Drilling trajectory automatic control method based on bezier curve and uncertainty ellipse

Abstract

The present application relates to a kind of drilling trajectory automatic control method based on Bezier curve and uncertainty ellipse, according to geological target and engineering constraint, the bending section of wellbore trajectory is optimized and designed using cubic Bezier curve model, and global planning trajectory is generated;During drilling, based on measurement-while-drilling data, the confidence region representing the uncertainty of drill bit position is calculated in real time;The offset vector of actual drilling position relative to global planning trajectory is calculated, and according to the comparison result of offset vector and confidence region, it is determined whether significant deviation occurs;If it is determined that significant deviation occurs, the current actual drilling position is taken as the starting point, the front target point of global planning trajectory is taken as the terminal point, the correction path is generated using cubic Bezier curve model, and the downhole tool is controlled to execute the correction path.The present application significantly improves the trajectory control accuracy, operation safety and drilling efficiency.

Description

Technical Field

[0001] This invention relates to the field of directional drilling engineering technology in oil and gas exploration and development, specifically to an automatic drilling trajectory control method and system that integrates advanced geometric trajectory planning and real-time uncertainty quantification technology for actual drilling operations. Background Technology

[0002] Directional drilling technology is one of the key technologies in modern oil and gas exploration and development. Its core lies in the precise control of the wellbore trajectory, enabling it to bypass obstacles, accurately reach the target formation, and maximize the reservoir contact area. As drilling targets become increasingly complex (such as extended reach wells, branch wells, and three-dimensional obstacle bypass wells), higher demands are placed on the smoothness and accuracy of trajectory design, as well as the real-time control capabilities during the drilling process.

[0003] Traditional wellbore trajectory design methods are mainly based on simple geometric models, such as the "tangent-circular arc-tangent" method or the minimum curvature method. These methods are relatively simple in calculation, but they have obvious shortcomings when dealing with complex three-dimensional trajectories: the trajectory is not smooth enough, which easily produces a large dogleg severity, increasing drill string friction, torque and the risk of stuck pipe; and they lack flexibility and are difficult to optimize locally.

[0004] In terms of drilling process control, current methods mainly rely on the driller or directional engineer manually adjusting parameters such as tool face angle based on measurement-while-drilling data. This approach is inefficient, highly dependent on personnel experience, and struggles to handle complex downhole conditions and measurement uncertainties. It often leads to deviations from the designed path, requiring multiple corrections and increasing non-productive time and operating costs.

[0005] In recent years, the development of rotary steerable systems (RSS) and measurement-while-drilling (MWD) systems has laid the foundation for automated drilling. However, achieving fully automated closed-loop drilling still faces two major challenges: first, how to design a theoretical trajectory that satisfies geological and engineering constraints (such as target point and dogleg limitations) while being as smooth and efficient as possible; and second, how to detect trajectory deviations in real time and accurately and automatically generate safe and effective correction commands in the presence of uncertainties such as tool errors and formation heterogeneity.

[0006] Existing patented technologies (such as CN115142791A) propose an automatic curvature control method for rotary steering systems, focusing on solving well depth sensing and single control algorithm problems. However, they do not address overall trajectory optimization design based on advanced geometric models, nor do they provide quantitative processing of measurement uncertainties or intelligent correction mechanisms based on them. Furthermore, fully physical testing of control algorithms is costly, while fully digital simulation lacks sufficient confidence. Therefore, a systematic solution integrating advanced trajectory planning algorithms with high-confidence real-time simulation correction capabilities is urgently needed. Summary of the Invention

[0007] This invention provides an automatic drilling trajectory control method and system that deeply integrates cubic Bézier curve geometric optimization and real-time uncertainty quantification. This addresses the technical bottlenecks in existing drilling trajectory control technologies, such as poor trajectory smoothness, reliance on human experience for control, lack of quantitative risk warning, and secondary problems caused by crude correction methods.

