CN-121976745-A - Top drive torsion pendulum parameter control method, device and storage medium

CN121976745ACN 121976745 ACN121976745 ACN 121976745ACN-121976745-A

Abstract

The application relates to the technical field of drilling engineering, in particular to a top drive torsional pendulum parameter control method, a device and a storage medium. The method comprises the steps of determining a first depth of a well bottom drill string reactive torque and a well bottom torque neutral point which reserves a safety margin based on well hole track, top drive torque and screw tool differential pressure data, determining a top drive torque theoretical value by combining a well hole track, a drill string structure, a well body structure and operation parameters, establishing a relation model of torsion angles and torque, calculating a theoretical total torsion angle, obtaining actual measured torque by on-site control, minimizing deviation between the actual measured torque and a calculated value of a material mechanical model containing an optimizable coefficient through a Bayesian optimization algorithm, obtaining an optimized total torsion angle, inputting drilling state parameters such as the optimized angle and the like into a control model based on reinforcement learning training, generating a top drive torsion angle control instruction, continuously updating model parameters according to execution feedback, and realizing intelligent calculation and control of torsion parameters.

Inventors

ZHU ZHAOPENG
WANG YIWEI
SONG XIANZHI
LI GENSHENG
ZHANG CHENGKAI
LIU JIAO
LIU MUCHEN
YANG YANLONG

Assignees

中国石油大学(北京)

Dates

Publication Date: 20260505
Application Date: 20260309

Claims (10)

