Intelligent control method and device for remote operation of magnetic steering tool

Abstract

The invention belongs to the technical field of underground detection, and particularly relates to an intelligent control method and device for remote operation of a magnetic steering tool. The method comprises the following steps: carrying out analog simulation before magnetic steering operation before well drilling, and compiling an initial scheme of a magnetic steering tool according to an analog simulation result; standard well operation is carried out based on the initial scheme, and in-drilling data is obtained; then, data mining is conducted by combining calibration data, drilling data and a machine learning algorithm, an optimal decision-making model of the magnetic steering tool is determined, and an intelligent decision-making scheme generated by the optimal decision-making model is issued to the magnetic steering tool for actual operation; the initial scheme of the magnetic steering tool and the intelligent decision model are evaluated according to the actual operation data condition, and the initial scheme of the magnetic steering tool is optimized according to the evaluation result. The method is combined with a machine learning algorithm to realize intelligent decision making of the magnetic steering tool, a traditional empirical formula mode is replaced, the positioning error is reduced, and the positioning accuracy is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of downhole detection, and particularly relates to a magnetic guide tool remote operation intelligent control method and device. BACKGROUND

[0002] A complex structure well is an advanced well type for efficient development of oil and gas fields, and needs to use magnetic guide drilling technology. However, the current magnetic guide tool still cannot meet the demand of more complex and more heavy workload, the underlying data of the existing single machine version software system cannot be communicated, engineers need to stay on site for tracking, the laboratory calibration results are not shared, there are problems such as lack of core personnel, poor information, and the data cannot fully play its role, and iteration and upgrading are imminent.

[0003] From the operation process of the magnetic guide tool, before drilling, the drilling scheme is prepared by relying on manual, in terms of efficiency, there is a gap with the intelligent model based on big data; in the drilling process, the trajectory is determined by relying on the experience formula, in terms of accuracy, there is also a gap with the intelligent decision method of data mining; in addition, there is still a lack of post-drilling evaluation system and method. Although some patents use simulation technology to prepare the scheme before drilling, for example, the patent with the publication number CN115628009A proposes an intelligent control method for well trajectory based on reinforcement learning, and the patent with the publication number CN108694258A proposes a drilling downhole virtual simulation method and system for construction scheme pre-play optimization. However, the above methods are limited to simulation before drilling, although they can roughly predict the operation scheme during the operation process to some extent, and can improve the operation efficiency, but they do not learn and simulate the actual drilling scheme, and the obtained scheme cannot actually guide the operation.

[0004] Therefore, the magnetic guide tool still faces problems such as non-uniform data standards and poor information sharing, cannot realize the fusion with the machine learning algorithm, and the intelligent development of the remote decision system is slow, and still relies on the on-site engineers for implementation and guidance. Therefore, how to realize the remote integrated processing of the magnetic guide tool data, timely and efficiently give an intelligent decision scheme, and form a perfect operation system and method is a problem to be solved. SUMMARY

[0005] In view of the above problems, the application provides a magnetic guide tool remote operation intelligent control method and system to realize the remote integrated processing of the magnetic guide tool data, timely and efficiently give a remote intelligent decision scheme, and form a perfect operation system and method. The specific technical scheme is as follows:

[0006] In the first aspect, the application provides a magnetic guide tool remote operation intelligent control method, including the following steps:

[0007] Simulate the magnetic guidance operation before drilling, and prepare an initial scheme of the magnetic guidance tool according to the simulation result;

[0008] Perform standard well operation based on the initial scheme of the magnetic guidance tool and obtain drilling data; the drilling data includes primary data and secondary data;

[0009] Perform data mining by combining the calibration data, the drilling data and a machine learning algorithm, and determine an optimal decision model of the magnetic guidance tool, and issue an intelligent decision scheme generated by the optimal decision model to the magnetic guidance tool for actual operation;

[0010] Evaluate the execution process of the initial scheme and the intelligent decision model of the magnetic guidance tool according to the actual operation data, and optimize the next operation scheme.

[0011] Further, the simulation of the magnetic guidance operation before drilling and the preparation of the initial scheme of the magnetic guidance tool according to the simulation result include the following steps:

[0012] Input first data of a target drilling area into a pre-drilling simulation system, the first data including formation parameters, tool parameters and boundary conditions;

[0013] Calculate a theoretical trajectory path of the magnetic guidance tool under given conditions according to the first data, and project the theoretical trajectory path into a simulation model;

[0014] Simulate the electromagnetic field distribution of the magnetic guidance measurement under different conditions by using the simulation model to obtain a theoretical value of the magnetic guidance electromagnetic field;

[0015] Compare the theoretical value with an actual value of the magnetic guidance electromagnetic field obtained by measurement, and calculate the error between the theoretical value and the actual value;

[0016] Adjust the first data by using a computer trial calculation method until the error between the theoretical value and the actual value is within a preset range, output a simulation result, and prepare an initial scheme of the magnetic guidance tool, the initial scheme including key information such as a measurement trajectory and measurement points.

