Intelligent dynamic inversion method for friction resistance coefficient of drill stem and well wall

Abstract

The invention discloses an intelligent dynamic inversion method for a friction resistance coefficient between a drill stem and a well wall. The method comprises the following steps: S1, identifying a drilling working condition based on real-time data; s2, extracting time sequence data before and after bit feeding based on drilling working conditions; s3, lift-off well bottom idling torque and lift-off well bottom hanging load data are obtained; s4, according to a drill string and well wall friction resistance coefficient inversion method, utilizing idling torque data and no-load hanging load data of the lift-off well bottom and combining a friction resistance torque calculation model to invert a drill string and well wall friction resistance coefficient; in the real-time drilling process, the current working condition is judged once every 10 minutes, if the current working condition is the drilling working condition, the step S3 is repeated to extract the lift-off well bottom idling torque data and the no-load hanging load data, then the step S4 is repeated to invert the drill column and well wall friction resistance coefficient, and the computer dynamically inverts the drill column and well wall friction resistance coefficient; and if not, the circulation is continued until the drilling working condition is reached. According to the method, the friction resistance coefficient of the drill stem and the well wall can be intelligently and dynamically inverted by a computer without manual intervention.

Description

Technical Field

[0001] The invention relates to the technical field of oil and gas drilling, in particular to an intelligent dynamic inversion method for friction coefficient between a drill string and a well wall. Background Art

[0002] The friction coefficient between the drill string and the well wall is a key parameter for the force analysis of the drilling pipe. It can be used to predict the transmission efficiency of drilling pressure torque, feasibility analysis of the pipe running and early warning of drill sticking, so the accurate calculation of the friction coefficient is of great significance. The friction coefficient between the drill string and the well wall is currently mainly set by experience or inverted according to field drilling data. These methods are mainly calculated for a single data point, and the idle torque and no-load hanging weight off the bottom of the well need to be manually calibrated. However, since the friction coefficient between the drill string and the well wall changes dynamically, in order to improve the calculation reliability, the friction coefficient between the drill string and the well wall needs to be dynamically corrected during the drilling process, which requires the staff to frequently calibrate the data. The work is cumbersome, inefficient, and the degree of digitalization and intelligence is low. Therefore, how to realize the computer intelligent dynamic inversion of the drill string and friction coefficient is a key problem that needs to be solved urgently, and it is also a key technology to support future intelligent drilling. Summary of the invention

[0003] In view of the problem that the current calculation method of the friction coefficient between the drill string and the well wall requires workers to frequently calibrate data, which is complicated and labor-intensive, the present invention provides an intelligent dynamic inversion method for the friction coefficient between the drill string and the well wall, which effectively integrates the drilling dynamic data, the drilling physical model and the artificial intelligence algorithm, so that the computer has the ability to autonomously and dynamically invert the friction coefficient between the drill string and the well wall.

[0004] The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall provided by the present invention comprises the following steps:

[0005] S1. Identify drilling conditions based on real-time data;

[0006] The drilling condition identification method specifically comprises: calculating and judging parameters according to real-time drilling data, including the change value of the big hook height, the change value of the well depth, the change value of the drill bit depth, the change value of the hanging weight, and the previous data condition; identifying the drilling condition in combination with the drill bit well depth, well depth, standpipe pressure and wellhead speed data. The drilling conditions include pulling out, drilling down, forward reaming, reverse reaming, single root connection, rotary drilling, sliding drilling, lifting off the bottom of the well for idling, lifting off the bottom of the well for circulation, sitting, circulation and stationary.

[0007] S2, extracting time series data before and after drilling based on drilling conditions;

[0008] Specifically, according to the identification result of the drilling condition, based on the calculation point time, the data of the idling condition after lifting off the bottom of the well, the circulation condition after lifting off the bottom of the well, the rotary drilling condition and the sliding drilling condition before the calculation point time are extracted.