1. The top drive torsion pendulum parameter control method is characterized by comprising the following steps of: acquiring wellbore trajectory data, top drive torque data and screw pressure difference data; calculating reactive torque of a screw motor based on the screw differential pressure data, and determining a first depth of a well bottom torque neutral point retaining a safety margin; calculating a theoretical value of top drive torque according to the soft rod model and the sectional infinitesimal method based on the borehole track data, the drill string structure data, the well structure data and the operation parameters; Establishing a relation model of the torsion angle and the top drive torque according to the theoretical value of the top drive torque and the first depth, and calculating a theoretical total torsion angle; the theoretical total torsion angle is applied to on-site top drive control to obtain actual measured torque, a Bayesian optimization algorithm is utilized to minimize deviation between the actual measured torque and a torque theoretical value calculated by a material mechanics model containing an optimizable coefficient, and the optimized total torsion angle is obtained; And inputting the drilling state parameters containing the optimized total torsion angle into a control model trained based on a reinforcement learning algorithm to generate a top drive torsion angle control instruction, and updating the parameters of the control model according to state feedback after the control instruction is executed to realize self-adaptive closed-loop control of the top drive torsion parameter.
2. The method of claim 1, wherein calculating screw motor reactive torque based on the screw differential pressure data and determining a first depth of a bottom hole torque neutral point that retains a safety margin comprises: According to the screw differential pressure data, the screw motor reactive torque is determined by combining the screw motor efficiency, the screw motor per-revolution displacement and the screw motor efficiency, and the screw motor reactive torque is determined based on the following formula: Wherein, the For the reaction torque of the screw motor, For each revolution of displacement of the screw motor, For the efficiency of the screw motor to be mentioned, Data for the screw differential pressure; Determining a torque neutral point depth based on the screw motor reactive torque such that drill string friction torque from downhole to the torque neutral point depth is equal to the screw motor reactive torque; subtracting a preset safety margin on the basis of the torque neutral point depth, and determining a first depth of the bottom hole torque neutral point with the safety margin reserved.
3. The method of claim 1, wherein calculating theoretical values of top drive torque based on the borehole trajectory data, drill string structure data, well bore structure data, and operating parameters according to a soft rod model and a segmented micro-element method comprises: Acquiring the well track data, the drill string structure data, the well bore structure data and the operation parameters, wherein the well track data at least comprise well depth, well inclination angle and azimuth angle, the drill string structure data at least comprise pipe column outer diameter and pipe column dead weight, the well bore structure data at least comprise segmented well bore diameter for confirming contact relation, and the operation parameters at least comprise circumferential friction coefficient; Dividing a drill string into a plurality of micro-element sections along the depth direction of a well according to the length of a preset micro-element section, acquiring the well inclination angle and the azimuth angle of each micro-element section based on the track data of the well bore, determining the well inclination angle change rate and the azimuth angle change rate, and calculating the contact force between each micro-element section and the well wall according to the axial force at the lower end of the micro-element section and the component of the dead weight of the pipe column along the normal direction of the well wall, wherein the contact force is determined based on the following formula: Wherein, the In order for the contact force to be such that, For the axial force of the lower end of the micro-element section, For the rate of change of the azimuth angle, For the rate of change of the well inclination angle, The weight of the pipe column is a component along the normal direction of the well wall; Acquiring the torque at the top end of the micro-segment and the torque at the bottom end of the micro-segment, calculating friction torque increment from the first depth to the ground segment by segment based on a micro-segment torque transmission formula according to the contact force, the outer diameter of the pipe column and the circumferential friction coefficient, and determining the theoretical value of the top drive torque in an accumulated manner, wherein the friction torque increment is determined based on the following formula: Wherein, the For the tip torque of the micro-segment, For the torque at the bottom end of the infinitesimal section, For the circumferential friction coefficient of the above-mentioned friction coefficient, For the outer diameter of the pipe string, For the length of the micro-segment.
4. The method of claim 1, wherein the establishing a relation model between the torsion angle and the top drive torque according to the theoretical value of the top drive torque and the first depth, and calculating the theoretical total torsion angle comprise: determining a micro-segment torsion angle based on a micro-segment tip torque, a micro-segment length, a shear modulus of a drill string material, and a polar moment of inertia of the drill string, wherein the polar moment of inertia is calculated from an outer diameter of the drill string and an inner diameter of the drill string; The theoretical total twist angle is determined based on integrating the twist angles of all the hogels in the well depth direction from the first depth to the surface.
5. The method of claim 4, wherein the micro-segment torsion angle at well depth x is determined based on the following formula: Wherein, the For the micro-segment twist angle at the well depth x, For the tip torque of the micro-segment, For the length of the micro-segment, For the shear modulus of the drill string material, A polar moment of inertia for the drill string; Wherein the polar moment of inertia of the drill string is determined based on the following formula: Wherein, the For the outer diameter of the drill string, An inner diameter of the drill string; The theoretical total torsion angle is determined based on the following formula: Wherein, the For the theoretical total torsion angle to be described, Is the first depth.
6. The method of claim 1, wherein the applying the theoretical total torsion angle to the field top drive control to obtain the actual torque, minimizing a deviation between the actual torque and a theoretical torque value calculated by a material mechanics model including an optimizable coefficient using a bayesian optimization algorithm, and obtaining the optimized total torsion angle includes: Driving the torsion pendulum of the drill string based on the theoretical total torsion angle, and collecting corresponding actual measurement torque through a top drive torque sensor; establishing an objective function for measuring the deviation degree of the actual torque and the theoretical torque, and iteratively adjusting the optimizable coefficient by using a Bayesian optimization algorithm to minimize the value of the objective function; determining a total torsion angle calculation formula based on an optimizable coefficient minimized by an objective function, and calculating an optimized total torsion angle, wherein the optimized total torsion angle is determined based on the following formula: Wherein, the For the optimized total torsion angle, At the first depth of the first channel, In order to be able to optimize the coefficients, Is the torque at the top end of the micro-element section, For the shear modulus of the drill string material, As the polar moment of inertia of the drill string, Is a micro-segment length.
7. The method of claim 1, wherein inputting the drilling state parameter including the optimized total torsion angle into a control model trained based on a reinforcement learning algorithm to generate a top drive torsion angle control command, and updating the parameter of the control model according to state feedback after executing the control command, and implementing adaptive closed-loop control of the top drive torsion parameter comprises: taking a current drilling state vector as the input of a strategy network, outputting the mean value and standard deviation of the top drive torsion angle action through the strategy network, and generating continuous torsion angle action by utilizing a heavy parameterization skill, wherein the drilling state vector comprises a borehole track parameter and an optimized total torsion angle output by a Bayesian optimization module; Estimating a state cost function through a cost network, and determining a dominance function based on the state cost function and instant rewards obtained after performing the action; constructing a trust domain objective function based on generalized dominance estimation, and updating strategy network parameters by maximizing the trust domain objective function; Taking the actually measured torque, the borehole track parameter and the tool face state which are obtained after the torsion swing angle action is executed as state feedback for updating the control model parameter of the next round; in the control process, real-time acquisition data and historical working condition data are fused to construct a two-dimensional experience pool for model training and parameter adjustment, and closed-loop updating and control of the top drive torsion parameters are achieved.
8. The method for controlling the top drive torsional pendulum parameter according to claim 7, wherein the dominance function is constructed based on a state cost function: Wherein, the As a function of the advantage of the above-mentioned method, In order to obtain an instant prize after the action is performed, Is in state of Is a value estimate of (1), Is in state of Value estimation of (2); the trust domain objective function that maximizes the generalized dominance estimate is: Wherein, the For the trust domain objective function, The probability of action a is generated for the current policy network, The probability of action a is generated for the old policy network, In order to tailor the function of the object, Clipping parameters for a step size for control policy update.
9. The utility model provides a top drive torsional pendulum parameter control device which characterized in that includes: a memory configured to store instructions; a processor configured to invoke the instructions from the memory and when executing the instructions is capable of implementing a top drive torsion parameter control method according to any one of claims 1 to 8.
10. A machine-readable storage medium having stored thereon instructions for causing a machine to perform the method of top drive torsional parameter control according to any one of claims 1 to 8.