[0008] The technical solution of this invention is as follows: An automatic drilling trajectory control method based on Bézier curves and uncertain ellipses is proposed, as follows: Based on geological objectives and engineering constraints, the curved section of the wellbore trajectory is optimized using a cubic Bézier curve model to generate a global planning trajectory. During actual drilling, confidence regions characterizing the uncertainty of drill bit position are calculated in real time based on measurement-while-drilling data; Calculate the offset vector of the actual drilling position relative to the global planning trajectory, and determine whether a significant offset has occurred based on the comparison result of the offset vector and the confidence region; If a significant deviation is determined, a correction path is generated using a cubic Bézier curve model, starting from the current actual drilling position and ending at the target point ahead of the global planned trajectory, and the downhole tool is controlled to execute the correction path. The specific process of obtaining the global planning trajectory is as follows: a cubic Bézier curve model is used to model the curved section of the wellbore trajectory to generate an initial global planning trajectory; an optimization problem including optimization variables and optimization objectives is constructed, and the optimization problem is solved under geological objectives and engineering constraints to obtain the global planning trajectory.

[0009] The geological targets include target coordinates; the engineering constraints include: (1) the instantaneous dogleg degree of any point on the curved section of the wellbore trajectory. Maximum allowable degree of sycophancy (2) The spatial distance between any actual drilling measurement point on the curved section of the well trajectory and the corresponding theoretical reference point in the global planning trajectory is ≤ the allowable error radius of the corresponding target point. (3) Obstacle avoidance distance ≥ collision avoidance safety distance .

[0010] This also includes setting the distance of the attraction point at the starting point of the curved section of the wellbore trajectory. and the distance to the attraction point at the end of the curved section of the wellbore trajectory The range of values ​​is used to model and optimize the curved section of the wellbore trajectory with the goal of minimizing the arc length of the cubic Bézier curve.

[0011] The confidence region is an uncertain ellipse; its semi-axis length and direction are determined by eigenvalue decomposition of the covariance matrix of the uncertainty of the current measurement point position. The semi-axis length is the product of the square root of the eigenvalue and the confidence factor, and the direction is the same as the direction of the eigenvector.

[0012] The specific process for determining whether a significant shift has occurred is as follows: project the shift vector onto the plane where the confidence region is located to obtain a plane shift vector, and calculate the magnitude of the plane shift vector; calculate the semi-axis length of the confidence region in the direction of the plane shift vector; and determine the significant shift by comparing the magnitude and the semi-axis length.

[0013] The specific process for generating the correction path is as follows: On the global planning trajectory, find the target point located in the actual direction of the wellbore and in front of the current actual drilling position; with the current actual drilling position as the starting point and the target point as the ending point, construct a cubic Bézier curve as a candidate correction path; perform engineering constraint verification on the candidate correction path; if the verification passes, the candidate correction path becomes the final correction path; if the verification fails, a new target point is determined, a new candidate correction path is generated and verified again, until a final correction path that meets the engineering constraint verification is obtained.

[0014] The tangential direction at the start and end points of the candidate correction path / new candidate correction path is consistent with the actual wellbore direction and the global planned trajectory direction.

[0015] The cubic Bézier curve model Specifically: In the formula: This represents the proportion of the position from the starting point to the ending point in the cubic Bézier curve formula; This is the first control point; This is the second control point; This marks the end point of the curved section of the wellbore trajectory; for The unit tangent vector at that point; for The unit tangent vector at that point; The distance from the attraction point at the starting point of the curved section of the wellbore trajectory; This is the distance to the attraction point at the end of the curved section of the wellbore trajectory.

[0016] The technical effects of this invention are as follows: This invention employs cubic Bézier curves for trajectory planning and correction, generating a smooth wellbore with continuous curvature and controlled dogleg degree, reducing friction and stuck drill bit risks. Based on the ISCWSA error model, an uncertainty ellipse is constructed in real time to quantify the drill bit position confidence range. Scientific offset judgment is achieved by comparing the offset vector magnitude with the ellipse radius, avoiding miscorrection. This forms a fully automated closed loop of "planning-monitoring-judgment-correction," significantly improving trajectory control accuracy, operational safety, and drilling efficiency. Attached Figure Description

[0017] Figure 1 This is a diagram of the architecture of the present invention.

[0018] Figure 2 This is a flowchart for the global trajectory planning.

[0019] Figure 3 A path diagram for determining whether a significant shift has occurred and for correction.

[0020] Figure 4 This is a schematic diagram illustrating the optimization effect of the Bezier curve on the curved section of the wellbore trajectory.