[0009] S3, lifting off the idle torque at the bottom of the well and lifting off the hanging weight at the bottom of the well;

[0010] The data before and after drilling are clustered according to the K-Means algorithm and divided into three categories: one is the hanging weight data and torque data before drilling, one is the hanging weight data and torque data after drilling, and the other is the hanging weight data and torque data with changes during drilling; the average value of the hanging weight data and torque data before drilling is the idling torque data and no-load hanging weight data lifted from the bottom of the well.

[0011] Among them, the K value of the K-Means algorithm is 3. The data before and after drilling can be clustered according to the data characteristics before and after drilling, and the time series of idling torque after lifting off the bottom of the well and the time series of no-load suspended weight after lifting off the bottom of the well can be extracted.

[0012] The data before and after the drill is from time t1 to t n ~t k ~t m Time series of moments, including hanging weight time series and torque time series Each time series is represented as a collection of 3 small time series: and It means from time t1 to time t n moment, the suspended weight time series and torque time series before drilling, and Indicates that from t n+1 Time~t k At the moment, there are variable hanging weight time series and torque time series in the drilling process. and Indicates that from t k+1 Time~t m Moment, suspended weight time series and torque time series after drilling.

[0013] When lifting the bottom hole idle torque data and no-load suspended weight data, it is first necessary to use the time series moving average filter to filter the torque time series. and hanging weight time series Perform outlier removal and smoothing.

[0014] S4. According to the inversion method of the friction coefficient between the drill string and the well wall, the friction coefficient between the drill string and the well wall is inverted by using the idling torque data and the no-load suspended weight data after lifting off the bottom of the well and combining the friction torque calculation model.

[0015] The drill string and well wall friction coefficient inversion method is divided into sliding drilling condition inversion and rotation drilling condition inversion.

[0016] In rotary drilling conditions, first determine the drill tool assembly data, wellbore trajectory data, and drilling fluid density data, set the initial friction coefficient, and then calculate the theoretical idling torque of the drill string through the friction torque model until the theoretical idling torque and the actual idling torque meet the error requirements and stop iteration. The friction coefficient at this time is the friction coefficient between the drill string and the well wall.

[0017] The theoretical idling torque is the friction torque between the drill string unit and the well wall, and the calculation formula is as follows:

[0018] T i =μ ri N i r i

[0019] Where, T i is the friction torque between the drill string unit and the well wall, N·m; μ ri is the circumferential friction coefficient between the drill string unit and the well wall; r i is the drill string unit radius, m; N i is the positive pressure of the drill string unit on the wellbore wall, N.

[0020] In sliding drilling conditions, the theoretical no-load hanging weight and the actual no-load hanging weight are calculated until the theoretical no-load hanging weight and the actual no-load hanging weight meet the error requirements, and the iteration is stopped. The friction coefficient at this time is the friction coefficient between the drill string and the well wall.

[0021] The theoretical no-load hanging weight can be calculated by the following formula:

[0022]

[0023]

[0024] In the formula, F i ,F i-1 is the tension at both ends of the drill string unit, N; μ i is the axial friction coefficient between the drill string unit and the well wall, dimensionless; α i and β i are the average well inclination angle and dogleg angle at both ends of the drill string unit, rad; L i is the drill string unit length, m; Δα i is the inclination increment of the drill string unit, rad; q m is the buoyancy weight of the drill string unit, N / m; N i is the positive pressure of the drill string unit on the wellbore wall, N.

[0025] S5, dynamic inversion friction coefficient;

[0026] During the real-time drilling process, the current working condition is judged every 10 minutes. If it is a drilling condition, step S3 is repeated to extract the idling torque data and no-load suspended weight data after lifting off the bottom of the well, and then step S4 is repeated to invert the friction coefficient between the drill string and the well wall, so as to realize the computer dynamic inversion of the friction coefficient between the drill string and the well wall; if it is not a drilling condition, the cycle is continued until the drilling condition is reached.