[0017] Further, the formation parameters include the electrical conductivity, magnetic permeability and thickness of the formation;

[0018] The tool parameters include the size, shape and magnet strength of the magnetic guidance tool;

[0019] The boundary conditions include ground conditions, surrounding environment and initial conditions of the electromagnetic field.

[0020] Further, the initial scheme of the magnetic guidance tool includes tool selection, drilling parameter setting and drilling strategy.

[0021] Furthermore, the initial scheme based on the magnetic steering tool for performing standard well operations and acquiring drilling data includes the following steps:

[0022] The well locations of standard wells are pre-defined in the target area, including the location relationship between the distributed test wells and the target wells;

[0023] The relative distance from each test well to the target well, calculated using MWD and geometric scanning methods, is used as calibration data.

[0024] Based on the initial scheme of the magnetic steering tool, the operation is carried out to obtain drilling operation data of the magnetic steering tool at different depths in the test well, and the drilling operation data is processed to obtain primary data;

[0025] Based on the primary data, the relative distance from each test well to the target well at different depths is predicted using traditional empirical formulas, and the predicted values ​​are used as secondary data.

[0026] Further, processing the drilling operation data to obtain primary data includes the following steps:

[0027] Configure test parameters, and use these parameters to eliminate erroneous or irrelevant data in the drilling operation data to obtain primary data; after filtering and normalizing the primary data, secondary data is obtained.

[0028] The completeness and accuracy of the secondary data are checked, and primary data is obtained after confirming that there are no errors.

[0029] Furthermore, the drilling operation data includes magnetic flux, acceleration, probe attitude, magnetic field amplitude, and relative orientation.

[0030] Furthermore, the traditional empirical formula is as follows:

[0031]

[0032] Where r represents the predicted relative distance between the test well and the target well at a preset well depth, μ0 represents the vacuum permeability, I is the magnitude of the accumulated current in the standard well, and H... x H y H z These represent the amplitude components of the magnetic field signal along the X, Y, and Z axes at a preset depth in the test well. This represents the average error of the traditional empirical formula.

[0033] Furthermore, the process of using calibration data, drilling data, and machine learning algorithms to perform data mining and determine the optimal decision model for the magnetic guidance tool includes the following steps:

[0034] Combining drill data, calibration data, and machine learning algorithms to perform data mining and self-learning of decision models;

[0035] The calibration data is compared with the predicted well spacing data output by the decision model, and the optimal decision model is determined based on the comparison results.

[0036] Furthermore, the process of combining drill data, calibration data, and machine learning algorithms for data mining and self-learning of decision models includes the following steps:

[0037] The original dataset is constructed using the primary data, secondary data, and calibration data;

[0038] Feature selection is performed on the primary and secondary data in the original dataset to obtain feature quantity data;

[0039] A dataset is constructed by dividing feature data and calibration data, and the dataset is further divided into a training set and a test set; the training set includes feature data and calibration data; the test set includes feature data.

[0040] Multiple machine learning algorithms are selected to define the decision model for the magnetic guidance tool, and the decision model is continuously optimized by changing the model parameters;

[0041] The decision model is self-learned using the training set to obtain the learned decision model.

[0042] Furthermore, feature selection on the original dataset includes:

[0043] The types of feature quantities for all primary and secondary data in the original dataset are statistically analyzed.

[0044] Different feature elimination methods are used to comprehensively determine the importance relationship between the types of features in the original dataset and the target variable well spacing;

[0045] Features that rank in the top N in importance across multiple feature elimination methods will be retained, where N is at least equal to 2.

[0046] Furthermore, the machine learning algorithms include linear regression, multinomial regression, support vector machine, decision tree, multilayer perceptron, and convolutional neural network;

[0047] The model parameters include the loss function, optimizer, training epochs, and learning rate.

[0048] Furthermore, the process of comparing the calibration data with the predicted well spacing data output by the decision model, and determining the optimal decision model based on the comparison results, includes the following steps:

[0049] The test set is input into the learned decision model to predict well spacing, and the predicted well spacing data is output.

[0050] The predicted well spacing data is compared with the calibration data, and the prediction effect of the decision model is evaluated based on the comparison results.

[0051] The optimal decision model is selected based on the evaluation results.

[0052] Furthermore, the evaluation includes evaluating the prediction accuracy, fluctuation ratio, robustness, and generalization ability of the decision-making model.

