mechanical characteristics of the grouting stone body.

5. The rock mass sensing system based on advanced drilling and testing as described in claim 1, characterized in that, When the five sub-models use the particle swarm optimization algorithm to optimize hyperparameters, they are trained using historical data. After verification, the trained models are used to identify and predict the surrounding rock level.

6. The surrounding rock grouting stone body sensing system based on advanced drilling and testing as described in claim 1, characterized in that, When the five sub-models use the particle swarm optimization algorithm to optimize hyperparameters, the particles iteratively update their position and velocity by tracking their individual historical best and the global best of the group. The core parameters include the learning factor and the inertia weight.

7. A method for operating the surrounding rock grouting stone body sensing system based on any one of claims 1-6, characterized in that, Includes the following steps: The data acquisition module acquires real-time tunneling and geological parameter data during the drilling process; The data correction module uses a clustering algorithm to remove outliers from the acquired data, while the data transformation module uses the Z-Score method to standardize the tunneling parameter data. The weighting module uses a function model to determine the weights of tunneling parameters and geological parameters in the final analysis model. Using the Extreme Learning Machine sub-model, the Backpropagation Artificial Neural Network sub-model, the Radial Basis Function Artificial Neural Network sub-model, the Random Forest sub-model, and the K-Nearest Neighbors algorithm sub-model, all five sub-models employ the Particle Swarm Optimization algorithm for hyperparameter optimization, and based on the tunneling parameters, the mechanical characteristics of the grouting stone body are perceived. The multimodal fusion module integrates the feature results perceived by the five sub-models after optimization using DS evidence theory to obtain the final features of the surrounding rock grouting stone body.

Images