[0027] Compared with the prior art, the present invention is beneficial in that:

[0028] (1) The present invention proposes an intelligent dynamic inversion method for the friction coefficient between the drill string and the well wall, which effectively integrates the drilling dynamic data, the drilling physical model and the artificial intelligence algorithm, so that the computer has the ability to autonomously and dynamically invert the friction coefficient between the drill string and the well wall without human intervention, which is of great significance for realizing intelligent drilling.

[0029] (2) The method of the present invention can be directly applied on the basis of existing conventional drilling equipment without the need for additional downhole measurement tools.

[0030] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The present invention is a flow chart of the intelligent dynamic inversion method of the drill string and well wall friction coefficient.

[0032] Figure 2 This is a diagram of the drilling condition identification method.

[0033] Figure 3 It is the time series data before and after the typical drilling in the specific implementation case, where (a) and (b) are the hanging weight and torque diagrams, respectively.

[0034] Figure 4 This is the flow chart of the friction coefficient inversion method.

[0035] Figure 5 This is the dynamic inversion result diagram of the friction coefficient. DETAILED DESCRIPTION

[0036] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0037] like Figure 1-5 As shown, the intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall of the present invention has the following specific steps:

[0038] S1. Determine the real-time drilling conditions based on real-time data identification: The drilling conditions are determined based on the change in the hook height, the change in the well depth, the change in the drill bit depth, the change in the hanging weight, the previous data condition, the drill bit well depth, the well depth, the riser pressure and the wellhead speed. For specific determination methods, refer to Figure 2 .

[0039] S2. According to the identification result of the drilling condition, based on the calculation point time, extract the data of the idling condition after lifting off the bottom of the well, the circulation condition after lifting off the bottom of the well, the rotary drilling condition and the sliding drilling condition before the calculation point time.

[0040] During the real-time drilling process, the current working condition is determined every 10 minutes. If it is a drilling condition, the process proceeds to step S3 and step S4 in sequence to start inversion of the friction coefficient between the drill string and the well wall. If it is not a drilling condition, the process continues to cycle and wait until the drilling condition is reached.

[0041] S3, lifting off the idle torque at the bottom of the well and lifting off the hanging weight at the bottom of the well;

[0042] The data before and after the drill is from time t1 to t n ~t k ~t m Time series of moments, including hanging weight time series and torque time series Each time series can be represented as a collection of 3 small time series: Respectively represent the time from t1 to t n moment, the suspended weight time series and torque time series before drilling, Respectively represent from t n+1 Time~t k At the moment, there are variable hanging weight time series and torque time series in the drilling process; Respectively represent from t k+1 Time~t m The K-Means algorithm is used to analyze the time series of the suspended weight and torque after the drill is delivered. and torque time series Clustering is performed, and the K value of the K-Means algorithm is set to 3. The clustering results are 3 categories, among which the suspension weight time series The clustering results Specifically, it can be expressed as:

[0043]

[0044] Torque time series The clustering results Specifically, it can be expressed as:

[0045]

[0046] In the formula t k The hanging weight value at the moment; t k The torque value at the moment.

[0047] When lifting the idle torque and the suspended weight from the bottom of the well, the torque time series needs to be filtered based on the time series moving average filter. and hanging weight time series Perform outlier removal and smoothing.

[0048] The idle torque from the bottom of the hole is a time series The average value of The time series of the empty suspended weight lifted from the bottom of the well The average value of Figure 3 The following are the time series data before and after the typical drilling in the following implementation case, where (a) and (b) are the hanging weight and torque diagrams, respectively.