[0053] Secondly, this invention proposes an intelligent control device for remote operation of magnetically guided tools, comprising:

[0054] The pre-drilling simulation module is used to simulate the magnetic steering operation before drilling and to develop an initial plan for the magnetic steering tool based on the simulation results.

[0055] The drilling data acquisition module is used to perform standard well operations and acquire drilling data based on the initial scheme of the magnetic steering tool; the drilling data includes primary data and secondary data;

[0056] The intelligent remote decision-making module is used to perform data mining by combining calibration data, drilling data and machine learning algorithms, and to determine the optimal decision model for the magnetic guidance tool. The intelligent decision scheme generated by the optimal decision model is then sent to the magnetic guidance tool for actual operation.

[0057] The post-drilling performance evaluation system is used to evaluate the initial plan and the execution process of the intelligent decision-making model of the magnetic guidance tool based on actual operation data, and to optimize the next operation plan.

[0058] Thirdly, the present invention proposes an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0059] Memory, which stores computer programs;

[0060] The processor, when executing the program stored in the memory, implements the intelligent remote operation control method for the magnetically guided tool.

[0061] Fourthly, the present invention proposes a computer-readable storage medium storing a computer program, which, when run, executes the aforementioned intelligent remote operation control method for magnetically guided tools.

[0062] The beneficial effects of this invention are:

[0063] This invention utilizes remote integrated data processing technology to achieve remote monitoring and analysis of magnetic steering operation data, significantly reducing the need for technicians to be physically present at the drilling site, thereby improving personnel efficiency and utilization. This remote operation mode not only saves labor costs but also accelerates decision-making, making operational adjustments more rapid and flexible.

[0064] This invention combines machine learning algorithms with data mining to automatically analyze calibration and drilling data, optimize the operating parameters of magnetic guidance tools, and form an optimal decision model. This intelligent decision-making process not only replaces traditional empirical formulas, reducing positioning errors caused by human factors, but also improves positioning accuracy and operational success rate. Actual operations are then conducted based on the plan output by the decision model, and feedback optimization of the initial plan developed through simulation is performed based on actual operation data, further improving the accuracy of the simulation plan. This invention, through its intelligent decision-making scheme, enables magnetic guidance tools to more flexibly cope with complex and changing underground environments, ensuring the stability and safety of the operation process. Implementing the technical solution of this invention can further advance drilling operations towards intelligence and automation, bringing new technological breakthroughs to the fields of oil, gas, and other energy extraction. It can not only improve operational efficiency and quality but also reduce operational risks and costs, playing a positive role in promoting the sustainable development of the entire industry.

[0065] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 A flowchart of a remote intelligent collaborative method for magnetic guidance tools proposed in an embodiment of the present invention is shown;

[0068] Figure 2 A flowchart illustrating the pre-drilling process in an embodiment of the present invention is shown;

[0069] Figure 3 The diagram shows the simulated electromagnetic field distribution before drilling in an embodiment of the present invention.

[0070] Figure 4A flowchart of drilling data acquisition in an embodiment of the present invention is shown;

[0071] Figure 5 A standard well location distribution map is shown in an embodiment of the present invention;

[0072] Figure 6 This invention illustrates a flowchart of processing drilling operation data to obtain primary data in an embodiment of the present invention;

[0073] Figure 7 This illustrates a flowchart of the data mining and decision model training self-learning process in an embodiment of the present invention.

[0074] Figure 8 A schematic diagram illustrating the importance of variable features in an embodiment of the present invention is shown;

[0075] Figure 9 A flowchart for determining the optimal decision model in an embodiment of the present invention is shown;

[0076] Figures 10A-10F The following figures illustrate the fitting results of linear regression, multinomial regression, SVR, CART, MLP, and CNN learning algorithms used in embodiments of the present invention.

[0077] Figure 11 A schematic diagram of the intelligent decision-making algorithm in an embodiment of the present invention is shown;

[0078] Figure 12 A schematic diagram of a remote intelligent control device for magnetic guidance tools proposed in an embodiment of the present invention is shown;

[0079] Figure 13 A schematic diagram of an electronic device according to an embodiment of the present invention is shown. Detailed Implementation

[0080] 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. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0081] This invention proposes an intelligent control method for remote operation of magnetically guided tools to address the problems of inconsistent downhole tool data leading to processing difficulties and low efficiency, and the inadequacy of intelligent remote decision-making systems in meeting production demands. The specific process is as follows: Figure 1 As shown, it includes the following steps:

[0082] S1: Conduct simulations before drilling to guide magnetic steerable drilling operations, and develop an initial plan for the magnetic steerable tool based on the simulation results; such as... Figure 2 As shown, the specific steps include:

[0083] S11. Input the first data of the target drilling area into the pre-drilling simulation system, including the necessary formation parameters, tool parameters and boundary conditions.