[0021] Figure 5 This is a schematic diagram for determining the uncertainty ellipse and offset. Detailed Implementation

[0022] Example 1—An automatic control method for drilling trajectory based on Bézier curves and uncertain ellipses, the method is as follows.

[0023] Step 1: Based on geological objectives and engineering constraints, optimize the wellbore trajectory using a parametric curve model to generate a global planning trajectory; 1. Input and Initialization: Input target sequence ;in, For the wellhead, each target point contains its spatial coordinates. And the corresponding engineering constraints; wherein, the engineering constraints include: (1) the instantaneous dogleg degree at any point on the curved section of the well trajectory. Maximum allowable degree of sycophancy (2) The spatial distance between any actual drilling measurement point on the curved section of the well trajectory and the corresponding theoretical reference point in the global planning trajectory is ≤ the allowable error radius of the corresponding target point. (3) Obstacle avoidance distance ≥ collision avoidance safety distance ; 2. Intelligent segment recognition: For adjacent target points and Calculate its design well inclination angle Change and design azimuth angle Change Simultaneously set the threshold angle. .like If the condition is met, the segment is identified as a wellbore azimuth-maintaining segment and estimated as a straight line; otherwise, it is identified as a curved segment of the wellbore trajectory. 3. Modeling and optimization of Bézier curves for curved sections of wellbore trajectory: a) Model establishment: For each curved segment of the well trajectory, a cubic Bézier curve model is defined. : In the formula: This represents the proportion of the position from the starting point to the ending point in the cubic Bézier curve formula; This marks the starting point of the curved section of the wellbore trajectory; This is the first control point to be optimized; This is the second control point to be optimized; This marks the end point of the curved section of the wellbore trajectory; b) Parameterization of control points to be optimized: To ensure the continuity of the tangent at the connection point of the curved section of the wellbore trajectory (i.e., smooth wellbore trajectory), the first control point to be optimized... Second control point to be optimized Determined by the following formula: In the formula: for The unit tangent vector at that point; for The unit tangent vector at that point; The distance from the attraction point at the starting point of the curved section of the wellbore trajectory; The distance to the attraction point at the end of the curved section of the wellbore trajectory; The design inclination angle and design azimuth angle at the starting point of the curved section of the well trajectory are calculated. The design inclination angle and design azimuth angle at the end of the curved section of the well trajectory are calculated as follows: ; c) Optimization under constraints: Construct an optimization problem with the objective function of minimizing the arc length. The decision variable is the distance from the attraction point at the starting point of the curved section of the wellbore trajectory. Distance of the attraction point at the end of the curved section of the wellbore trajectory The constraints include: 1) Instantaneous dogleg degree at any point on the curved section of the wellbore trajectory Maximum allowable degree of sycophancy ; From cubic Bézier curves Calculation of the first and second derivatives; 2) Spatial obstacle avoidance constraint: The minimum distance between the curved section of the wellbore trajectory and a known obstacle (such as the trajectory of an adjacent well) is greater than... That is, obstacle avoidance distance ≥ collision avoidance safety distance ; 3) The spatial distance between any actual drilled measuring point on the curved section of the well trajectory and the corresponding theoretical reference point in the global planning trajectory is ≤ the allowable error radius of the corresponding target point. ; d) Solution and Trajectory Generation: The above optimization problem is solved using sequential quadratic programming or a genetic algorithm to obtain the attraction point distance at the starting point of the curved section of the optimal wellbore trajectory. Distance from the attraction point at the end of the curved section of the optimal wellbore trajectory Thus, the optimal first control point is obtained. and the optimal second control point This leads to the global planning trajectory.

[0024] Step 2: During actual drilling, based on the measurement-while-drilling data, the confidence region characterizing the uncertainty of the drill bit position is calculated in real time; 1. Input and Initialization: Receives real-time data from actual drilling measurement points obtained from the measurement-while-drilling tool. Data, including depth measurements Design well inclination angle Design azimuth angle And possibly raw readings of tool facet angle, accelerometer and magnetometer; 2. Calculation of Uncertain Ellipses: a) Error source modeling: Consider the main systematic error sources, including: depth scale factor error Acceleration scaling factor error Accelerometer zero bias error Magnetometer scale factor error Magnetometer zero bias error And errors caused by wellbore geometry effects; b) Covariance matrix calculation: at the current actual drilling point At that location, based on its designed well inclination angle Design azimuth angle Error model, calculate the covariance matrix of its positional uncertainty in the northeast vertical coordinate system. The covariance matrix It characterizes the two-dimensional (in a plane perpendicular to the borehole direction) probability distribution of the true position of the drill bit around the measurement position; c) Generation of Uncertainty Ellipses: Covariance Matrix eigenvalues ​​( , The uncertainty ellipse is defined by the eigenvectors and the major and minor axes. and , of which Confidence factor ( =3 corresponds to a 99.7% confidence level. The direction of the uncertainty ellipse is determined by the eigenvector. This uncertainty ellipse varies with each new drilled point. Dynamically updated.

