A data-driven multi-impact directional drilling parameter optimization method

Abstract

This invention discloses a data-driven method for optimizing multi-dimensional impact directional drilling parameters, comprising: constructing a database matching drilling parameters with rock formation characteristics, the database being obtained through numerical simulation and drilling tests; performing data preprocessing on the database, including wavelet denoising and standardization; inputting the preprocessed data into a machine learning algorithm to train a mechanical drilling rate prediction model; and optimizing drilling pressure and rotational speed based on an intelligent optimization algorithm with mechanical drilling rate, mechanical energy density, and cost per minute as optimization objectives. This invention shortens the drilling cycle, reduces drilling costs, and improves drilling efficiency.

Description

Technical Field

[0001] This invention belongs to the field of horizontal directional drilling technology, and particularly relates to a data-driven method for optimizing multi-element impact directional drilling parameters. Background Technology

[0002] In recent years, tunnel construction has witnessed numerous major disasters such as water and mud inrushes and collapses, resulting in significant casualties, severe project delays, and substantial economic and property losses. Long-distance directional drilling using directional drilling rigs for remote reinforcement and grouting of disaster sites has become an effective means of refined tunnel exploration and disaster prevention. However, because directional drilling trajectories are mostly located near the surface, where the geological environment is complex and lithology varies, the directional drilling process faces challenges such as slow drilling speeds. In recent years, domestic researchers have proposed a new composite impact rock-breaking method combining axial and torsional impact, and have successively developed axial and torsional composite impact drilling tools. These methods have achieved good application results in the field and are gradually developing into a key technology for rapid drilling.

[0003] Meanwhile, different drilling parameters (drilling pressure, rotation speed, etc.) have different effects on the drilling process. Optimizing drilling parameters can effectively shorten the drilling cycle, reduce drilling costs, and improve drilling efficiency. However, there is currently a lack of databases for matching multi-dimensional impact horizontal directional drilling parameters with post-disaster composite strata characteristics. This patent constructs a drilling parameter-strata attribute database based on rock characteristic difference classification and a large data source of drilling parameters. It uses deep learning methods to establish a drilling rate prediction algorithm model and uses intelligent optimization algorithms to optimize drilling parameters, forming a data-driven multi-dimensional impact horizontal directional drilling parameter optimization method. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a data-driven method for optimizing multi-element impact directional drilling parameters, which shortens the drilling cycle, reduces drilling costs, and improves drilling efficiency.

[0005] To achieve the above objectives, this invention provides a data-driven method for optimizing multi-element impact directional drilling parameters, comprising: A database matching drilling parameters with rock strata characteristics is constructed, which is obtained through numerical simulation and drilling tests; Machine learning algorithms are used to invert formation strength characteristics from drilling parameters, and intelligent optimization algorithms are used to optimize the drilling parameters.

[0006] Optionally, the process of building the database includes: A numerical simulation model for multi-element impact horizontal directional drilling was established, which includes the drill string, wellbore, impactor, PDC drill bit, and rock. Set the boundary conditions for drilling parameters such as drill pressure, rotation speed, axial impact load, axial impact frequency, torsional impact load, and torsional impact frequency; Obtain simulated data on mechanical drilling rate, torque, and rock fragmentation volume.

[0007] Optionally, the database construction process may also include: Establish a drilling test system, including a drilling power system, a rock support structure, and a data acquisition system; The data acquisition system includes a wire-type displacement sensor to measure drilling displacement, a pressure sensor to measure drilling pressure, a dynamic torque sensor to measure rotational speed and torque, and a flow meter to measure flow rate.

[0008] Optionally, the process of inverting formation strength characteristics based on drilling parameters using machine learning algorithms includes: The drilling parameters are preprocessed, including wavelet denoising to remove outliers from the fused data. Correlation analysis was performed on drilling parameters and mechanical drilling rate; Data standardization and normalization are used to unify parameter scales; The model was trained using machine learning algorithms such as random forest, support vector machine, and backpropagation neural network; The algorithm with the smallest error is selected by evaluating MSE, RMSE, and R² values. Optimize model parameters based on rock strength prediction results.