[0084] Specifically, formation parameters refer to the electromagnetic properties of the formation, including its electrical conductivity, magnetic permeability, and thickness; tool parameters include the size, shape, and magnet strength of the magnetically guided tool; boundary conditions include ground conditions, the surrounding environment, and the initial conditions of the electromagnetic field, specifically including terminals and grounding necessary for simulation.

[0085] S12. Calculate the theoretical trajectory path of the magnetic guidance tool under given conditions based on the first data, and project the theoretical trajectory path into the simulation model; specifically, a digital twin drilling simulation system can be applied to project the calculated trajectory path into the simulation model according to formulas, theories and realities; the method of calculating the trajectory path is not specifically limited in this invention, for example, it can be the minimum curvature method.

[0086] S13. Use a simulation model to simulate the electromagnetic field distribution of magnetic guidance measurement under different conditions and obtain the theoretical value of the magnetic guidance electromagnetic field; specifically, simulate the electromagnetic field distribution of magnetic guidance measurement under different distances, formations, drilling fluids and other conditions, and calculate the error between the theoretical value and the actual value;

[0087] S14. Reduce errors through computer-aided planning and simulation, output simulation results, and develop an initial scheme for the magnetic steering tool. The optimal initial scheme is the one with the smallest error between theoretical and actual values, including tool selection, drilling parameter settings, and drilling strategy.

[0088] In an exemplary embodiment of the present invention, the electromagnetic field distribution and measurement results of magnetic guidance measurement under different conditions such as distance, formation, and drilling fluid were simulated according to the methods in S11-S13 to assist in risk identification. (See also...) Figure 3 As shown, Figure 3 The simulation results show the results of different dimensions of magnetic flux density in the probe during magnetic steering measurement. Based on these simulation results, drilling operations can be assisted. In addition, they can be used to compare the actual drilling results, providing basic reference data for post-drilling evaluation and optimization of feedback control.

[0089] S2: Perform standard well operations based on the initial scheme of the magnetic steering tool and acquire drilling data; the drilling data includes primary data and secondary data;

[0090] In one embodiment of the present invention, the specific process of obtaining drilling data is as follows: Figure 4As shown, it includes the following steps:

[0091] S21, Pre-determine the well location distribution of standard wells in the target area, including the positional relationship between the distributed test wells and the target wells;

[0092] In one embodiment of the present invention, the positional relationship between the test well and the target well is as follows: Figure 5 As shown: Seven vertical wells, each 150m long, were drilled in a rectangular open area. A steel casing was installed in well T as the target well, and PE plastic pipes (drilled holes) were installed in wells O, A, M, L1, L2, and L3 as test wells.

[0093] S22, the relative distance from each test well to the target well, calculated using MWD and geometric scanning methods, is used as calibration data;

[0094] S23, based on the initial scheme of the magnetic guidance tool, perform the operation, obtain the drilling operation data of the magnetic guidance tool at different depths in the test well, and process the drilling operation data to obtain primary data; the drilling operation data includes magnetic flux, acceleration, probe attitude, magnetic field amplitude and relative orientation.

[0095] like Figure 6 As shown, processing the drilling operation data to obtain primary data includes the following steps:

[0096] S231. During the drilling process, standard well operations are performed based on the initial scheme of the magnetic steering tool, and drilling operation data is collected; specifically, the drilling data acquisition module in the drilling simulation decision system can be connected to the magnetic steering tool on site, and drilling operation data from the magnetic steering tool can be collected in real time in an end-to-end manner, and traditional empirical formulas can be used for calculation.

[0097] S232. Configure test parameters to exclude obviously erroneous or irrelevant drilling operation data;

[0098] Test parameters include data range limits and outlier detection standards to ensure the quality of data in subsequent analyses;

[0099] S233. Select drilling operation data that meet preset conditions and quality standards as primary data. After confirming the completeness and accuracy of the primary data, upload it to the cloud storage data platform. The selection process may include removing data that exceeds the normal range, eliminating duplicate data, and selecting based on specific conditions such as time window, tool status, etc.

[0100] In an exemplary embodiment of the present invention, the operational data collected in S21 includes, but is not limited to, geological information, drilling parameters, tool information, etc.

[0101] S24. Based on the primary data, the relative distance from each test well to the target well at different depths is predicted using traditional empirical formulas, and the predicted values ​​are used as secondary data.