Description

Top drive torsion pendulum parameter control method, device and storage medium Technical Field The application relates to the technical field of drilling engineering, in particular to a top drive torsional pendulum parameter control method, a device and a storage medium. Background In the current oil and gas exploration and exploitation, a top drive (namely a top drive drilling device) is an important drilling rod driving system widely applied, and can remarkably improve the capacity and efficiency of drilling operation. In the sliding drilling process of the directional well, the drilling pressure applied by the ground is difficult to be effectively transferred to the drill bit due to larger friction resistance between the drilling tool and the well wall or the casing, and the pressure supporting phenomenon is easy to occur. The underpressure not only affects cuttings carry over and wellbore cleaning, but also increases the risk of differential sticking. Once the backing pressure is suddenly released, the accumulated axial force and torque can be released instantaneously, so that the tool face is severely fluctuated, the fatigue damage of the downhole tool is aggravated, and the drilling efficiency is severely restricted. Currently, one common approach to solving such problems is the top drive torsional pendulum technique. The top drive torsion pendulum technology drives the drill string to periodically twist left and right by controlling the top drive, so that friction resistance is reduced, and transmission of drilling pressure is improved. The existing top drive torsion pendulum technology relies on subjective experience, and parameters are calculated and set manually, so that the problems of lack of real-time adjustment capability, large error, multiple errors and the like exist. Disclosure of Invention The embodiment of the application aims to provide a top drive torsion pendulum parameter control method, a device and a storage medium, which are used for solving the problems that in the prior art, the top drive torsion pendulum parameter depends on manual experience setting, and the deviation between field actual measurement torque and theoretical prediction is larger due to the lack of real-time feedback and self-adaptive adjustment capability. In order to achieve the above object, a first aspect of the present application provides a method for controlling a top drive torsional pendulum parameter, including: acquiring wellbore trajectory data, top drive torque data and screw pressure difference data; calculating reactive torque of the screw motor based on the screw differential pressure data, and determining a first depth of a well bottom torque neutral point retaining a safety margin; Calculating a theoretical value of top drive torque according to the soft rod model and the sectional infinitesimal method based on the borehole track data, the drill string structure data, the well structure data and the operation parameters; Establishing a relation model of the torsion angle and the top drive torque according to the theoretical value of the top drive torque and the first depth, and calculating a theoretical total torsion angle; applying the theoretical total torsion angle to the field top drive control to obtain actual measured torque, and minimizing deviation between the actual measured torque and a torque theoretical value calculated by a material mechanics model containing an optimizable coefficient by using a Bayesian optimization algorithm to obtain an optimized total torsion angle; And inputting the drilling state parameters containing the optimized total torsion angle into a control model trained based on a reinforcement learning algorithm to generate a top drive torsion angle control instruction, and updating the parameters of the control model according to state feedback after the control instruction is executed to realize self-adaptive closed-loop control of the top drive torsion parameters. In the embodiment of the application, based on screw differential pressure data, calculating the reactive torque of the screw motor and determining the first depth of the well bottom torque neutral point for reserving the safety margin comprises the steps of determining the reactive torque of the screw motor according to the screw differential pressure data by combining the screw motor efficiency, the displacement per revolution of the screw motor and the screw motor efficiency, wherein the reactive torque of the screw motor is determined based on the following formula: Wherein, the For the reaction torque of the screw motor,For each revolution of the screw motor,For the efficiency of the screw motor,The method comprises the steps of obtaining screw differential pressure data, determining a torque neutral point depth based on screw motor reactive torque, enabling drill string friction torque from a bottom hole to the torque neutral point depth to be equal to the screw motor reactive torque, subtracting