[0049] S4. Inverse the friction coefficient between the drill string and the well wall based on the friction torque model;

[0050] like Figure 4 As shown in the figure, the inversion of the friction coefficient between the drill string and the wellbore first determines the drill tool assembly data, wellbore trajectory data, and drilling fluid density data, and sets the initial friction coefficient. If it is a rotary drilling condition, the drill string idling torque is calculated according to the friction torque model until the theoretical idling torque and the actual idling torque meet the error of less than 0.1kN·m, and the iteration is stopped. The friction coefficient at this time is the friction coefficient between the drill string and the wellbore. If it is a sliding drilling condition, the theoretical no-load hanging weight and the actual no-load hanging weight are calculated until the theoretical no-load hanging weight and the actual no-load hanging weight meet the error of less than 1kN, and the iteration is stopped. The friction coefficient at this time is the friction coefficient between the drill string and the wellbore.

[0051] The theoretical no-load suspended weight is calculated using formula (7):

[0052]

[0053] In the formula, F i ,F i-1 is the tension at both ends of the drill string unit, N; μ i is the axial friction coefficient between the drill string unit and the well wall, dimensionless; α i , β i are the average well inclination angle and dogleg angle at both ends of the drill string unit, rad; L i is the drill string unit length, m; Δα i is the inclination increment of the drill string unit, rad; qm is the buoyancy weight of the drill string unit, N / m; N i is the positive pressure of the drill string unit on the wellbore wall, N.

[0054] The theoretical idling torque is calculated using formula (9):

[0055] T i =μ ri N i r i (9)

[0056] Where, T i is the friction torque between the drill string unit and the well wall, N·m; μ ri is the circumferential friction coefficient between the drill string unit and the well wall; r i is the drill string unit radius, m.

[0057] During the real-time drilling process, the current working condition is determined every 10 minutes. If it is a drilling condition, step S3 is repeated to extract the idle torque and no-load suspended weight data of the bottom hole, and step S4 is repeated to invert the friction coefficient, so as to realize the computer dynamic inversion of the friction coefficient between the drill string and the well wall. If it is not a drilling condition, the cycle continues until the drilling condition is reached.

[0058] Specific application cases:

[0059] The simulation data in this implementation case are taken from a horizontal well in a shale gas block in Sichuan. The wellbore trajectory, drilling tool assembly, and wellbore structure data are shown in Tables 1-3. The real-time data including hanging weight, drilling pressure, rotation speed, torque, etc. are shown in Table 4.

[0060] 1. The basic data are as follows:

[0061] (1) Actual drilling wellbore trajectory data, see Table 1.

[0062] Table 1 Wellbore trajectory data

[0063]

[0064]

[0065] (2) Actual drilling tool combination data, see Table 2.

[0066] Table 2 Drilling tool assembly data

[0067]

[0068] (3) Actual drilling wellbore structure data, see Table 3.

[0069] Table 3 Wellbore structure data

[0070]

[0071] (4) Real-time data before and after drilling, see Table 4.

[0072] Table 4 Real-time logging data

[0073]

[0074] 2. Extract idling torque data and no-load hook load data

[0075] Based on the method in step S3, clustering the data in Table 4 can obtain the time series of suspended weight and torque before drilling. Figure 3 Shown:

[0076] Suspended weight time series: H = {1185, 1187, 1193, 1192, ..., 1188, 1181, 1178, 1183}

[0077] Torque time series: T = {10,10,10,11,…,11,11,10,10}

[0078] The idling torque when lifted off the bottom of the well is 10.5 kN·m, and the no-load suspended weight when lifted off the bottom of the well is 1182 kN.

[0079] 3. Inversion of friction coefficient between drill string and well wall

[0080] According to the inversion of the friction coefficient in step S4, when the friction coefficient is 0.17 after cyclic iteration, the theoretical idling torque is 10.4 kN·m, and the no-load hanging weight is 1183 kN, which meets the error requirement. Therefore, the friction coefficient between the drill string and the well wall is 0.17.