[0025] Step 3: Calculate the offset vector of the actual drilling position relative to the global planned trajectory. Based on the comparison between the offset vector and the uncertainty ellipse, determine whether a significant offset has occurred. 1. Trajectory Comparison: For each actual drilling point Find its measured depth on the global planning trajectory. Same theoretical reference point ; 2. Offset Vector Calculation: Calculated from the actual drilled measuring point To the theoretical reference point Spatial offset vector ; 3. Significance assessment: a) Spatial offset vector Projected onto actual drilling points The plane offset vector is obtained on the plane containing the uncertain ellipse at the point. ; b) Calculate the offset vector along the plane on this plane. Direction, distance from the center of the uncertain ellipse to the boundary ; c) Judgment logic: If Then it is determined that the actual drilling point is at the test point. A significant shift has occurred, meaning the shift exceeds the range that the current measurement error can explain, requiring correction. Otherwise, it is considered to be within the normal fluctuation range.

[0026] Step 4: If a significant shift is determined, then use the current actual drilling measurement point. Starting from the target point ahead of the globally planned trajectory and ending at the target point ahead, a cubic Bézier curve model is used. Generate a correction path and control the downhole tool to execute the correction path: 1. Correction trigger and target point search: When continuous When a drilled point is determined to be significantly offset, the system triggers a correction. The last drilled point to be offset is taken as the offset point. This is used as the starting point for correction; on the globally planned trajectory, the offset measurement point is used as the starting point. Current depth is ahead of the curve Starting from the current position, search forward for a suitable target return point. Target return point It should be from the offset measuring point Return to target point The angle between the straight line direction and the tangent direction at the global planning trajectory is the smallest; 2. Generation of cubic Bézier curves for the correction path: using offset measurement points Starting point Target return point End point Using the cubic Bézier curve model Generate candidate correction paths. Its first control point to be optimized. Second control point to be optimized The determination of the starting point must meet the following conditions: The tangential direction is consistent with the current wellbore direction, and the endpoint is... The tangential and global planned trajectory at the target return point The tangential direction is consistent at the location; 3. Path security and feasibility verification: a) Dogleg degree check: Calculate the maximum dogleg degree on the candidate correction path. Must meet Maximum allowable degree of sycophancy ; b) Collision avoidance check: Ensure that the distance between the candidate correction path and all known obstacles is greater than the collision avoidance safety distance. That is, obstacle avoidance distance ≥ collision avoidance safety distance ; c) The spatial distance between any actual drilled measuring point on the curved section of the wellbore trajectory and the corresponding theoretical reference point in the global planning trajectory is ≤ the allowable error radius of the corresponding target point. ; d) Engineering feasibility verification: Verify whether the tool's azimuth / or azimuth adjustment capability required for the candidate correction path is within the actual capability range of the downhole tool; 4. Command Generation and Execution: If the verification passes, the system discretizes the candidate correction path into a series of dense guide target points, and uses inverse kinematics model or lookup table method to calculate and issue the corresponding rotation guide tool command (tool face angle) in real time. Guidance rate If the verification fails, continue forward along the globally planned trajectory to select the next target return point. Repeat steps 2-3 until the final correction path is found; 5. Closed-loop monitoring and state switching: The system continuously monitors the drill bit's tracking along the final correction path. When the system detects that the drill bit's actual position has entered the uncertainty ellipse of the preceding point on the global planning trajectory and has stably followed it beyond the predetermined advance, the correction is considered complete. The system automatically switches the control target from the final correction path back to the original global planning trajectory, achieving seamless transition.