[0009] Optionally, the process of optimizing drilling parameters using intelligent optimization algorithms includes: The first approach is to set drilling parameters with a single objective of maximizing mechanical drilling speed, minimizing mechanical specific energy, or minimizing drilling cost, or a combination thereof to form multiple objectives. The optimal combination of drilling pressure and rotation speed is determined using an intelligent optimization algorithm. The second approach involves constructing a proxy model for the drilling parameters. An intelligent optimization algorithm is used to solve the surrogate model and obtain the optimal combination of drilling pressure and rotation speed.

[0010] Optionally, the data acquisition process of the drilling test system includes: A wire-type displacement sensor is installed at the end of the drill rig's power head to measure displacement. A pressure sensor is installed between the rock box and the rock support base to measure drilling pressure; A dynamic torque sensor is installed between the impact screw and the hollow short section to measure torque and speed. The flow meter is installed in the drill string and pump workshop to measure flow rate.

[0011] Technical advantages of this invention: This invention discloses a data-driven method for optimizing parameters in multi-element impact directional drilling, establishes a numerical simulation model for multi-element impact horizontal directional drilling, and develops a test system for multi-element impact horizontal directional drilling. Based on this, a database matching drilling parameters with rock formations is established. Furthermore, the data-driven optimization method for multi-element impact directional drilling can effectively shorten drilling cycles, reduce drilling costs, and improve drilling efficiency. Attached Figure Description

[0012] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating a data-driven multi-element impact directional drilling parameter optimization method according to an embodiment of the present invention. Figure 2 A schematic diagram of a multi-element impact horizontal directional drilling model including a drill string and an impactor is provided for an embodiment of the present invention. Figure 3 This is a schematic diagram of the drilling test system according to an embodiment of the present invention; Figure 4 This is a schematic diagram showing the mechanical drilling speed, torque, drilling pressure, rotational speed, and flow rate during the actual drilling process of an embodiment of the present invention. Detailed Implementation

[0013] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0014] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0015] like Figure 1 As shown, this embodiment provides a data-driven method for optimizing multi-element impact directional drilling parameters, including: A database matching drilling parameters with rock strata characteristics is constructed, which is obtained through numerical simulation and drilling tests; Based on machine learning algorithms, formation strength characteristics are inverted from drilling parameters, and intelligent optimization algorithms are used to optimize drilling parameters. Develop an intelligent optimization system for drilling parameters to achieve optimized parameter output.

[0016] Furthermore, the process of building the database includes: A numerical simulation model for multi-element impact horizontal directional drilling was established, which includes the drill string, wellbore, impactor, PDC drill bit, and rock. Set the boundary conditions for drilling parameters such as drill pressure, rotation speed, axial impact load, axial impact frequency, torsional impact load, and torsional impact frequency; Obtain simulated data on mechanical drilling rate, torque, and rock fragmentation volume.

[0017] Furthermore, the process of building a database also includes: Establish a drilling test system, including a drilling power system, a rock support structure, and a data acquisition system; The data acquisition system includes a wire-type displacement sensor to measure drilling displacement, a pressure sensor to measure drilling pressure, a dynamic torque sensor to measure rotational speed and torque, and a flow meter to measure flow rate.

[0018] Furthermore, the process of inverting formation strength characteristics based on drilling parameters using machine learning algorithms includes: The drilling parameters are preprocessed, including wavelet denoising to remove outliers from the fused data. Correlation analysis was performed on drilling parameters and mechanical drilling rate; Data standardization and normalization are used to unify parameter scales; The model was trained using machine learning algorithms such as random forest, support vector machine, and backpropagation neural network; The algorithm with the smallest error is selected by evaluating MSE, RMSE, and R² values. Optimize model parameters based on rock strength prediction results.

[0019] Furthermore, the process of optimizing drilling parameters using intelligent optimization algorithms includes: The first approach is to set drilling parameters with a single objective of maximizing mechanical drilling speed, minimizing mechanical specific energy, or minimizing drilling cost, or a combination thereof to form multiple objectives. The optimal combination of drilling pressure and rotation speed is determined using an intelligent optimization algorithm. The second approach involves constructing a proxy model for the drilling parameters. An intelligent optimization algorithm is used to solve the surrogate model and obtain the optimal combination of drilling pressure and rotation speed.

[0020] Furthermore, the data acquisition process of the drilling test system includes: A wire-type displacement sensor is installed at the end of the drill rig's power head to measure displacement. A pressure sensor is installed between the rock box and the rock support base to measure drilling pressure; A dynamic torque sensor is installed between the impact screw and the hollow short section to measure torque and speed. The flow meter is installed in the drill string and pump workshop to measure flow rate.