[0081] 4. Dynamic inversion of friction coefficient

[0082] Taking the real-time data of the well from August 7, 2022 to September 6, 2022 as an example, the dynamic inversion of the friction coefficient was carried out. The data generation process was simulated by receiving and sending data, and the working conditions were judged in real time to enable the computer to automatically invert the friction coefficient. The inversion results of the friction coefficient are as follows: Figure 5 shown.

[0083] The above description is only a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. An intelligent dynamic inversion method for the friction coefficient between drill string and well wall, characterized in that: The following steps are involved: S1. Identify drilling conditions based on real-time data; S2, extracting time series data before and after drilling based on drilling conditions; S3, lifting off the idle torque at the bottom of the well and lifting off the hanging weight at the bottom of the well; The data before and after drilling are clustered according to the K-Means algorithm, and the data before and after drilling are divided into three categories: one is the suspended weight data and torque data before drilling, one is the suspended weight data and torque data after drilling, and the other is the suspended weight data and torque data with changes during drilling; the average value of the suspended weight data and torque data before drilling is the idling torque data and no-load suspended weight data lifted from the bottom of the well; S4. According to the inversion method of the friction coefficient between the drill string and the well wall, the friction coefficient between the drill string and the well wall is inverted by using the idling torque data and the no-load suspended weight data after lifting off the bottom of the well and combining the friction torque calculation model; S5, dynamic inversion friction coefficient; During the real-time drilling process, the current working condition is judged every 10 minutes. If it is a drilling condition, step S3 is repeated to extract the idling torque data and no-load suspended weight data after lifting off the bottom of the well, and then step S4 is repeated to invert the friction coefficient between the drill string and the well wall, so as to realize the computer dynamic inversion of the friction coefficient between the drill string and the well wall; if it is not a drilling condition, the cycle is continued until the drilling condition is reached.

2. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 1 is characterized in that: Specifically, step S2 is to extract data of idling condition, circulation condition, rotary drilling condition and sliding drilling condition before the calculation point time based on the identification result of the drilling condition and the calculation point time.

3. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 1 is characterized in that: The K value of the K-Means algorithm is 3.

4. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 3 is characterized in that: The data before and after the drill is from time t1 to t n ~t k ~t m Time series of moments, including hanging weight time series and torque time series Each time series is represented as a collection of 3 small time series: and It means from time t1 to time t n moment, the suspended weight time series and torque time series before drilling, and Indicates that from t n+1 Time~t k At the moment, there are variable hanging weight time series and torque time series in the drilling process. and Indicates that from t k+1 Time~t m Moment, suspended weight time series and torque time series after drilling.

5. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 4 is characterized in that: When lifting the bottom hole idle torque and no-load suspended weight data, it is first necessary to use the time series moving average filter to filter the torque time series. and hanging weight time series Perform outlier removal and smoothing.

6. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 1, characterized in that: The drill string and well wall friction coefficient inversion method is divided into sliding drilling condition inversion and rotation drilling condition inversion.

7. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 6 is characterized in that: In rotary drilling conditions, first determine the drill tool assembly data, wellbore trajectory data, and drilling fluid density data, set the initial friction coefficient, and then calculate the theoretical idling torque of the drill string through the friction torque model until the theoretical idling torque and the actual idling torque meet the error requirements and stop iteration. The friction coefficient at this time is the friction coefficient between the drill string and the well wall.

8. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 6, characterized in that: In sliding drilling conditions, the theoretical no-load hanging weight and the actual no-load hanging weight are calculated until the theoretical no-load hanging weight and the actual no-load hanging weight meet the error requirements, and the iteration is stopped. The friction coefficient at this time is the friction coefficient between the drill string and the well wall.

9. The intelligent dynamic inversion method of the friction coefficient between the drill string and the well wall according to claim 1, characterized in that: In step S1, the drilling conditions include pulling out the drill, drilling down the drill, forward reaming, reverse reaming, single-root connection, rotary drilling, sliding drilling, lifting off the bottom of the well for idling, lifting off the bottom of the well for circulation, sitting, circulation and stationary.

Images