While-drilling attitude measurement method for processing measurement bias of sensor

Abstract

The invention discloses a while-drilling attitude measurement method for processing sensor measurement bias. The method comprises the following steps: S1, establishing an evolution relation of a sensor measurement bias vector theta (k) along with time; s2, establishing a dynamic while-drilling attitude state model based on an attitude angle and sensor bias; s3, establishing a measurement model based on the drilling tool data and the sensor measurement offset vector theta (k); and S4, based on the dynamic while-drilling attitude state model and the measurement model, obtaining the optimal state estimation at each moment, and calculating the optimal attitude parameters of the drilling tool. According to the invention, the problem that the attitude parameter measurement is inaccurate when the sensor while drilling is used for measuring the offset is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of drilling measurement, and particularly to a method for measuring the attitude while drilling for processing sensor measurement biases. Background Art

[0002] Drilling operations are an essential part of oil and gas exploration. Obtaining accurate and reliable drill string attitude parameters (i.e., well inclination angle, azimuth angle, and tool face angle) plays an important role in improving drilling efficiency and quality. The signals required for traditional drill string attitude measurement mainly rely on a variety of near-bit sensors, including acceleration sensors, magnetometers, and gyroscopes. However, during drilling, factors such as high temperature, high pressure, strong vibration, and strong magnetic field underground have varying degrees of influence on a variety of near-bit sensors, making the signals output by the sensors carry a large amount of random interference noise and systematic biases, resulting in inaccurate measurement of the attitude parameters of the drilling tool.

[0003] The Chinese patent document with the publication number CN108387205A discloses a measurement method for a drill string attitude measurement system based on multi-sensor data fusion, including the installation methods of various near-bit sensors, the data fusion processing method, and the attitude calculation method. However, it cannot process sensor signals with biases, and there are singularities in the attitude calculation method.

[0004] Most of the attitude parameter measurements of domestic and foreign steering drilling tools when dealing with sensor measurement biases adopt the method of separately calibrating or filtering the signals output by each sensor for attitude calculation. This method has high requirements for sensor accuracy, the process of separately calibrating and filtering the sensor output signals is complex, sacrificing cost and drilling efficiency, and it cannot process the biases and random noises in the measurement simultaneously, resulting in inaccurate or even unmeasurable attitude measurement.

[0005] In response to this, the present invention provides a method for estimating the attitude while drilling for processing sensor measurement biases. Summary of the Invention

[0006] Based on the technical problems existing in the background art, the present invention proposes a method for measuring the attitude while drilling for processing sensor measurement biases, which solves the problem of inaccurate measurement of attitude parameters when the sensor measurement while drilling has biases.

[0007] The method for measuring the attitude while drilling for processing sensor measurement biases proposed by the present invention has the following steps:

[0008] S1: Establish the evolution relationship of the sensor measurement bias vector θ(k) over time;

[0009] S2: Establish a dynamic attitude state model while drilling based on the attitude angle and sensor bias

[0010]

[0011] Wherein, s0(k), s1(k), s2(k), and s3(k) are quaternions; x(k) is the state quantity to be estimated at time k; H ω (k) is the state transition matrix corresponding to the quaternion; H θ (k) is the state transition matrix corresponding to the sensor bias; U(k) is the state equation noise.

[0012] S3: Establish a measurement model based on the drill string data and the sensor measurement bias vector θ(k);

[0013] S4: Based on the dynamic attitude state model while drilling and the measurement model, obtain the optimal state estimation at each moment, and solve the optimal attitude parameters of the drill string.

[0014] Preferably, the measurement model established in S3 is as follows:

[0015]

[0016] Wherein, Z(k) is the sensor measurement vector collected at time k; g(k) is the ideal sensor measurement vector at time k; V(k) is the measurement noise; H x (k), H y (k), H z (k) are the ideal triaxial magnetometer measurements; G x (k), G y (k), G z (k) are the ideal triaxial accelerometer measurements; C is the local magnetic intensity; δ is the local magnetic dip; g is the local acceleration due to gravity.