[0021] Specifically, the implementation process of this embodiment includes: This embodiment obtains a data source that matches drilling parameters to rock formations through a combination of numerical simulation and experimentation. Based on artificial intelligence, it establishes a mechanical drilling rate prediction model and a drilling parameter optimization model, forming a data-driven multi-element impact directional drilling parameter optimization method. The technical route is as follows: Figure 1 As shown.

[0022] Drilling parameters were matched with rock formation data obtained through numerical simulation and drilling tests. Numerical simulation of drilling: To obtain a more realistic picture of the motion and force during multi-element impact horizontal directional drilling, a multi-element impact horizontal directional drilling model including the drill string and impactor is established based on the actual drilling process, such as... Figure 2 As shown. The model includes: drill string, wellbore, impactor, PDC drill bit, and rock, wherein: ① The drill string is discretized using beam elements, the wellbore is set as an analytical steel body, the impactor and rock are discretized using C3D8R mesh, and the PDC drill bit is discretized using C3D10M mesh. To obtain more accurate results, the rock mesh is refined in the contact area between the drill bit and the rock. ② The drill string, impactor, and wellbore use a general contact setting, and the drill bit and rock use surface-to-surface contact with the contact property set to penalized contact. ③ Hinge elements are used to connect the impactor to the upper drill string to simulate the relative rotation of the screw motor and the drill string. ④ The model boundary conditions include: drill pressure, rotational speed, axial impact load, axial impact frequency, torsional impact load, and torsional impact frequency. ⑤ Based on the parameter post-processing module, the numerical simulation results of multivariate impact horizontal directional drilling are obtained: mechanical drilling speed, torque, rock fragmentation volume, etc.

[0023] Drilling test: The drilling test system as a whole includes: drilling power system, rock support structure, and data acquisition system, such as... Figure 3As shown. The drilling power system includes a mud pump truck, drilling rig, drill string, hollow sub, impact screw, and PDC drill bit. The rock support structure includes a rock box, rock support base, and guide rails. The data acquisition system includes a wire-type displacement sensor, pressure sensor, dynamic torque sensor, and flow meter. The wire-type displacement sensor is installed at the power head end of the drilling rig to measure the drilling displacement; the pressure sensor is installed between the rock box and the rock support base to measure the drilling pressure in real time; the dynamic torque sensor is installed between the impact screw and the hollow sub to measure the rotational speed of the impact screw and the counter-torque generated by the PDC drill bit when breaking the rock in real time; the flow meter is installed between the drill string and the pump truck to measure the water flow rate during drilling. Drilling parameters are collected, displayed in real time, and saved using a data acquisition system. During the test, the drilling rig provides the drilling pressure, and the mud pump truck provides power to the impact screw through drilling fluid, causing the screw drill string to drive the drill bit to rotate. Simultaneously, drilling fluid flows out from the drill bit nozzle, effectively cleaning the bottom of the well. The impact screw motor converts the pressure energy of the liquid into mechanical energy to drive the drill bit to rotate. Simultaneously, its internal hammer and spring provide impact load to the drill bit, achieving axial impact functionality. A wire-type displacement sensor, pressure sensor, dynamic torque sensor, and flow meter measure and collect data on displacement, drilling pressure, rotational speed, torque, and flow rate during the test. The actual drilling process's mechanical drilling speed, torque, drilling pressure, rotational speed, and flow rate are recorded and saved by a data acquisition instrument. Figure 4 As shown.

[0024] Data obtained from numerical simulations and drilling tests are integrated into a single data table, ensuring structural consistency, to establish a database that matches drilling parameters with rock strata.

[0025] Data obtained from drilling simulations and tests underwent preprocessing: wavelet denoising, data standardization, and normalization were performed to remove outliers, and correlation analysis was conducted on drilling parameters and mechanical drilling rate. The preprocessed data was then used to train a model using machine learning algorithms such as random forest, support vector machine, and backpropagation neural network. By evaluating the MSE, RMSE, and R² values ​​of different algorithms, the algorithm with the smallest model error was selected for rock strength inversion.

