A clustering trajectory correction method for directional drilling based on acceleration sensing

Abstract

The present invention discloses a directional drilling cluster trajectory correction method based on acceleration sensing. First, the acceleration signal data of each measuring point during the drilling process is obtained through an acceleration sensor; then, the signal data is denoised using a variational mode decomposition method to obtain the denoised acceleration signal data; then, the acceleration signal data is calculated to obtain the azimuth angle, inclination angle and drilling depth of the borehole when passing through each measuring point during the drilling process; the actual centroid vector position is generated in real time using the above-obtained data, and the actual centroid vector position is compared and analyzed with the initial centroid vector position of the preset trajectory using a K-means optimization algorithm, thereby adjusting the direction of the drill bit during the drilling process and realizing the correction of the trajectory during the drilling process. Through the above method, the present invention can monitor the drilling trajectory in real time and adjust the drilling direction in a timely manner, thereby ensuring that the drilling process conforms to the preset trajectory and meeting the needs of subsequently deploying other equipment in the borehole.

Description

Technical Field

[0001] The present invention relates to the field of drilling efficiency and safety of drilling rigs in the drilling field, and in particular to a directional drilling cluster trajectory correction method based on acceleration sensing. Background Art

[0002] The drilling process primarily consists of three phases: acceleration, uniform motion, and deceleration. Rotating the drill rod during drilling, as well as advancing and retreating after adding the drill rod, can cause vibrations and, consequently, deviations in the drilling trajectory. Furthermore, the drill bit and the drill rod tend to tilt toward softer lithology during drilling. A significant deviation in the drilling trajectory can lead to increased operation time, increased drilling risk, and deviation from the intended target. This can also result in poor conditions for subsequent drilling operations using equipment such as transient electromagnetic drilling.

[0003] Existing technologies primarily rely on stabilizers or horizontal directional drilling (HDD), but the installation cost of a guide drill bit for HDD is high. The stabilizer principle involves attaching three to five stabilizers (also known as centralizers) approximately the same diameter as the drill bit to the drill pipe within 10 to 20 meters of the drill bit. By adjusting the position and number of stabilizers and matching drilling parameters, the drill pipe exerts a force to control the drill bit, moving it upward, downward, or maintaining its straightness. However, this technology only supports short-distance drilling operations and fails to maintain effective drill pipe sinking for long-distance drilling. Furthermore, if the stabilizers deviate, subsequent drilling direction correction to the preset trajectory is difficult.

[0004] In order to solve the problem of drilling trajectory deviation that is prone to occur during drilling, how to provide a new method that can monitor the drilling trajectory in real time and adjust the drilling direction to ensure that the drilling process conforms to the preset trajectory is the research direction that this industry needs to study. Summary of the Invention

[0005] In response to the problems existing in the above-mentioned prior art, the present invention provides a directional drilling cluster trajectory correction method based on acceleration sensing, which can monitor the drilling trajectory in real time and adjust the drilling direction, thereby ensuring that the drilling process conforms to the preset trajectory.

[0006] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is: a directional drilling cluster trajectory correction method based on acceleration sensing, the specific steps are as follows:

[0007] Step 1: Install the acceleration sensor in the drill rod near the drill bit, pre-set the required drilling trajectory, and set multiple measurement points at different drilling depths. After completion, use the drill bit to start drilling according to the set trajectory. Each time a measurement point is reached at a drilling depth, the acceleration signal data is obtained through the acceleration sensor.

[0008] Step 2: The acceleration signal data obtained in step 1 is first analyzed by autocorrelation function to determine whether the collected signal data contains noise. This judgment process is a well-known technology and noise is generally present in the drilling process. If it is determined that noise exists, the signal data is denoised using variational mode decomposition to obtain denoised acceleration signal data; if not, directly proceed to step 3;

[0009] Step 3: Using the acceleration signal data obtained in step 2, the azimuth angle, inclination angle and drilling depth of the borehole when passing through each measuring point during the drilling process are calculated;

[0010] Step 4: Generate the actual center of mass vector position in real time based on the data obtained in step 3, and use the K-means Optimizer (KO) algorithm to compare and analyze the actual center of mass vector position with the initial center of mass vector position of the preset trajectory, so as to adjust the direction of the drill bit during drilling and realize the correction of the trajectory during drilling.