[0017] Preferably, the method steps for obtaining the optimal attitude parameters in S4 are as follows:

[0018] S41: Initialize the cubature Kalman filter algorithm using the maximum likelihood estimation method, and obtain the optimal state estimation at each moment according to the cubature Kalman filter algorithm

[0019] S42: Based on the initial state and the covariance matrix P0, for each moment k, obtain the state prediction at time k + 1 according to the dynamic attitude state model while drilling and the predicted covariance matrix P k+1|k , and calculate the predicted measurement according to the state prediction at time k + 1

[0020] S43: Based on the state prediction the predicted measurement and the actual output measurement Z of the drill string sensor k+1 ​, obtain the optimal state estimate at time k+1 and calculate the corresponding error covariance matrix P k+1 ;

[0021] S44: Resolve the quaternion components in the optimal state estimate into attitude angles to obtain the optimal attitude parameters of the drill string.

[0022] Preferably, the initialization method in S41 is: Based on the initial measurements of the sensors of the drill string and the condition that the noise is Gaussian independent, establish a maximum likelihood estimation problem, solve the maximum likelihood estimation problem through the interior point method, and use the obtained solution as the initialization of the cubature Kalman filter algorithm.

[0023] Preferably, the drill string data includes the output data of the accelerometer and the magnetometer; the sensor measurement bias includes the accelerometer bias and the magnetometer bias.

[0024] Advantageous technical effects of the present invention:

[0025] By taking the bias as an estimated quantity and forming a state quantity together with the quaternion, and adopting the cubature Kalman filter algorithm, the present invention filters out the noise interference of the state equation and the measurement equation, obtains the optimal estimation of the quaternion related to the attitude angle, and thus resolves the optimal attitude parameters; compared with the existing measurement methods for attitude while drilling, the method of the present invention can solve the problem that the attitude angles of the drill string cannot be accurately measured due to the bias in the measurements of various near-bit attitude sensors. Description of the Drawings

[0026] Figure 1 is a flowchart of the measurement method for attitude while drilling for processing sensor measurement bias proposed by the present invention;

[0027] Figure 2 is a simulation result diagram of the measurement method for attitude while drilling for processing sensor measurement bias proposed by the present invention;

[0028] Figure 3 is a schematic diagram of the measurement system for attitude while drilling for processing sensor measurement bias proposed by the present invention;

[0029] Figure 4 is an application flowchart of the measurement system for attitude while drilling for processing sensor measurement bias proposed by the present invention.

[0030] In the figure: 1 - drill string, 2 - accelerometer, 3 - magnetometer, 4 - gyroscope. Detailed Embodiments

[0031] The present invention will be further explained below in conjunction with specific embodiments.

[0032] Refer to Figure 3, which is the measurement-while-drilling (MWD) attitude measurement system used in the embodiments of the present invention. The measurement tool includes a drill string 1, a three-axis accelerometer 2, a three-axis magnetometer 3, and a three-axis gyroscope 4 installed orthogonally.

[0033] It should be understood that the harsh environments of high temperature, high pressure, strong shock vibration, and electromagnetic interference underground have different effects on different types of sensors. Using a single near-bit sensor results in a large measurement error and does not meet the existing MWD attitude measurement indicators. Therefore, a measurement system using multiple sensors is used.

[0034] Multiple near-bit sensors on the drill string are installed according to the front (X)-right (Y)-down (Z) coordinate system, and with reference to the north-east-down (O-NED) geographical coordinate system, the rotation direction follows the right-hand rule. That is, the data measured by the accelerometer and magnetometer on the drill string can be transformed into the geographical coordinate system through a rotation matrix. Specifically:

[0035]

[0036] In the formula, is the rotation matrix; C is the local magnetic intensity; δ is the local magnetic dip angle; g is the local acceleration due to gravity.

[0037] Referring to Figure 1 , the MWD attitude measurement method for processing sensor measurement bias proposed by the present invention is as follows:

[0038] Step 1: Establish an MWD attitude measurement system, including establishing a front (X)-right (Y)-down (Z) orthogonal drill string coordinate system with reference to the north-east-down (O-NED) geographical coordinate system, and installing three groups of different near-bit sensors, including an accelerometer, a magnetometer, and a gyroscope. Specifically, see Figure 3 .

[0039] Step 2: Based on the rotational speed measurement of the gyroscope, establish a dynamic MWD attitude state model for the evolution of the drill string attitude over time. The state quantity x(k) to be estimated at time k consists of the attitude angle and the sensor bias. Specifically:

[0040] Based on the sensor measurement bias vector θ(k) of the accelerometer or magnetometer sensor, establish the evolution relationship with time:

[0041] θ(k + 1) = H θ (k)θ(k) (4)

[0042] In the formula, H θ (k) is the state transition matrix of the sensor bias at time k.