[0102] In one embodiment of the present invention, the conventional empirical formula is as follows:

[0103]

[0104] Where r represents the predicted relative distance between the test well and the target well at a preset well depth, μ0 represents the vacuum permeability, I is the magnitude of the accumulated current in the standard well, and H... x H y H z These represent the amplitude components of the magnetic field signal along the X, Y, and Z axes at a preset depth in the test well. This represents the average error of the traditional empirical formula, which, in one embodiment of the present invention, is 35.375m.

[0105] S3: Data mining is performed using calibration data, drilling data, and machine learning algorithms to determine the optimal decision-making model for the magnetic guidance tool. The intelligent decision-making scheme generated by the optimal model is then distributed to the magnetic guidance tool for actual operation. This includes the following steps:

[0106] S31 combines drill-in data, calibration data, and machine learning algorithms to perform data mining and self-learning of decision models; such as... Figure 7 As shown, the specific steps include:

[0107] S311, Construct the original dataset using the primary data, secondary data, and calibration data;

[0108] S312, Perform feature selection on the primary and secondary data in the original dataset to obtain feature quantity data; including the following steps:

[0109] The types of feature quantities for all primary and secondary data in the original dataset are statistically analyzed.

[0110] Different feature elimination methods are used to comprehensively determine the importance relationship between the types of features in the original dataset and the target variable well spacing;

[0111] Feature data that ranks highly in importance among various feature elimination methods will be retained.

[0112] In specific implementation, the number of features is adjusted according to requirements. In one embodiment of the present invention, the original dataset contains multiple features, some of which are of varying importance, such as... Figure 8As shown (with variable parameters set to x1 to x10_2), the results in the figure are the features that are important in different model training processes selected through F test, mutual information test and recursive feature elimination method.

[0113] S313, after feature selection, the feature quantities are divided to construct a dataset, and the dataset is divided into a training set and a test set; the training set includes feature quantity data and calibration data; wherein, the feature quantity consists of some primary data and all secondary data.

[0114] Specifically, the feature quantities are divided into data from each test well using a random partitioning and single-well testing method to construct a dataset. The dataset includes feature quantities such as magnetic flux, acceleration, and probe attitude at different depths within the range, as well as the corresponding true values ​​of relative distances.

[0115] In one embodiment of the present invention, according to Figure 5 The well locations were distributed, and the data of the six wells in the dataset were named respectively. The actual values ​​of the relative distances of the six test wells to the target well were 55-59m, 8-10m, 29-31m, 55-59m, 60-68m, and 4-6m respectively.

[0116] S314, Select multiple machine learning algorithms to define the decision model of the magnetic guidance tool, and continuously optimize the decision model by changing the model parameters;

[0117] Specifically, a decision model can be defined using machine learning algorithms, and the model can be optimized by continuously changing its parameters. In this embodiment, candidate machine learning algorithms include mainstream algorithms such as linear regression, multinomial regression, support vector machine (SVR), decision tree (CART), multilayer perceptron (MLP), and convolutional neural network (CNN). Model parameters include loss function, optimizer, training epochs, and learning rate.

[0118] S315, The decision model is self-learned using the dataset to obtain the learned decision model.

[0119] S32, compare the calibration data with the predicted well spacing data output by the decision model, and determine the optimal decision model based on the comparison results; such as Figure 9 As shown, it includes the following steps:

[0120] S321, Input the test set into the learned decision model to predict well spacing, and output the predicted well spacing data;

[0121] S322, The predicted well spacing data is compared with the calibration data, and the prediction effect of the decision model is evaluated based on the comparison results;

[0122] S323, Select the optimal decision model as the best decision model based on the evaluation results.

[0123] In an exemplary embodiment of the present invention, the feature data in the dataset is randomly divided into a training set and a test set in a 4:1 ratio, which serve as input data for the decision model. The training set is input into the algorithm for self-learning to obtain the final trained model; the test set is input into the final decision model for well spacing prediction, and the predicted well spacing data is output and compared with the calibration data to obtain the fitting effect, as can be found in [reference needed]. Figures 10A-10F As shown, this allows for a comparison of the performance of decision models corresponding to different machine learning algorithms. Taking neural networks as an example, the training and prediction process of the model algorithm can be illustrated as follows... Figure 11 As shown.