[0026] Two methods are proposed to optimize drilling parameters using intelligent optimization algorithms. The first method uses mechanical drilling rate, mechanical energy specificity, and cost per minute as optimization objectives. It employs intelligent optimization algorithms, such as particle swarm optimization, to optimize multiple parameters including drill pressure and rotational speed, aiming to achieve the optimal drilling speed. The second method constructs a Kriging surrogate model using drilling parameters from a database. Intelligent optimization algorithms, such as genetic algorithms, are then used to solve the surrogate model to obtain the optimal drilling parameters under the maximum mechanical drilling rate.

[0027] This invention discloses a data-driven method for optimizing parameters in multi-element impact directional drilling. A numerical simulation model for multi-element impact horizontal directional drilling is established, and a test system for multi-element impact horizontal directional drilling is developed. Based on this, a database matching drilling parameters with rock formations is established. Furthermore, a data-driven optimization method for multi-element impact directional drilling is developed, which can effectively shorten drilling cycles, reduce drilling costs, and improve drilling efficiency.

[0028] This invention establishes a multi-element impact horizontal directional drilling test system and collects drilling parameters during the drilling process; establishes a database that matches drilling parameters with rock formations; trains the drilling parameter-rock formation data using machine learning algorithms to obtain a mechanical drilling rate prediction model; and optimizes drilling pressure and rotation speed using intelligent optimization algorithms with mechanical drilling rate, mechanical specific energy, and cost per minute as optimization objectives to achieve the optimal drilling speed.

[0029] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data-driven method for optimizing parameters in multi-element impact directional drilling, characterized in that, include: A database matching drilling parameters with rock strata characteristics is constructed, which is obtained through numerical simulation and drilling tests; Machine learning algorithms are used to invert formation strength characteristics from drilling parameters, and intelligent optimization algorithms are used to optimize the drilling parameters.

2. The data-driven multi-element impact directional drilling parameter optimization method as described in claim 1, characterized in that, The process of building a database includes: A numerical simulation model for multi-element impact horizontal directional drilling was established, which includes the drill string, wellbore, impactor, PDC drill bit, and rock. Set boundary conditions for drilling parameters such as drill pressure, rotation speed, axial impact load, axial impact frequency, torsional impact load, and torsional impact frequency; Obtain simulated data on mechanical drilling rate, torque, and rock fragmentation volume.

3. The data-driven multi-element impact directional drilling parameter optimization method as described in claim 2, characterized in that, The process of building a database also includes: Establish a drilling test system, including a drilling power system, a rock support structure, and a data acquisition system; The data acquisition system includes a wire-type displacement sensor to measure drilling displacement, a pressure sensor to measure drilling pressure, a dynamic torque sensor to measure rotational speed and torque, and a flow meter to measure flow rate.

4. The data-driven multi-element impact directional drilling parameter optimization method as described in claim 1, characterized in that, The process of inverting formation strength characteristics based on drilling parameters using machine learning algorithms includes: The drilling parameters are preprocessed, including wavelet denoising to remove outliers from the fused data. Correlation analysis was performed on drilling parameters and mechanical drilling rate; Data standardization and normalization are used to unify parameter scales; The model was trained using machine learning algorithms such as random forest, support vector machine, and backpropagation neural network; The algorithm with the smallest error is selected by evaluating MSE, RMSE, and R² values. Optimize model parameters based on rock strength prediction results.

5. The data-driven multi-element impact directional drilling parameter optimization method as described in claim 1, characterized in that, The process of optimizing drilling parameters using intelligent optimization algorithms includes: The first approach is to set drilling parameters with a single objective of maximizing mechanical drilling speed, minimizing mechanical specific energy, or minimizing drilling cost, or a combination thereof to form multiple objectives. The optimal combination of drilling pressure and rotation speed is determined using an intelligent optimization algorithm. The second approach involves constructing a proxy model for the drilling parameters. An intelligent optimization algorithm is used to solve the surrogate model and obtain the optimal combination of drilling pressure and rotation speed.

6. The data-driven multi-element impact directional drilling parameter optimization method as described in claim 3, characterized in that, The data acquisition process of the drilling test system includes: A wire-type displacement sensor is installed at the end of the drill rig's power head to measure displacement. A pressure sensor is installed between the rock box and the rock support base to measure drilling pressure; A dynamic torque sensor is installed between the impact screw and the hollow short section to measure torque and speed; The flow meter is installed in the drill string and pump workshop to measure flow rate.

Images