[0011] Furthermore, the step 2 uses variational mode decomposition to denoise the signal data. The specific process is as follows:

[0012] Perform variational mode decomposition on the acceleration signal data:

[0013]

[0014] Where: {u k}={u1,…,u K} represents the kth IMF component, {ω k}={ω1,…,ω K} represents the center frequency of the kth IMF component, f(t) represents the original acceleration signal data;

[0015] In order to obtain the optimal solution of the constrained variational problem, the Lagrange multiplication operator λ(t) and the quadratic penalty factor α are used to transform the constrained variational problem into an unconstrained variational problem. The Lagrange multiplication operator λ(t) can ensure the strictness of the constraints, and the quadratic penalty factor α can ensure the accuracy of signal reconstruction in a Gaussian noise environment. The expanded Lagrange expression is:

[0016]

[0017] The alternating direction method of multipliers (ADMM) is used to continuously update the IMF components and their center frequencies to solve the saddle points in the Lagrange expression, and u is updated through a cyclic process. kn+1 、ω k n+1 ,λ k n+1 The update formula is:

[0018]

[0019] Where: ^ represents Fourier transform; τ represents the time step;

[0020] Initialize {w k 1}、{u k 1}、λ 1 , n, iteratively update u k , ε, λ, until the allowable error ε is met, the decomposition stops, and finally K IMF components are output. The allowable error discriminant is as follows:

[0021]

[0022] The K IMF components finally output from the stripped data include 1 denoised acceleration signal data and K-1 noise components. After deleting the noise components, the acceleration signal data under noise-free conditions is obtained to obtain more accurate inclination data, providing more accurate original data for the next step of intelligent clustering correction of drilling trajectory.

[0023] Furthermore, in step 3, the inclination angle of the borehole at each measuring point is obtained. The specific process is as follows: the inclination angle is determined by measuring the projection of the acceleration signal data on the X-axis of the acceleration sensor according to the denoised acceleration signal data. If the acceleration value measured on the X-axis is ax, the inclination angle a is calculated by the following formula:

[0024] a=arcsin(ax / g)

[0025] where g is the acceleration due to gravity.

[0026] Furthermore, the specific process of the K-means optimization algorithm in step 4 is as follows:

[0027] The formula for setting the Euclidean distance weight is as follows:

[0028]

[0029] The K-means optimization algorithm is used to establish the centroid vector of the clustering area, specifically:

[0030] 1. The random initialization formula is described as follows:

[0031] X i=lb+(ub-lb)·rand(0,1,[1,D]),i=1,2,…,N

[0032] Where: Xi represents the ith candidate solution, lb and ub represent the lower and upper bounds of the search space respectively; rand is a random number uniformly distributed in the range [0,1]; N represents the number of solutions;

[0033] 2. Update process of centroid vector position

[0034] The K-means optimization algorithm uses a linear population size reduction method. As the number of iterations increases, the population size in the data will decrease in a linear function.