[0124] In this embodiment, the evaluation comparison categories include prediction accuracy, fluctuation percentage, robustness, and generalization. In the steps of evaluating and comparing model performance: the traditional empirical formula achieved an accuracy of 37% with a randomly partitioned dataset; linear regression, multinomial regression, SVR, CART, MLP, and CNN improved by 48%, 48%, 0%, 60%, 61%, and 59% respectively compared to the traditional method with random partitioning. The traditional empirical formula had a 10% fluctuation percentage of 0% with a randomly partitioned dataset; linear regression, multinomial regression, SVR, CART, MLP, and CNN improved by 46%, 57%, 22%, 97%, 100%, and 74% respectively compared to the traditional method with random partitioning. In summary, when prediction accuracy is used as the evaluation metric, all six algorithms showed varying degrees of improvement; when the 10% fluctuation percentage is used as the evaluation metric, all five methods except SVR showed varying degrees of improvement. Among them, MLP has the most significant and stable improvement effect, and its robustness and generalization are also the highest compared with other algorithms. Therefore, MLP is the optimal choice for magnetic guidance far-field localization data mining algorithm.

[0125] S33, the intelligent decision-making scheme generated using the optimal decision-making model is sent to the magnetic guidance tool to guide its actual operation. When sending the remote decision-making scheme, the data decision-making scheme can be updated, and the types of messages sent include explanatory reports, guidance instructions, and electronic archives.

[0126] S4. Evaluate the execution process of the initial plan and intelligent decision-making model of the magnetic guidance tool based on the actual operation data, and optimize the next operation plan.

[0127] After completing this stage of drilling, the post-drilling effect evaluation system in this embodiment retains the operation data and provides real-time feedback on the operation effect based on the data. It evaluates the pre-drilling and drilling effects by using the overlapping points of the magnetic steering operation connections. In this embodiment, the evaluation system indicates that the magnetic steering operation connection overlap point is 0. The pre-drilling simulation effect and the actual drilling measurement show a consistency rate exceeding 90%. The post-drilling evaluation system then optimizes and generates an immediate effect evaluation, which is fed back to the pre-drilling simulation system. This effectively improves the reliability of the pre-drilling simulation effect for the next stage to 95%. Thus, the post-drilling effect evaluation system achieves a closed-loop feedback mechanism for supporting and controlling the next stage of drilling operations.

[0128] Based on the above method, another embodiment of the present invention proposes an intelligent control device for remote operation of magnetically guided tools, see [link to relevant documentation]. Figure 12 As shown. This system is a standard "pre-drilling-drilling-post-drilling" system for magnetic guidance technology, including a pre-drilling simulation system, a drilling simulation decision-making system, and a post-drilling effect evaluation system.

[0129] The pre-drilling simulation system is used to simulate magnetic steering operations before drilling and to develop an initial plan for the magnetic steering tool based on the simulation results. Specifically, the pre-drilling simulation system is a simulation system for magnetic steering operations, applying the concept of digital twins, and includes an input terminal, an electromagnetic simulation window, and an output terminal. The input terminal is used to input necessary formation, trajectory, tool parameters, and boundary conditions; the electromagnetic simulation window, through a built-in interface with numerical simulation software in the interpretation terminal software, displays the electromagnetic field distribution of magnetic steering measurements under different distances, formations, drilling fluids, and other conditions; the output terminal outputs simulation results to assist in risk warning and other guidance for on-site operations.

[0130] The drilling simulation decision-making system includes a drilling data acquisition module and an intelligent remote decision-making module;

[0131] The drilling data acquisition module is used to perform standard well operations and acquire drilling data based on the initial scheme of the magnetic steering tool. The drilling data includes primary and secondary data. Specifically, the drilling data acquisition module in the drilling simulation decision system can be connected to the field magnetic steering tool to collect actual operational data from the magnetic steering tool in real time via an end-to-end approach (processed and recorded as primary data). Secondary data is then calculated using traditional empirical formulas. The drilling data acquisition module needs to configure test parameters, generate data based on actual operational scenarios, exclude obviously erroneous or irrelevant actual operational data, and further filter the data according to preset conditions to obtain primary data for uploading. During the real-time data transmission process, data meeting preset conditions and quality standards should be uploaded to the cloud storage platform to optimize storage space and data processing efficiency. The drilling data acquisition module also needs to check the integrity and accuracy of the data to ensure that valid data is uploaded normally, and finally, the uploaded data is stored and processed.

[0132] The intelligent remote decision-making module utilizes calibration data, drilling data, and machine learning algorithms to perform data mining and determine the optimal decision model for the magnetic steering tool. The intelligent decision plan generated by this optimal model is then distributed to the magnetic steering tool for actual operation. Specifically, the intelligent remote decision-making module unifies secondary data and combines it with machine learning algorithms for data mining, used for simulation, inversion, and training of the decision model. Supervised learning of the machine learning algorithm is performed using calibration data as a baseline. The output of the decision model is compared with the calibration data. The decision model predicts well spacing data to assist in optimizing on-site operations. For difficult problems, explanation reports and steering commands are issued to guide far-field operations, thereby reducing errors and improving drilling positioning accuracy. Data is remotely updated to generate intelligent decision plans to guide drilling operations.