[0027] As a further optimization of this method, an online replanning function is also included: when logging-while-drilling data indicates a significant difference between the geological model and the design, or when unforeseen obstacles are encountered, replanning can be triggered manually or automatically. In this case, the global planning trajectory and optimization module uses the current actual bottom hole offset measurement point. Starting from the new point, with the remaining target points as the objective, and taking into account the latest geological model and all constraints, step S1 is re-executed online to generate and optimize the new global planning trajectory for subsequent well sections. After confirmation, the new global planning trajectory replaces the corresponding part of the original global planning trajectory, thereby realizing the dynamic adaptive adjustment of the global planning trajectory.

[0028] Specific Application Case 1—Taking a land-based three-dimensional directional well as an example, this demonstrates the entire process of the method of the present invention from design to actual drilling closed-loop control. The well design has three target points, A, B, and C, and requires bypassing an adjacent well.

[0029] 1. Trajectory Planning Stage Target coordinates (northeast vertical coordinate system, unit: meters): Wellhead O: (N0=0, E0=0, V0=0) Target A: (N_A=500, E_A=300, V_A=1200) Target B: (N_B=800, E_B=800, V_B=2000) Target C: (N_C=500, E_C=1200, V_C=2500) Engineering constraints: Maximum dogleg degree Permissible error radius of the target point Collision avoidance safety distance ; Error model parameters: Using the ISCWSA model, configure the systematic error parameters corresponding to the measurement while drilling tools used, such as depth scale factor error. Acceleration scaling factor error wait; The trajectory planning and optimization module is started: a) Section Identification: Calculate the wellbore inclination and azimuth changes for each segment from wellhead O to target point A, target point A to target point B, and target point B to target point C. The segment from wellhead O to target point A requires the completion of inclination building and azimuth adjustment from vertical to directional, and is identified as "Wellbore Trajectory Curved Segment 1"; the segment from target point A to target point B requires continued inclination building and azimuth adjustment, and is identified as "Wellbore Trajectory Curved Segment 2"; the segment from target point B to target point C is a stable inclination and azimuth segment, and is identified as "Maintained Segment"; b) Optimization of the curved section of the well trajectory: Taking "curved section 1 of the well trajectory" as an example. The wellhead O (well inclination 0°) is the endpoint of the curved section of the wellbore trajectory. For target point A, set the distance from the attraction point at the starting point of the initial curved section of the wellbore trajectory. Distance of the attraction point at the end of the curved section of the wellbore trajectory ;calculate unit tangent vector at point (Vertically downward); Calculate the design well inclination angle based on the coordinates of target point A. and design azimuth ,get unit tangent vector at point ; Identify the control points to be optimized: Constructing cubic Bézier curves And calculate its arc length and ; c) Constraint Optimization and Solution: A sequential quadratic programming algorithm is used to optimize and solve the problem. and Optimize within the specified range. The optimization objective is to minimize the arc length, and the constraints are as follows: The collision avoidance safety distance is no less than 30m. Through iteration, the optimal attraction point distance at the starting point of the curved section of the wellbore trajectory is obtained. The distance from the attraction point at the end of the curved section of the optimal wellbore trajectory At this point, the maximum dogleg angle is 9.8° / 30m, which meets the requirements. Similarly, the optimization of "wellbore trajectory curvature segment 2" is completed. The final result is as follows: Figure 2 The global planning trajectory shown is the blue curve.