[0035]

[0036] where t is the current iteration, T max is the maximum number of iterations; therefore, the population size is N at the first iteration t = 1 and T at the last iteration max The overall size is 4;

[0037] 3. Optimization strategy

[0038] In order to diversify the search space of the K-means optimization algorithm, in addition to utilizing three centroid vector positions in each iteration, the K-means optimization algorithm also adds a new search area called the average centroid vector position As shown in the following formula:

[0039]

[0040] The Euclidean distance between the actual center of mass vector position during drilling and the initial center of mass vector position in the preset trajectory can be more accurately obtained by averaging the center of mass vector position. The initial center of mass position is determined as α and the average center of mass vector position is determined as β according to the following formula. If the drill bit drills along the preset trajectory during drilling, d(α, β) approaches 0; if d(α, β) continues to increase with drilling time, it means that the drilling direction has deviated from the trajectory. At this time, the drilling direction of the drill bit is adjusted until d(α, β) approaches 0, thereby keeping the subsequent drilling trajectory consistent with the preset trajectory. The specific formula is:

[0041]

[0042] In the above formula, d(α,β) is the Euclidean distance between any drilling time α and any drilling time β, and x α 、y α 、z α are the x, y, and z coordinates of the α event, respectively. β 、y β 、zβ are the x, y, and z coordinates of the β event, respectively.

[0043] Compared with the prior art, the present invention first obtains the acceleration signal data of each measuring point of the drilling depth through an acceleration sensor; then uses the variational mode decomposition method to denoise the signal data to obtain the denoised acceleration signal data; then the denoised acceleration signal data is calculated to obtain the azimuth angle, inclination angle and drilling depth of the borehole when passing through each measuring point during the drilling process; the actual centroid vector position is generated in real time using the above-mentioned data, and the actual centroid vector position is compared and analyzed with the initial centroid vector position of the preset trajectory using the K-means optimization algorithm, thereby adjusting the direction of the drill bit during the drilling process and realizing the correction of the trajectory during the drilling process. Through the above method, the present invention can monitor the drilling trajectory in real time and adjust the drilling direction in time, thereby ensuring that the drilling process conforms to the preset trajectory and meeting the needs of subsequently deploying other equipment in the borehole. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the overall process of the present invention;

[0045] Figure 2 It is a schematic diagram of the installation of the acceleration sensor in the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below.

[0047] like Figure 1 As shown, the specific steps of the present invention are:

[0048] Step 1: Install the acceleration sensor in the drill rod near the drill bit. Figure 2 As shown, the required drilling trajectory is pre-set, and multiple measuring points are set at different drilling depths. After completion, the drill bit starts drilling according to the set trajectory, and each time a measuring point of a drilling depth is reached, the acceleration signal data is obtained through the acceleration sensor;

[0049] Step 2: The acceleration signal data obtained in step 1 is first analyzed using an autocorrelation function to determine whether the collected signal data contains noise. This determination process is a well-known technique. If noise is determined to be present, the signal data is denoised using variational mode decomposition to obtain denoised acceleration signal data. If noise is not present, the process proceeds directly to step 3. The specific denoising process is as follows:

[0050] Perform variational mode decomposition on the acceleration signal data:

[0051]

[0052] Where: {uk}={u1,…,u K} represents the kth IMF component, {ω k}={ω1,…,ω K} represents the center frequency of the kth IMF component, f(t) represents the original acceleration signal data;

[0053] In order to obtain the optimal solution of the constrained variational problem, the Lagrange multiplication operator λ(t) and the quadratic penalty factor α are used to transform the constrained variational problem into an unconstrained variational problem. The Lagrange multiplication operator λ(t) can ensure the strictness of the constraints, and the quadratic penalty factor α can ensure the accuracy of signal reconstruction in a Gaussian noise environment. The expanded Lagrange expression is:

[0054]

[0055] The alternating direction method of multipliers (ADMM) is used to continuously update the IMF components and their center frequencies to solve the saddle points in the Lagrange expression, and u is updated through a cyclic process. k n+1 、ω k n+1 ,λ k n+1 The update formula is:

[0056]

[0057] Where: ^ represents Fourier transform; τ represents the time step;

[0058] Initialize {w k 1}、{u k 1}、λ 1 , n, iteratively update u k , ε, λ, until the allowable error ε is met, the decomposition stops, and finally K IMF components are output. The allowable error discriminant is as follows:

[0059]

[0060] The K IMF components finally output from the stripped data include 1 denoised acceleration signal data and K-1 noise components. After deleting the noise components, the acceleration signal data under noise-free conditions is obtained to obtain more accurate inclination data, providing more accurate original data for the next step of intelligent clustering correction of drilling trajectory.