[0133] The post-drilling performance evaluation system assesses the execution process of the initial plan and intelligent decision-making model of the magnetically guided tool based on actual operational data, and optimizes the next operational plan. Specifically, it can evaluate the pre-drilling and in-drilling performance by working backward from the overlapping points of the magnetically guided operation. It can retain data and provide real-time feedback on operational performance, generating immediate performance evaluations to support the next stage of drilling. This makes the subsequently generated decision-making model more accurate, achieving collaborative support for remote intelligent drilling using magnetically guided tools.

[0134] Another exemplary embodiment of the present invention provides an electronic device. For example... Figure 13 As shown, the electronic device includes at least one processor 1301, at least one communication interface 1302, at least one memory 1303 and at least one communication bus 1304; wherein, the processor 1301, the communication interface 1302 and the memory 1303 communicate with each other through the communication bus 1304.

[0135] Memory 1303 stores computer programs;

[0136] The processor 1301, when executing the program stored in the memory 1303, implements the aforementioned intelligent remote operation control method for magnetic guide tools.

[0137] Optionally, the communication interface can be an interface of a communication module, such as the interface of a GSM module; the processor may be a CPU, an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The memory may include high-speed RAM and may also include non-volatile memory, such as at least one disk storage device. The memory stores a program, and the processor calls the program stored in the memory to execute some or all of the above-described method embodiments.

[0138] Based on the same inventive concept, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed, implements some or all of the above-described method embodiments. Optionally, the storage medium may be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0139] 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 remote intelligent control method for magnetically guided tools, characterized in that, Includes the following steps: Before drilling, conduct simulations of magnetic guidance operations and develop an initial plan for the magnetic guidance tool based on the simulation results. Standard well operations are performed based on the initial scheme of the magnetic steering tool, and drilling data is acquired; the drilling data includes primary data and secondary data; Data mining is performed by combining calibration data, drilling data, and machine learning algorithms to determine the optimal decision model for the magnetic guidance tool. The intelligent decision scheme generated by the optimal decision model is then sent to the magnetic guidance tool for actual operation. The initial plan and execution process of the intelligent decision-making model of the magnetic guidance tool are evaluated based on actual operation data, and the next operation plan is optimized.

2. The intelligent remote control method for magnetically guided tools according to claim 1, characterized in that, The simulation before drilling magnetic guidance operation, and the preparation of an initial plan for the magnetic guidance tool based on the simulation results, includes the following steps: Input the first data of the target drilling area into the pre-drilling simulation system. The first data includes formation parameters, tool parameters, and boundary conditions. The theoretical trajectory path of the magnetic guide tool under given conditions is calculated based on the first data, and the theoretical trajectory path is projected into the simulation model; The electromagnetic field distribution of magnetic guidance measurement under different conditions was simulated using a simulation model to obtain the theoretical value of the magnetic guidance electromagnetic field; The theoretical value is compared with the measured actual value of the magnetically guided electromagnetic field, and the error between the theoretical and actual values ​​is calculated. The first data is adjusted using computer-aided calculation methods until the error between the theoretical and actual values ​​is within a preset range. Then, the simulation results are output, and an initial scheme for the magnetic guidance tool is prepared. The initial scheme includes the measurement trajectory and measurement point information.

3. The intelligent remote control method for magnetically guided tools according to claim 2, characterized in that, The formation parameters include the electrical conductivity, magnetic permeability, and thickness of the formation; Tool parameters include the size, shape, and magnet strength of the magnetically guided tool; Boundary conditions include ground conditions, surrounding environment, and initial conditions of the electromagnetic field.

4. The intelligent remote control method for magnetically guided tools according to claim 1, characterized in that, The initial design of the magnetically guided tool includes tool selection, drilling parameter settings, and drilling strategy.

5. The intelligent remote control method for magnetically guided tools according to claim 1, characterized in that, The initial scheme based on the magnetic steering tool for standard well operations and obtaining drilling data includes the following steps: The well locations of standard wells are pre-defined in the target area, including the location relationship between the distributed test wells and the target wells; The relative distance from each test well to the target well, calculated using MWD and geometric scanning methods, is used as calibration data. Based on the initial scheme of the magnetic steering tool, the operation is carried out to obtain drilling operation data of the magnetic steering tool at different depths in the test well, and the drilling operation data is processed to obtain primary data; Based on the primary data, the relative distance from each test well to the target well at different depths is predicted using traditional empirical formulas, and the predicted values ​​are used as secondary data.