[0030] 2. Actual drilling closed-loop control and dynamic correction When drilling reaches 1500 meters (in the initial stage of "wellbore trajectory curvature section 2"), monitor the status in real time; a) Monitoring while drilling and quantification of uncertainty: Uploading measured points from the measurement while drilling tool ( Design well inclination angle Design azimuth angle The real-time monitoring and decision-making module calculates the covariance matrix of the point location based on the ISCWSA model. Take the confidence factor. Generate a major semi-axis short half shaft An ellipse with a specific orientation of uncertainty, displayed in a 3D view (e.g.) Figure 3 (gray ellipse) b) Offset Intelligent Judgment: The system finds a theoretical reference point at a depth of 1500m on the globally planned trajectory. (Design well inclination angle) Design azimuth angle ). Calculate the spatial offset vector And project it onto the plane containing the uncertain ellipse, to obtain , module length Calculate the uncertainty ellipse in Semi-axis length in the direction Since 3.8m > 2.9m, it is determined to be a significant offset; c) Dynamic correction: The system continuously detects significant shifts at three measuring points, triggering correction; Target point search: using the current offset measurement point ( Starting from a depth of 1540m (30m ahead), search forward along the globally planned trajectory to find the target return point. (Located on the global planning trajectory) At this point, the angle between the tangent of the global planned trajectory and the current wellbore direction is the smallest; Path generation and verification: using offset measurement points and target return point Using the starting and ending points, candidate correction paths (yellow dashed lines) of cubic Bézier curves are generated. After verification, their maximum dogleg degree is determined. < The collision avoidance safety distance is 35m > 30m. Verification passed. Command Execution and Return: The calibration path is converted into a sequence of guiding commands and issued to the measurement-while-drilling (MWD) tool. The drill bit smoothly drills along the yellow dashed line path. When drilling reaches... When the actual drilling point enters the uncertainty ellipse of the corresponding point on the global planning trajectory, it will automatically switch back to the original blue global planning trajectory after confirmation, thus completing the closed-loop correction.

[0031] Specific Application Case 2: Intelligent Safety Correction and Adaptive Control for Unexpected Rigid Interlayers

[0032] This specific application case 2 aims to demonstrate in detail the adaptive control process of the system of the present invention in dealing with sudden and challenging geological conditions (such as encountering hard interlayers that cause temporary degradation of tool performance), showcasing its advanced early warning, intelligent decision-making, and safety-first approach. This scenario deeply tests the system's comprehensive decision-making capabilities under engineering constraints and uncertainties.

[0033] 1. Scene setting and initial state a) Well type and target: A horizontal well for unconventional oil and gas development. The designed well depth structure is as follows: vertical section (0-1500m) → build-up section (1500-2200m, build-up rate 8° / 30m, well inclination increasing from 0° to 85°) → horizontal section (>2200m, maintaining 85° well inclination); b) Engineering constraints: Maximum permissible dogleg degree Target entry requirements: At a vertical depth of 2200m ± 0.5m, the well inclination must reach 85° ± 1° to ensure accurate target entry. The upper limit of the drilling fluid performance and the drilling tool assembly's build-up capacity is 10.5° / 30m. c) System Initialization: The global trajectory planning and optimization module has generated a smooth global planning trajectory using cubic Bézier curves, based on the above design. The error model has been configured, and the uncertainty elliptic confidence factor has been set. (Corresponding to approximately 95% confidence level).

[0034] 2. Geological mutations and early system perception a) Event: Drilling to the measured depth At that time (the designed well inclination angle should be 49.1°), the drill bit encountered a dense calcareous sandstone interlayer (quartz content >80%) with a thickness of about 2 meters, which was not clearly identified in the previous seismic data. This dense calcareous sandstone interlayer has extremely poor drillability, which leads to a decrease in the drill bit's rock breaking efficiency. The actual effective build-up rate of the downhole tools decreased from the designed 8° / 30m to about 5.0-5.5° / 30m. b) System data flow and early warning: Data acquisition: The measurement-while-drilling tool uploads data at a rate of one point per meter. From the measuring points, the system obtains: the designed well inclination angle. Design azimuth angle ; Uncertainty Ellipse Calculation: The real-time monitoring and decision-making module calculates the covariance matrix of the measuring point based on the current wellbore geometric parameters and error model. After eigenvalue decomposition, an uncertain ellipse with a major axis of approximately 3.5m and a minor axis of approximately 1.8m is generated. c) Offset Trend Analysis: The system found the corresponding path in the global trajectory planning. Theoretical points (Design well inclination angle 49.3°); calculate the offset vector and project it to obtain the modulus. ; Calculate the length of the semi-axis of the uncertain ellipse in this direction At this point, 1.2m < 2.8m, and the "significant offset" alarm was not triggered; However, the trend analysis submodule within the system is activated. This module analyzes the offset vector sequence of the five most recent measurement points. The linear regression slope was calculated. Analysis revealed a stable linear increasing trend in the offset, with the direction of increase consistent with the "insufficient well deviation growth" predicted by the geological steering model. A persistent negative offset trend was detected, which may be related to formation changes or fluctuations in tool performance.