[0043] The sensor measurement bias vector θ(k) can be a fixed systematic error, a linearly varying systematic error, or a complexly varying systematic error, that is, Hθ (k) = 1, H θ (k) = C or H θ (k) is the complex function F(·).

[0044] Based on the above-established dynamic attitude state model while drilling, it is as follows:

[0045]

[0046] In the formula, s0(k), s1(k), s2(k), and s3(k) are quaternions at time k; H ω (k) is the state transition matrix corresponding to the quaternion; U(k) is the state equation noise.

[0047] Among them, is the state transition matrix, including the three-axis angular velocities ω x (k), ω y (k), ω z (k) measured by the gyroscope, and the sensor bias H θ (k). Let be the discretized time unit, then there is:

[0048]

[0049] In the formula, t s is the discretized time unit.

[0050] Step 3: Construct the measurement equation according to the output data of the three-axis accelerometer and the three-axis magnetometer:

[0051]

[0052]

[0053] In the formula, Z(k) is all the measurement vectors collected by the sensor at time k; g(k) is the ideal measurement vector; V(k) is the measurement noise; H x (k), H[[ID= fifty - two]] y (k), H z (k) are the ideal three - axis magnetometer data; G x (k), G y (k), G z (k) are the ideal three - axis accelerometer data.

[0054] Step 4: Use the maximum likelihood estimation method to initialize the cubature Kalman filter algorithm, and obtain the optimal state estimate at each moment according to the cubature Kalman filter algorithm Based on the initial state and the covariance matrix P0, for each moment k, according to the dynamic attitude state model while drilling Convert it into 2n (n is the dimension of the state variable) volume points Propagate to obtain the state prediction at time k+1 And the predicted covariance matrix P k+1|k And based on the state prediction at time k+1 Calculate the predicted measurement Specifically as follows:

[0055]

[0056] In the formula, S k Is obtained by performing Cholesky decomposition on the covariance matrix P k ; ξ i Is the i-th unit volume point vector; Q k Is the process noise covariance matrix; H ω Is the state transition matrix; F is the observation matrix.

[0057] Step 5: Based on the state prediction The predicted measurement And the actual output measurement Z of the drill tool sensor k+1 , Obtain the optimal state estimate at time k+1 And calculate the corresponding error covariance matrix P k+1 , Specifically as follows:

[0058]

[0059] In the formula, K k+1 Is the weight coefficient, Is the measurement error covariance matrix; Is the cross-covariance matrix; R k Is the measurement noise covariance matrix.

[0060] Step 6: For the optimal state estimate at each time k in Step 5 Use its quaternion components to perform attitude calculation, and obtain the three attitude angle estimates at time k as:

[0061]

[0062] The present invention forms the state variable together with the quaternion by taking the bias as the quantity to be estimated, and adopts the cubature Kalman filter algorithm to filter out the noise interference of the state equation and the measurement equation, and obtains the optimal estimate of the quaternion related to the attitude angle, so as to calculate the optimal attitude parameters; compared with the existing measurement methods of attitude while drilling, the method of the present invention can solve the problem that the measurement of the attitude angle of the drill tool is inaccurate due to the bias of the measurement of various near-bit attitude sensors.

[0063] Embodiment

[0064] The application flowchart of the method for measuring the attitude while drilling for processing sensor measurement bias is referred to Figure 3 .

[0065] Figure 4 is the mean square error diagram of the drill string attitude angle obtained by the method of the present invention under 200 Monte Carlo experiments, where the standard deviation of the process noise U(k) is fixed at 0.01 rad, and the standard deviation of the observation noise V(k) is fixed at 0.01 m / s 2 , the initial attitude angles of the drill string are set to I0 = 45°, A0 = 30° and T0 = 60°, where the acceleration measurement along the drill pipe is g x with a bias θ = 1 m / s 2 , and the rotational speeds of the three axes of the gyroscope are ω x (k) = 2π rad / s, ω y (k) = 0 rad / s, ω z (k) = 0 rad / s, and the discretization time unit t of the state equation s = 1 s.

[0066]

[0067] It can be seen that the estimation results of the three attitude angles converge rapidly, and the final mean square errors do not exceed 0.2°, reaching a relatively high accuracy.