6. The intelligent control method for remote operation of magnetic guidance tools according to claim 5, characterized in that, Processing the drilling operation data to obtain primary data includes the following steps: Configure test parameters, and use these parameters to eliminate erroneous or irrelevant data in the drilling operation data to obtain primary data; after filtering and normalizing the primary data, secondary data is obtained. The completeness and accuracy of the secondary data are checked, and primary data is obtained after confirming that there are no errors.

7. The intelligent remote control method for magnetically guided tools according to claim 5, characterized in that, The drilling operation data includes magnetic flux, acceleration, probe attitude, magnetic field amplitude, and relative orientation.

8. The intelligent remote control method for magnetically guided tools according to claim 5, characterized in that, The traditional empirical formula is as follows: Where r represents the predicted relative distance between the test well and the target well at a preset well depth, μ0 represents the vacuum permeability, I is the magnitude of the accumulated current in the standard well, and H... x H y H z These represent the amplitude components of the magnetic field signal along the X, Y, and Z axes at the preset depth of the test well, respectively, while E represents the average error of the traditional empirical formula.

9. The intelligent remote control method for magnetically guided tools according to claim 1, characterized in that, The process of using calibration data, drilling data, and machine learning algorithms to perform data mining and determine the optimal decision model for the magnetic guidance tool includes the following steps: Combining drill data, calibration data, and machine learning algorithms to perform data mining and self-learning of decision models; The calibration data is compared with the predicted well spacing data output by the decision model, and the optimal decision model is determined based on the comparison results.

10. The intelligent remote control method for magnetically guided tools according to claim 9, characterized in that, The process of combining drill data, calibration data, and machine learning algorithms for data mining and self-learning of decision models includes the following steps: The original dataset is constructed using the primary data, secondary data, and calibration data; Feature selection is performed on the primary and secondary data in the original dataset to obtain feature quantity data; A dataset is constructed by dividing feature data and calibration data, and the dataset is further divided into a training set and a test set; the training set includes feature data and calibration data; the test set includes feature data. Multiple machine learning algorithms are selected to define the decision model for the magnetic guidance tool, and the decision model is continuously optimized by changing the model parameters; The decision model is self-learned using the training set to obtain the learned decision model.

11. The intelligent control method for remote operation of magnetically guided tools according to claim 10, characterized in that, Feature selection of the original dataset includes: The types of feature quantities for all primary and secondary data in the original dataset are statistically analyzed. Different feature elimination methods are used to comprehensively determine the importance relationship between the types of features in the original dataset and the target variable well spacing; Features that rank in the top N in importance across multiple feature elimination methods will be retained, where N is at least equal to 2.

12. The intelligent control method for remote operation of magnetically guided tools according to claim 9, characterized in that, The machine learning algorithms include linear regression, multinomial regression, support vector machine, decision tree, multilayer perceptron, and convolutional neural network; The model parameters include the loss function, optimizer, training epochs, and learning rate.

13. The intelligent control method for remote operation of magnetically guided tools according to claim 9, characterized in that, The process of comparing calibration data with the predicted well spacing data output by the decision model, and determining the optimal decision model based on the comparison results, includes the following steps: The test set is input into the learned decision model to predict well spacing, and the predicted well spacing data is output. The predicted well spacing data is compared with the calibration data, and the prediction effect of the decision model is evaluated based on the comparison results. The optimal decision model is selected based on the evaluation results.

14. The intelligent remote control method for magnetically guided tools according to claim 13, characterized in that, The evaluation includes assessing the prediction accuracy, fluctuation ratio, robustness, and generalization ability of the decision-making model.

15. A remote intelligent control device for magnetically guided tools, characterized in that, include: The pre-drilling simulation module is used to simulate the magnetic steering operation before drilling and to develop an initial plan for the magnetic steering tool based on the simulation results. The drilling data acquisition module is used to perform standard well operations and acquire drilling data based on the initial scheme of the magnetic steering tool; the drilling data includes primary data and secondary data; The intelligent remote decision-making module is used to perform data mining by combining calibration data, drilling data and machine learning algorithms, and to determine the optimal decision model for the magnetic guidance tool. The intelligent decision scheme generated by the optimal decision model is then sent to the magnetic guidance tool for actual operation. The post-drilling performance evaluation system is used to evaluate the initial plan and the execution process of the intelligent decision-making model of the magnetic guidance tool based on actual operation data, and to optimize the next operation plan.

16. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, which stores computer programs; A processor, when executing a program stored in a memory, implements the intelligent remote operation control method for magnetically guided tools as described in any one of claims 1-14.

17. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is run, it executes the intelligent remote operation control method for magnetic guidance tools as described in any one of claims 1-14.

Images