[0035] 3. Offset confirmation and failure of the first calibration attempt a) Offset Confirmation: Continue drilling until The measurement point data is: actual well inclination angle. (The design value should be 51.0°). At this time... , .satisfy A significant shift has occurred. b) Initial correction path generation and security veto: Calibration trigger: Activate the dynamic calibration module; Initial target point selection: based on the current offset measurement point ( Starting from the default parameters, a leading search is performed on the global trajectory planning. Nearby points, select target return point ( (Design well inclination angle 53.0°). Path generation: using the current offset measurement point return point to target Using the start and end points, a cubic Bezier correction path is constructed for the first time. ; Security verification: Dog leg severity verification: Calculation Instantaneous dogleg degree along the route; found at the beginning of the path (approximately =0.1), in order to quickly compensate for the current 1.5° well inclination difference and catch up with the design slope in a short distance, instantaneous Reaching 14.2° / 30m; Constraint comparison: 14.2° / 30m> (12.0° / 30m), and this value also exceeds the upper limit of the design build-up capacity of the measurement while drilling tool (10.5° / 30m). Intelligent veto: Strictly adhering to the principle of safety first, the decision is made. Not feasible, and the reason for rejection was recorded: "The maximum dogleg angle of the path (14.2° / 30m) exceeds the safety threshold (12.0° / 30m) and tool capability (10.5° / 30m)." This decision avoided the possibility of manual forced correction under on-site pressure.

[0036] 4. Intelligent search and execution of adaptive safe paths a) Expanding the search strategy: Activating the adaptive search algorithm. This algorithm dynamically increases the look-ahead distance. And assess the feasibility of reaching further forward targets; b) Iterative search process: attempts : Generation path The maximum dogleg angle is 13.1° / 30m, which still exceeds the limit; therefore, the attempt is rejected. : Generation path The maximum dogleg angle is 11.8° / 30m, which is lower than... However, it still exceeds the capability of the measurement-while-drilling tool by 10.5° / 30m, so it is rejected; an attempt was made. : Generation path Maximum dogleg angle 9.8° / 30m; c) Final verification and confirmation: [This refers to a specific verification process or confirmation.] Perform a full verification: Dogleg angle: 9.8° / 30m < 12.0° / 30m, and < 10.5° / 30m, pass; Collision avoidance scan: Collision avoidance safety distance > 50m; No collision, pass; Target Feasibility Verification: System Simulation from Starting from point A, drilling continued at the maximum design capacity of the measurement-while-drilling tool (10.5° / 30m) to verify whether the well inclination could be raised to 85° before reaching the target depth of 2200m. Simulation results: It can be completed, with a margin of safety. d) Path confirmation and execution: The system will The final safety correction path is determined (displayed as a highlighted green curve in the 3D view). Subsequently, it is discretized into a series of high-density target guidance points, and the corresponding tool face angle and guidance force command sequence are calculated through the inverse kinematics model and issued to the downhole rotary steering tool for execution.

[0037] 5. Correction effect and post-system evaluation a) Execution process: The drill bit begins to move along the green correction path. Drilling. This approach allows for a gradual increase in well inclination over a longer horizontal displacement, effectively mitigating the risks of forcing high build-up rates in hard interlayers; b) Performance Monitoring: Real-time monitoring shows that the actual dogleg angle is stable between 9.0-10.0° / 30m, and the tool operates smoothly. When drilling to approximately... At that time, it was detected that the actual drilling trajectory point had entered the uncertainty ellipse of the corresponding point on the global planning trajectory, and the subsequent 3 points maintained this trend; c) State Switching and Completion: The system determines that the "safety correction phase" objective has been achieved and automatically and smoothly returns control to the original global planning trajectory for tracking. At this point, the wellbore trajectory has safely and stably rejoined the optimal design trajectory from the "below design line" state through a long and gentle curve. d) Final Results: Although approximately 75 meters of additional drilling footage was achieved due to navigating around a hard interlayer, no complex situations such as tool face loss of control, stuck tools, or the formation of harmful doglegs occurred throughout the operation. Ultimately, the well accurately entered the target window at a vertical depth of 2200.3m and an inclination of 84.9°, fully meeting geological requirements. Post-operation analysis showed that the system's "safety veto" and "adaptive search" decisions successfully avoided a highly likely downhole failure (such as tool damage or keyway formation). Although this resulted in a slight increase in drilling footage, it ensured the overall safety of the operation and the achievement of the final objective, resulting in significant comprehensive economic benefits.