[0061] Step 3: Using the acceleration signal data obtained in step 2, the azimuth, inclination, and depth of the borehole at each measurement point during the drilling process are calculated. The inclination angle of the borehole at each measurement point is obtained by measuring the projection of the acceleration signal data on the X-axis of the acceleration sensor based on the de-noised acceleration signal data to determine the inclination angle. If the acceleration value measured on the X-axis is ax, the inclination angle a is calculated using the following formula:

[0062] a=arcsin(ax / g)

[0063] where g is the acceleration due to gravity.

[0064] Step 4: Based on the data obtained in step 3, the actual centroid vector position is generated in real time, and the K-means Optimizer (KO) algorithm is used to compare and analyze the actual centroid vector position with the initial centroid vector position of the preset trajectory, thereby adjusting the direction of the drill bit during the drilling process and realizing the correction of the trajectory during the drilling process. The specific process is as follows:

[0065] The formula for setting the Euclidean distance weight is as follows:

[0066]

[0067] The K-means optimization algorithm is used to establish the centroid vector of the clustering area, specifically:

[0068] 1. The random initialization formula is described as follows:

[0069] X i =lb+(ub-lb)·rand(0,1,[1,D]),i=1,2,…,N

[0070] Where: Xi represents the ith candidate solution, lb and ub represent the lower and upper bounds of the search space respectively; rand is a random number uniformly distributed in the range [0,1]; N represents the number of solutions;

[0071] 2. Update process of centroid vector position

[0072] The K-means optimization algorithm uses a linear population size reduction method. As the number of iterations increases, the population size in the data will decrease in a linear function.

[0073]

[0074] where t is the current iteration, T max is the maximum number of iterations; therefore, the population size is N at the first iteration t = 1 and T at the last iteration max The overall size is 4;

[0075] 3. Optimization strategy

[0076] In order to diversify the search space of the K-means optimization algorithm, in addition to utilizing three centroid vector positions in each iteration, the K-means optimization algorithm also adds a new search area called the average centroid vector position As shown in the following formula:

[0077]

[0078] The Euclidean distance between the actual center of mass vector position during drilling and the initial center of mass vector position in the preset trajectory can be more accurately obtained by averaging the center of mass vector position. The initial center of mass position is determined as α and the average center of mass vector position is determined as β according to the following formula. If the drill bit drills along the preset trajectory during drilling, d(α, β) approaches 0; if d(α, β) continues to increase with drilling time, it means that the drilling direction has deviated from the trajectory. At this time, the drilling direction of the drill bit is adjusted until d(α, β) approaches 0, thereby keeping the subsequent drilling trajectory consistent with the preset trajectory. The specific formula is:

[0079]

[0080] In the above formula, d(α,β) is the Euclidean distance between any drilling time α and any drilling time β, and x α 、y α 、z α are the x, y, and z coordinates of the α event, respectively. β 、y β 、z β are the x, y, and z coordinates of the β event, respectively.

[0081] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A directional drilling cluster trajectory correction method based on acceleration sensor, characterized in that: The specific steps are: Step 1: Install the acceleration sensor in the drill rod near the drill bit, pre-set the required drilling trajectory, and set multiple measurement points at different drilling depths. After completion, use the drill bit to start drilling according to the set trajectory. Each time a measurement point is reached at a drilling depth, the acceleration signal data is obtained through the acceleration sensor. Step 2: The acceleration signal data obtained in step 1 is first analyzed by autocorrelation function to determine whether the collected signal data contains noise. If it is determined that noise exists, the signal data is denoised using variational mode decomposition to obtain denoised acceleration signal data; if not, the process proceeds directly to step 3; Step 3: Using the acceleration signal data obtained in step 2, the azimuth angle, inclination angle and drilling depth of the borehole when passing through each measuring point during the drilling process are calculated; Step 4: Generate the actual centroid position in real time based on the data obtained in step 3, and use the K-means optimization algorithm to compare and analyze the actual centroid position with the initial centroid position of the preset trajectory. The specific process is as follows: The formula for setting the Euclidean distance weight is as follows: ； The K-means optimization algorithm is used to establish the centroid vector of the clustering area, specifically:

1. The random initialization formula is described as follows: ； in: X i represents the ith candidate solution, lb and ub They represent the lower and upper bounds of the search space respectively; rand is a random number uniformly distributed in the range [0,1]; N represents the number of solutions; 2. Update process of centroid vector position: The K-means optimization algorithm uses a linear population size reduction method. As the number of iterations increases, the population size in the data will decrease in a linear function. ； where t is the current iteration, T max is the maximum number of iterations; thus, the population size is N at the first iteration t = 1 and T at the last iteration max The overall size is 4; 3. Optimization strategy: In order to diversify the search space of the K-means optimization algorithm, in addition to utilizing three centroid vector positions in each iteration, the K-means optimization algorithm also adds a new search area called the average centroid vector position , as shown below: ； The Euclidean distance between the actual center of mass vector position during drilling and the initial center of mass vector position in the preset trajectory can be more accurately obtained by averaging the center of mass vector position. The initial center of mass position is determined as α and the average center of mass vector position is determined as β according to the following formula. If the drill bit drills along the preset trajectory during drilling, d (α, β) approaches 0; if d (α, β) continues to increase with drilling time, it means that the drilling direction has deviated from the trajectory. At this time, the drilling direction of the drill bit is adjusted until d (α, β) approaches 0, thereby keeping the subsequent drilling trajectory consistent with the preset trajectory and realizing the correction of the trajectory during drilling. The specific formula is: ； In the above formula, d (α, β) is the Euclidean distance between any drilling time α and any drilling time β, and x α 、y α 、z α are the x, y, and z coordinates of the α event, respectively. β 、y β 、z β are the x, y, and z coordinates of the β event, respectively.

2. The directional drilling cluster trajectory correction method based on acceleration sensor according to claim 1, characterized in that: The second step uses variational mode decomposition to denoise the signal data. The specific process is as follows: Perform variational mode decomposition on the acceleration signal data: ； in: represents the kth IMF component, represents the center frequency of the kth IMF component, and f(t) represents the original acceleration signal data; In order to obtain the optimal solution of the constrained variational problem, the Lagrange multiplication operator λ(t) and the quadratic penalty factor α are used to transform the constrained variational problem into an unconstrained variational problem. The expanded Lagrange expression is: ； The alternating direction multiplier method is used to continuously update each IMF component and its central frequency to solve the saddle point in the Lagrange expression, and the saddle point is updated through a cyclic process. 、 、 The update formula is: ； ； ； Where: ^ represents Fourier transform; τ represents the time step; Initialize {w k 1 }、{u k 1 }、λ 1 , n, iteratively update u k , ε, λ, until the allowable error ε is met, the decomposition stops, and finally K IMF components are output. The allowable error discriminant is as follows: ； The K IMF components finally output from the stripped data include 1 denoised acceleration signal data and K-1 noise components. After deleting the noise components, the acceleration signal data under noise-free conditions is obtained.

3. The directional drilling cluster trajectory correction method based on acceleration sensor according to claim 1, characterized in that: In step 3, the inclination angle of the drill hole at each measuring point is obtained. The specific process is as follows: the inclination angle is determined by measuring the projection of the acceleration signal data on the X-axis of the acceleration sensor according to the denoised acceleration signal data. If the acceleration value measured on the X-axis is ax, the inclination angle a is calculated by the following formula: ； where g is the acceleration due to gravity.

Images