[0068] The following is the data and algorithm implementation process of a Monte Carlo experiment corresponding to Figure 4 steps 50 to 51:

[0069] According to the filtering result of step 50

[0070] obtain the state prediction of step 51 in step 4 and the measurement prediction According to the actual measurement Z of step 51 51 = [8.0005, -3.3375, 5.9918, 49.2554, -18.6964, 0.2487] T , and combine with step 5 to obtain the state estimation of step 51

[0071]

[0072] Substitute into formulas (17), (18) and (19) in step 6 to obtain the optimal attitude parameters of the drill string in step 51:

[0073] I 51 = arcsin[0.8207×0.4365 - 0.3680×0.0243] = 44.3094°

[0074]

[0075] It should be understood that in the field of geological exploration, especially in the process of oil and gas exploration, the measurement-while-drilling (MWD) technology is mostly adopted, that is, a measurement-while-drilling device is installed behind the drill bit to timely measure the geological parameters of the formation around the drill bit, identify complex oil and gas layers with industrial exploitation value, and make the drill bit stop in the required reservoir in time; at the same time, monitor the position of the drill bit, the drilling trajectory, the well inclination, the well diameter and the vibration parameters, etc., so as to realize real-time positioning and timely deviation correction. In the prior art, when the signal of the drill string attitude sensor has an offset, it will cause inaccurate measurement of the attitude parameters, resulting in the inconsistency between the drilling trajectory and the target trajectory, and affecting the oil and gas exploitation process. However, a drill string attitude measurement method for processing the offset of the attitude sensor measurement in the embodiment of the present disclosure can solve such problems.

[0076] Although the embodiments of the present application have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present application. The scope of the present application is defined by the appended claims and their equivalents, and all of them should be included within the protection scope of the present application.

Claims

1. A measurement-while-drilling attitude measurement method for processing sensor measurement bias, characterized in that, The steps are as follows: S1: Establish the evolution relationship of the sensor measurement bias vector θ(k) over time; S2: Establish a dynamic while-drilling attitude state model based on attitude angle and sensor bias where \(s_0(k)\), \(s_1(k)\), \(s_2(k)\) and \(s_3(k)\) are quaternions; \(x(k)\) is the state quantity to be estimated at time \(k\); \(H\) ω (k) is the state transition matrix corresponding to the quaternion; \(H\) θ (k) is the state transition matrix corresponding to the sensor bias; \(U(k)\) is the state equation noise. S3: Establishing a measurement model based on the drilling tool data and the sensor measurement bias vector θ(k); S4: Based on the dynamic while-drilling posture state model and measurement model, the optimal state estimation at each moment is obtained, and the optimal posture parameters of the drilling tool are calculated.

2. The measurement-while-drilling attitude measurement method for processing sensor measurement bias according to claim 1, wherein The measurement model established in S3 is as follows: Where, Z(k) is all sensor measurements collected at time k; g(k) is the ideal sensor measurement vector at time k; V(k) is the measurement noise; H x (k), H y (k), H z (k) are the ideal triaxial magnetometer measurements; G x (k), G y (k), G z (k) are the ideal triaxial accelerometer measurements; C is the local magnetic intensity; δ is the local magnetic dip; g is the local acceleration due to gravity.

3. The measurement-while-drilling attitude measurement method for processing sensor measurement bias according to claim 1, wherein The steps for obtaining the optimal posture parameters in S4 are as follows: S41: Initialize the cubature Kalman filter algorithm using the maximum likelihood estimation method, and obtain the optimal state estimation at each moment according to the cubature Kalman filter algorithm S42: In the initialization state Based on the covariance matrix P0, for each moment k, obtain the state prediction at the moment k+1 according to the dynamic attitude state model while drilling and the predicted covariance matrix P k+1|k , and calculate the predicted measurement based on the state prediction at the moment k+1 ​ S43: Based on state prediction Predictive Measurement and the actual output measurement of the sensor of the drilling tool Z k+1 , get the optimal state estimate at time k+1 And calculate the corresponding error covariance matrix P k+1 ; S44: Calculate the quaternion components in the optimal state estimate as attitude angles to obtain the optimal attitude parameters of the drill string. ​ 4. The method for measuring the attitude while drilling for processing the sensor measurement bias according to claim 3, wherein, The initialization method in S41 is: based on the initial measurement of the drilling tool sensor and the condition that the noise is Gaussian independent, a maximum likelihood estimation problem is established, the maximum likelihood estimation problem is solved by the interior point method, and the obtained solution is used as the initialization of the cubature Kalman filter algorithm.

5. The method for measuring the attitude while drilling for processing the bias of the sensor measurement according to claim 1, wherein The drilling tool data includes output data of the accelerometer and the magnetometer; the sensor measurement bias includes the accelerometer bias and the magnetometer bias.

Images