[0038] This embodiment demonstrates in detail the complete decision chain of the system of the present invention, from "trend warning" to "safety veto", then to "intelligent search" and finally "safe execution", highlighting its core value of prioritizing engineering safety constraints and its powerful adaptive control capabilities when dealing with uncertainties in actual drilling.

Claims

1. An automatic drilling trajectory control method based on Bézier curves and uncertain ellipses, characterized in that, The method is as follows: Based on geological objectives and engineering constraints, the curved section of the wellbore trajectory is optimized using a cubic Bézier curve model to generate a global planning trajectory. During actual drilling, confidence regions characterizing the uncertainty of drill bit position are calculated in real time based on measurement-while-drilling data; Calculate the offset vector of the actual drilling position relative to the global planning trajectory, and determine whether a significant offset has occurred based on the comparison result of the offset vector and the confidence region; If a significant deviation is determined, a correction path is generated using a cubic Bézier curve model, starting from the current actual drilling position and ending at the target point ahead of the globally planned trajectory. The downhole tool is then controlled to execute the correction path.

2. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, The specific process of obtaining the global planning trajectory is as follows: a cubic Bézier curve model is used to model the curved section of the wellbore trajectory to generate an initial global planning trajectory; an optimization problem including optimization variables and optimization objectives is constructed, and the optimization problem is solved under geological objectives and engineering constraints to obtain the global planning trajectory.

3. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 2, characterized in that, The geological targets include target coordinates; the engineering constraints include: (1) the instantaneous dogleg degree of any point on the curved section of the wellbore trajectory. Maximum allowable degree of sycophancy (2) The spatial distance between any actual drilling measurement point on the curved section of the well trajectory and the corresponding theoretical reference point in the global planning trajectory is ≤ the allowable error radius of the corresponding target point. (3) Obstacle avoidance distance ≥ collision avoidance safety distance .

4. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, This also includes setting the distance of the attraction point at the starting point of the curved section of the wellbore trajectory. and the distance to the attraction point at the end of the curved section of the wellbore trajectory The range of values ​​is used to model and optimize the curved section of the wellbore trajectory with the goal of minimizing the arc length of the cubic Bézier curve.

5. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, The confidence region is an uncertain ellipse; its semi-axis length and direction are determined by eigenvalue decomposition of the covariance matrix of the uncertainty of the current measurement point position. The semi-axis length is the product of the square root of the eigenvalue and the confidence factor, and the direction is the same as the direction of the eigenvector.

6. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, The specific process for determining whether a significant shift has occurred is as follows: project the shift vector onto the plane where the confidence region is located to obtain a plane shift vector, calculate the magnitude of the plane shift vector, and calculate the semi-axis length of the confidence region in the direction of the plane shift vector. Significance offset is determined by comparing the magnitude of the modulus and the semi-shaft length.

7. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, The specific process of generating the correction path is as follows: on the global planning trajectory, find the target point located in the actual direction of the wellbore and in front of the current actual drilling position; Starting from the current actual drilling position and ending at the target point, a cubic Bézier curve is constructed as a candidate correction path; the candidate correction path is verified by engineering constraints; if the verification passes, the candidate correction path becomes the final correction path.

8. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 7, characterized in that, It also includes, if the verification fails, determining a new target point, regenerating a new candidate correction path and verifying it again, until a final correction path that meets the engineering constraint verification is obtained.

9. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, The tangential direction at the start and end points of the candidate correction path / new candidate correction path is consistent with the actual wellbore direction and the global planned trajectory direction.

10. The automatic drilling trajectory control method based on Bézier curves and uncertain ellipses according to claim 1, characterized in that, The cubic Bézier curve model Specifically: In the formula: This represents the positional proportion from the starting point to the ending point in the cubic Bézier curve formula; This marks the starting point of the curved section of the wellbore trajectory; This is the first control point to be optimized; This is the second control point to be optimized; This marks the end point of the curved section of the wellbore trajectory; for The unit tangent vector at that point; for The unit tangent vector at that point; The distance from the attraction point at the starting point of the curved section of the wellbore trajectory; This is the distance to the attraction point at the end of the curved section of the wellbore trajectory.

Images